即时简报

• ChatGPT 官方更新 · web:chatgpt-updates:2026-09-09:2e093e4813d8ec1a88969ad37c861ae7f57ddc43603324567bfe82be339e27ab · help.openai.com ↗ · 发布：2026-09-09T00:00:00+00:00 · 采集：2026-09-17T13:11:16.322077+00:00

## **Updated models and usage limits in ChatGPT Voice**

ChatGPT Voice can now use GPT-5.6 or GPT-6 Astra when it needs to search or reason through harder questions. Choose your model and reasoning effort using the same controls as text chat. Available models and their usage limits depend on your plan.

We’re also simplifying GPT-Live daily usage limits:

-   Go: Up to 3 hours with GPT-Live-1 mini, replacing GPT-Live-1 access.

-   Plus: Up to 3 hours with GPT-Live-1.

-   Pro ($100/month): Up to 15 hours with GPT-Live-1.

-   Pro ($200/month): Unlimited GPT-Live-1 usage.

Plus and Pro no longer switch to GPT-Live mini after reaching a Voice limit.

With these updates, the separate Instant/Medium/High Voice intelligence levels are deprecated.

Learn more: [https://help.openai.com/en/articles/20001274-chatgpt-voice](https://help.openai.com/en/articles/20001274-chatgpt-voice)

## **Share files and folders from your Library**

Library sharing lets you share files and folders in ChatGPT, choose who can access them, and work with shared content directly in your conversations.

### **What’s new**

-   **Share with specific people.** Invite recipients and assign Viewer or Editor access.

-   **Share across your workspace.** Make content available to everyone in your workspace.

-   **Manage access.** Review who has access, change permissions, or remove access from the sharing dialog.

-   **Find and use shared content.** Access items in **Shared with me** and use shared files in conversations with ChatGPT.

### **Shared folders and ownership**

Files uploaded to a shared folder belong to the folder’s owner. The person who uploaded each file is still recorded.

If you upload a file to someone else’s shared folder and they later remove your folder access, the file stays in their folder and you lose access to it.

## Deep Research in ChatGPT Work and Codex

Deep research is now available in ChatGPT Work and Codex. Research complex questions across the web, your files, and supported connected apps, then turn the findings into an editable document with citations. You can steer the research as it runs and request a document, presentation, spreadsheet, or Site, depending on available tools.
To start, type @Deep Research in Work or explicitly ask for deep research. Available to Plus, Pro, Business, Enterprise, and Edu users with Work access on web, desktop, iOS, and Android. It uses your existing Work/Codex allowance or credits; deep research limits in Chat are unchanged. [Learn more](https://help.openai.com/articles/10500283).

• Swift.org · rss:a76d8c3e94eba61fc46b5cc05fc51e5f6df1d0f46d901d32e9c27bb251f155ae:swift-blog:18 · swift.org ↗ · 发布：2026-09-04T16:30:00+00:00 · 采集：2026-09-17T07:07:56.208180+00:00

Welcome to “What’s new in Swift,” a curated digest of releases, videos, and discussions in the Swift project and community. Here’s an update from guest contributor Simon Leeb on Swift’s progress as a language for web scenarios: Hi, Simon here! I am the creator of the elementary-swift project, a collection of packages born from a simple wish: I want to build web UIs in Swift and ultimately help Swift become a first-class choice for the web. This journey began after I started using Swift for backend services. The web frontend, however, still lived in a separate ecosystem, and I really wanted it to feel as ergonomic, safe, and efficient as the Swift I was writing everywhere else. That led to the creation of Elementary: a modern and efficient HTML rendering library with a familiar declarative API, built for the web. It integrates easily with frameworks like Vapor and Hummingbird, and has become a practical option for server-rendered web UIs. Around that same time, years of community work in the swift-wasm project made compiling Swift to WebAssembly increasingly viable, while Embedded Swift was taking its first experimental steps. This made me wonder: “How hard can it be to use Embedded Swift and build a state-driven web UI framework that produces tiny WebAssembly binaries?” Turns out: quite hard, actually! But it was too late. Despite my better judgment, I was in the middle of creating what is now known as ElementaryUI. Where Elementary renders HTML on the server, ElementaryUI runs in the browser itself. You can watch my talk at Swift@FOSDEM 2026 if you want to know more about the why, what, and how. To showcase where the project is heading, I recently posted a small Full-Stack Swift on Cloudflare demo. It features Swift in the browser communicating with a Swift backend on an edge worker through shared message types. I hope it gives people a concrete sense of how much the core technologies and the surrounding tooling have advanced. ElementaryUI is still young, with plenty left to build. Visit elementary.codes to try it, share feedback, contribute, or sponsor its development. Let’s work together and make Swift a first-class choice for the web! Now on to other news about Swift: Videos to watch Saleem Abdulrasool joined the Empower Apps podcast to discuss Swift on Windows, server-side Swift, SwiftWin32, Swift’s C++ interop, and how to get started. Building memory-safe software? Write security-sensitive code in Swift covers how Swift guarantees safety across bounds, lifetimes, types, initialization, and concurrency, with primitives like Span and non-copyable types, plus how to audit unsafe code with strict memory safety and incrementally migrate existing C modules. Two short videos about running Embedded Swift on Raspberry Pi Pico: a video on getting started on macOS called Let it blink! and Fun with traffic lights in 60 seconds. Community highlights Building scalable backend apps in Swift shares an approach to structuring server-side Swift codebases, separating business logic from database and framework details so the code stays easier to test and change over time. The Swift Package Index blog explains what a package registry actually is, how it fits alongside SwiftPM and package indexes, and walks through an example of switching a dependency managed from Git to a Swift registry. Embedded Swift Improvements Coming in Swift 6.4 rounds up what’s ahead for Swift on microcontrollers and other constrained environments. The Browser Company’s Swift on Windows: A year of refinement reflects on a year of work making Swift on Windows more predictable. New package releases Write an interface once with SwiftTUI using a declarative, state-driven syntax, then ship it as a terminal app, as a native macOS or iOS app, as an Android app, or as a WASI build for the browser. Tired of hand-writing RawRepresentable and LosslessStringConvertible conformances? lexic generates them for you via macros, and runs the same on Linux as on Apple platforms. StructuredQueries, Point-Free’s SQLite query builder, now has fully type-safe support for JSON and JSONB columns, including a json_each table function, so nested data can be queried and updated by key path without leaving Swift’s type system. Swift Evolution The Swift project adds new language features through the Swift Evolution process. These are some of the proposals currently under review or recently accepted for a future Swift release. Under active review: ST-0029 Include additional issue metadata in event stream - Today, Swift Testing’s JSON event stream reports only bare-bones details when an issue occurs, making it hard to tell a thrown error apart from a manual Issue.record call. This proposal adds structured fields, including error, confirmationMiscount, exceededTimeLimit, and expression, so tools like Xcode and VS Code can show richer, more specific failure information. Recently accepted: SE-0544 Mutation and consumption in non-copyable type deinits - Non-copyable types that manage a resource, like a file handle or buffer, often need to run the same cleanup logic in their deinit that they use elsewhere, but until now self inside a deinit could only be borrowed, not mutated or consumed. This proposal lets a deinit mutate or consume its own stored properties directly, so existing cleanup methods can be reused instead of duplicated. ST-0028 Revise Swift Testing’s Attachment/Encodable interop - Swift Testing lets you attach extra data, like a screenshot or JSON snapshot, to a test for inspecting after a failure, but attaching custom types previously required extra setup code and offered no way to choose the encoding format. This proposal adds new Attachment initializers that let you attach Encodable or NSSecureCoding values directly, picking the format or supplying your own encoder. Recently accepted with modifications: SE-0536 Package Registry Search - To use a package from a registry today, you already have to know its exact identifier, since there’s no standard way to discover packages within a registry the way other package ecosystems allow. This proposal adds an optional /search endpoint to the registry specification and a swift package-registry search subcommand, letting you find packages by name, scope, author, and other criteria, with support for qualifiers like author:"Mona Lisa Octocat" and searches that span every configured registry at once. SE-0516 Iterable - Looping over a collection in Swift traditionally means copying out one element at a time, which doesn’t work for newer types that can’t be copied, like Span and InlineArray. This proposal introduces Iterable, a new way to loop over data without copying, and was renamed from BorrowingSequence and given support for typed throws before acceptance. ST-0026 TaskLocal test trait - Task-local values are like settings, such as a feature flag, that apply only within a single task. Overriding one in a test previously meant writing a custom trait from scratch, but this proposal adds a .taskLocal(_:_:) trait that does it in one line, like @Suite(.taskLocal(FeatureFlags.$isEnabled, true)).

• ChatGPT 官方更新 · web:chatgpt-updates:2026-09-03:2dff4d5ae213e2a9c0be4b5ac67d033c437fceeefcd798bab96043a47332d8eb · help.openai.com ↗ · 发布：2026-09-03T00:00:00+00:00 · 采集：2026-09-17T13:11:16.322077+00:00

## **Introducing GPT-6 Astra**

Today we’re introducing GPT-6 Astra, with improvements in coding, research, computer use, and complex, multi-step work. Astra can create documents, spreadsheets, and presentations that follow your templates and instructions, and adapt when you add requirements or change direction.

Access is rolling out to a limited set of organizations. Astra is not yet generally available. Broader availability is planned over the coming days.

Astra includes additional safety monitoring to look for cases where agents may not have interpreted your instructions correctly. If a potential case is detected, the conversation may be paused or stopped as a precaution for you to review and decide how to proceed.

## Zendesk & OneNote Plugins in ChatGPT and Codex \[Beta\]

Today, we added the Zendesk and OneNote plugins in the Plugin directory. The Zendesk plugin helps teams review support tickets and customer history, find relevant knowledge, and prepare replies in ChatGPT and Codex. Members connect their own Zendesk accounts to work with the support information they can access.

The OneNote plugin helps you find and summarize notes, gather decisions and action items, and create or update notes through supported actions in ChatGPT and Codex.

Read more: [Zendesk Plugin](https://help.openai.com/articles/20001512), [OneNote Plugin](https://help.openai.com/articles/20001511).

## Share Sites with people outside your workspace

Eligible Site owners can now share a live ChatGPT Site with named people outside their workspace, without making the Site public. External viewers can use the shared Site but cannot edit or publish it.
To share, open the Site, select **Share**, enter the recipient’s email, and save their viewer access. The recipient signs in with the account that was granted access. You can review or remove viewers in the Site’s sharing controls. Learn more in [Creating and managing ChatGPT Sites](https://help.openai.com/en/articles/20001339).

• ChatGPT 官方更新 · web:chatgpt-updates:2026-09-01:6d5eac972efc82c82783320b102bdcfe5f3f540015dcae7a7f8c543ce6885854 · help.openai.com ↗ · 发布：2026-09-01T00:00:00+00:00 · 采集：2026-09-17T13:11:16.322077+00:00

## **Healthcare Public Data in ChatGPT for Clinicians**

Eligible ChatGPT for Clinicians users in the United States can now use Healthcare Public Data in ChatGPT. The plugin brings together nine apps for searching public healthcare sources, including biomedical research, clinical trials, medication information, Medicare data, and provider records.
To get started, install the plugin from the **Plugin directory**, then connect the apps you want to use. These apps are read-only and do not access patient charts. Do not include protected health information in searches sent to public sources.
For details, see: [Using Healthcare Public Data in ChatGPT and Codex](https://help.openai.com/articles/20001489).

• ChatGPT 官方更新 · web:chatgpt-updates:2026-08-31:cc6152cf568c63c38d4e55339d2c8bc28ff4ebeb92eabb823744f746d745424f · help.openai.com ↗ · 发布：2026-08-31T00:00:00+00:00 · 采集：2026-09-17T13:11:16.322077+00:00

## **Create and share personalized sticker packs**

Turn an idea or a photo into a set of stickers you can use in your messages.

-   **Start with a prompt or template.** Ask ChatGPT to make a sticker pack, or choose a sticker template from the home screen or Images.

-   **Share your stickers in messaging apps.** Download your pack or add it to iMessage or WhatsApp on supported devices.

Available to all ChatGPT users on mobile globally. Update the ChatGPT app to see the new experience.

## **See Live voice content on your Lock Screen**

Content from a Live voice conversation can now appear on your iPhone’s Lock Screen and, on iPhones with Dynamic Island, in the Dynamic Island.

-   **Follow along outside the app.** Live Activities can show content from your voice conversation while you use another app or your phone is locked.

To continue a Voice conversation outside ChatGPT, turn on Background conversations in Settings → Voice.

[Learn more about ChatGPT Voice](https://help.openai.com/en/articles/20001274).

## **Use website tools in the desktop browser**

ChatGPT Work and Codex can now use tools that supported websites provide directly in the desktop app’s built-in browser.

-   **Work directly with a website’s tools.** ChatGPT can discover tools for your task without a separate connection. Select the arrow in the address bar to see the tools a page provides.

Site tools use WebMCP and require a supported account, model, and webpage. They are available in the built-in desktop browser, not through the Chrome extension. Existing website-access and sensitive-action confirmations still apply.

[Learn more about site tools](https://help.openai.com/en/articles/20001423-using-site-tools-in-the-chatgpt-desktop-app).

## **Use the browser extension in more browsers**

The ChatGPT browser extension now supports Microsoft Edge, Brave, Opera, and Vivaldi.

-   **Bring your browser into Work and Codex.** Mention open tabs to use them as context, or ask ChatGPT to complete a browser task from the desktop app.

Side chat is available in Edge, Brave, and Vivaldi. Opera supports tab mentions and browser control, but not side chat. Update the desktop app before setup; availability depends on your workspace settings.

[Set up the ChatGPT browser extension](https://learn.chatgpt.com/docs/chrome-extension).

## **Improved pronunciation help in ChatGPT**

When you ask how to pronounce a word or phrase, you can now tap to hear it aloud and see a phonetic breakdown

• Swift.org · rss:a76d8c3e94eba61fc46b5cc05fc51e5f6df1d0f46d901d32e9c27bb251f155ae:swift-blog:4 · swift.org ↗ · 发布：2026-03-02T14:05:00+00:00 · 采集：2026-09-17T07:07:56.208180+00:00

Welcome to this edition of “What’s New in Swift,” where we share interesting projects and updates in the Swift community that deserve visibility to a broader audience. FOSDEM is the largest open source conference in the world, and this year Swift was featured in talks across multiple tracks and a dedicated pre-conference event. Here’s a report from the Swift community: Hey there! Karen Chu here. 👋 I help grow and support the Swift open source community, and one of my favorite things has been increasing Swift’s participation in open source events. As a follow-up to Swift’s presence at FOSDEM last year, this year we expanded our approach by running our own Pre-FOSDEM fringe event, which included 11 talks across multiple platforms, in addition to four Swift talks across various devroom tracks. Though it was my first year at FOSDEM, it was delightful to meet several people who attended the Swift devroom last year and also came through again this year. With the vast array of attendee backgrounds (I met recent grads, former Swift mentees, Swift workgroup members, folks from the Swift Core Team, and more) and talks covering different platforms (embedded, server, BSD, Android, and more), that weekend showed me just how much momentum Swift has gained as a general purpose language that can benefit everyone. I wanted to highlight Simon Leeb’s talk, Swift in the Browser with ElementaryUI, as a popular talk from the event which showed how to run Swift applications natively in the browser with WebAssembly. FOSDEM may be over (for now), but I’m already excited about how the Swift community will show up next year! To catch up on all the talks at the Pre-FOSDEM event, check out the YouTube playlist. Now on to other news about Swift: New package releases Originally developed for searching financial databases for strings like stock tickers, FuzzyMatch brings high-performance fuzzy string matching to Swift. Now you can expose GraphQL APIs in your applications with new Vapor and Hummingbird packages. Using Microsoft SQL Server? SQLClient-Swift is a native client for iOS, macOS, and Linux that’s built using modern Swift concurrency. Developers now have more options to create terminal-based Swift applications with TuiKit, a declarative, SwiftUI-like framework for building Terminal User Interfaces. TuiKit joins Noora, which provides themeable components for CLIs. Lastly, Swift System Metrics 1.0 was announced, making the collection of process-level metrics easy. Videos to watch Interested in writing Swift using agents? At Swift Barcelona, a talk on Building Agentic Apps with MCP in Swift demonstrated several open source tools. And if you’re an Xcode user, check out a recent code-along covering coding intelligence in Xcode 26. Introducing the Swift SDK for Android covered the current state of Swift on Android, how to call Java and Kotlin APIs from Swift, UI strategies, and what’s actively being worked on. SwiftCrossUI: Swift apps, everywhere presented a UI solution for Swift applications that run on Android, Linux, and Windows, using native UI components on each platform. The project also recently published a development update in video form, covering contributions and recent changes. Ever wanted to control microcontrollers in Swift? CoreAVR + ArduinoKit: Swift on the Arduino, the Swift way introduces ArduinoKit, which acts as the Arduino standard library for Swift, and CoreAVR, a hardware abstraction layer for AVR microcontrollers written in Swift. Community highlights Swift is participating in Google Summer of Code 2026! 🎉 A recent forums post announced the process and outlined key dates to learn more. This is a great opportunity if you’re a student interested in contributing to Swift! A community member recently blogged about their experience profiling Swift applications on Windows and macOS with Tracy, covering how to visualize and analyze their application performance. Swift Evolution The Swift project adds new language features through the Swift Evolution process. These are some of the proposals currently under review or recently accepted for a future Swift release. Under active review: SE-0514 Hashable Conformance for Dictionary.Keys, CollectionOfOne and EmptyCollection - Three standard library collection types can’t currently be used in sets or as dictionary keys. This proposal adds Hashable conformance to all three, for consistency with the rest of the standard library. SE-0513 API to get the path to the current executable - There’s currently no portable, reliable way to get the path to the currently running executable in Swift. This proposal adds CommandLine.executablePath, a new property in the standard library that provides a consistent way to get this value across all platforms Swift supports. Recently accepted: SE-0506 Advanced Observation Tracking - @Observable types automatically track when their properties change. However, advanced use cases like developing middleware infrastructure or widget systems require more control and features. This proposal adds options to the existing withObservationTracking to control when/which changes are observed, and a continuous variant that re-observes automatically after coalesced events. SE-0502 Exclude private initialized properties from memberwise initializer - When you define a struct in Swift, the compiler automatically generates an initializer that takes each property as an argument. Currently, adding a private property with a default value to a struct forces that auto-generated initializer to become private, which breaks callers outside of the type. This proposal fixes that behavior. Have any suggestions for a future issue? Drop them in the Community Showcase on the Swift Forums. See you next time!

• Swift.org · rss:a76d8c3e94eba61fc46b5cc05fc51e5f6df1d0f46d901d32e9c27bb251f155ae:swift-blog:9 · swift.org ↗ · 发布：2026-05-01T10:30:00+00:00 · 采集：2026-09-17T07:07:56.208180+00:00

Welcome to “What’s new in Swift,” a curated digest of releases, videos, and discussions in the Swift project and community. The 1.0 release of valkey-swift was recently announced on the Valkey blog. We’ve invited one of the authors to be this month’s guest contributor: Hi, I’m Adam Fowler, an open source developer working in the Swift on server ecosystem. I am excited to announce the 1.0 release of valkey-swift - a production-grade Swift client for Valkey. Valkey is a high-performance datastore commonly used as a caching layer or message broker in server applications. It is an open source fork of Redis. Valkey-swift is a client library targeted at Valkey servers but is equally capable of working with Redis. It is built from the ground up with Swift 6 and structured concurrency. Every Valkey command returns typed responses checked at compile time, and strict concurrency checking is enabled throughout so that data races are caught by the compiler, not in production. Connections and subscriptions are all scoped through structured concurrency, so resources clean up automatically. The client covers every standard Valkey command, auto-generated from Valkey’s own command specifications to stay in sync as the server evolves. Previously, the de facto client library for Redis was RediStack, which was built on top of pre-concurrency concepts. Retrofitting structured concurrency would have been awkward and some of the new features in valkey-swift infeasible. Around the same time Redis changed its licensing structure and the open source fork Valkey was created. So it felt like a good time to make a clean break and build a new library. If you’re building server-side Swift and need a fast key-value store, add valkey-swift via Swift Package Manager, and you’re ready to go. If you are using RediStack to connect with a Redis server, we have a guide to help you migrate to valkey-swift. Complete documentation is available, and contributions are welcome on GitHub. Now on to other news about Swift: Videos to watch The try! Swift Tokyo 2026 conference featured not one, but two exciting talks about Embedded Swift: Getting started with Embedded Swift is a short and accessible introductory talk. Learn about writing Swift using embedded simulators, along with code examples to run Swift code on devices including the Game Boy Advance! Learn by Building: Bare-Metal Programming with Embedded Swift is a deeper look at Embedded Swift. Follow along by trying the sample code for five bare-metal Raspberry Pi Pico examples that are featured in the talk. Want to learn more about Swift concurrency from engineers who have designed and used its features? Check out this live online Q&A on Swift concurrency. Nil Coalescing published a new video on some of the lesser known options for working with optionals in Advanced Techniques for Working with Optionals in Swift. New package releases IndustrialKit is a framework for designing, programming, and controlling robotic systems that was recently discussed on the Swift forums. swift-tar is a pure Swift library for reading, writing, and extracting TAR archives. swift-tar is cross-platform, works without requiring any system framework or Foundation, and supports GNU & PAX extensions. Xylem is a pure Swift XML parser with zero dependencies, covering SAX, DOM, and XPath 1.0. Swift Evolution The Swift project adds new language features through the Swift Evolution process. These are some of the proposals currently under review or recently accepted for a future Swift release. Under active review: SE-0529 Add FilePath to the Standard Library - FilePath in the swift-system package parses platform-specific path syntax on the developer’s behalf, provides a normalized view of path components, and enables filesystem resolution. However, shipping in an external package means the standard library, Swift runtime, and toolchain libraries such as Foundation cannot depend on it. This proposal adds FilePath and its associated types to the Swift module, alongside essential functionality for construction, decomposition, resolution, and C interoperability. Recently accepted: Vision for Networking - Swift’s networking ecosystem is getting an overhaul. This vision document proposes three initial areas of focus: evolve HTTP APIs including new HTTP client and server implementations, define currency types to reduce duplicated effort and integration friction, and define a unified networking stack. SE-0517 UniqueBox - Sometimes in Swift it’s necessary to manually put something on the heap that wouldn’t otherwise live there. This proposal introduces a new type in the standard library, UniqueBox, which is a smart pointer type that uniquely owns a value on the heap. ST-0022 Custom reflection during testing - When a test fails, Swift Testing reflects the values involved to help diagnose the failure, but types have no way to customize what appears in that output. This proposal adds a customization point, CustomTestReflectable, for developers to specify exactly what should be included in test output, whether they want to simplify, obscure, extend, or reformat that information. One more thing Have you recently looked at the Swift.org community page? It includes updated content, plus a new How we work page that describes opportunities to get involved.

• ChatGPT 官方更新 · web:chatgpt-updates:2026-07-16:78df73fea798b133906ad1e0282ad31008cf40a782c3f9f7606586042a9da082 · help.openai.com ↗ · 发布：2026-07-16T00:00:00+00:00 · 采集：2026-09-17T13:11:16.322077+00:00

## **ChatGPT desktop app experience updates**

We’ve updated the ChatGPT desktop app to make it easier to choose between Chat and Work, find your conversations and Projects, and continue Work across devices. These updates are now live for all plans on macOS and Windows.

### **macOS and Windows**

-   **A clearer desktop layout:** A global switcher lets you choose between ChatGPT and Codex. In ChatGPT, choose Chat for quick questions and conversational help, or Work to complete tasks end to end.

-   **Unified Recents:** Chat and Work conversations now appear together in Recents, where you can sort, filter, and pin them.

-   **Projects in the desktop app:** Your existing ChatGPT Projects now appear in the app. You can start a Chat conversation inside a Project or begin a Work thread using Project context.

-   **Continue Work across devices:** Cloud Work conversations now sync across web, mobile, and desktop, so you can start on one surface and continue on another. Local conversations stay on your computer.

-   **Codex remains unchanged:** As part of this update, Codex remains a separate view, and its workflows and history are unchanged. Quick Chats remain available for fast conversations.

• ChatGPT 官方更新 · web:chatgpt-updates:2026-08-20:09da470760e511dd742d6d51fb5e8ef0d4a1d6dba304a0b0dd28f102482370d7 · help.openai.com ↗ · 发布：2026-08-20T00:00:00+00:00 · 采集：2026-09-17T13:11:16.322077+00:00

## **Codex and ChatGPT feature updates**

-   **Choose a new URL for a ChatGPT Site.** Site owners on Plus and Pro can change an existing Site’s ChatGPT-hosted URL without redeploying it. The previous address redirects to the new URL, including routes and query parameters, while custom domains remain separate and unchanged. In the Site’s settings, select **Change URL**. [Learn more about creating and managing ChatGPT Sites](https://help.openai.com/articles/20001339).

-   **Use Apple Messages from Codex and ChatGPT Work.** On Apple silicon Macs, the Apple Messages plugin in the ChatGPT desktop app can read and search iMessage, SMS, and RCS conversations and prepare or send messages through Messages. By default, ChatGPT asks the user to approve the message and recipients before sending. [Learn more about the Apple Messages plugin](https://developers.openai.com/codex/plugins?surface=app#app-use-apple-messages-from-codex).

-   **Use Computer History in more regions.** Computer History is now available in the EEA, Switzerland, and the United Kingdom for Pro users in the ChatGPT desktop app on macOS. It is off by default and requires Memories. [Learn more about Computer History](https://developers.openai.com/codex/customization/computer-history).

-   **Share a read-only snapshot of a Codex chat.** Personal share links can be opened by anyone with the link. A snapshot is static and does not include tool calls or shell input and output. Codex redacts known secret patterns, but users should review the snapshot because sensitive paths, diffs, images, or other content may remain. [Learn more about sharing a Codex chat](https://developers.openai.com/codex/use-chatgpt#share-a-read-only-snapshot-of-a-codex-thread).

-   **Keep pinned chats in sync.** Pinned chats now stay in sync between the ChatGPT desktop app and iOS when you use the same Codex account. Android is not included in this release. [Learn more about organizing projects and chats](https://developers.openai.com/codex/projects?surface=app#app-organize-projects-and-chats).

• ChatGPT 官方更新 · web:chatgpt-updates:2026-08-13:5976278f75fac909feead4e12c63612882e3033aaffba7d34051843ca5e3a280 · help.openai.com ↗ · 发布：2026-08-13T00:00:00+00:00 · 采集：2026-09-17T13:11:16.322077+00:00

## **Google Drive is now in Library**

If you have the Google Drive plugin connected, you can now see and browse your Google Drive files and folders directly from Library, including items shared directly with you. You can also quickly pull up a Drive file from the composer or with @mentions and add it to any chat—without uploading it again.

When you’re working with a file, you can now keep Google Docs, Sheets, and Slides open beside the conversation while asking ChatGPT to summarize, analyze, compare, or create something new from them. You can also select a folder and ask ChatGPT to work across the files it contains. Your content stays connected to Drive, so it’s easy to return to the original. Where supported and authorized, ChatGPT can update the source file directly.

### **Availability**

The initial experience includes My Drive and files and folders shared directly with you; Shared Drives aren’t included yet. Some Google Drive editing and collaboration features aren’t yet available in ChatGPT.

Rolling out to Plus, Pro, Enterprise, Edu, Healthcare and Business users on the web in both the Chat and Work toggles. Mobile support will follow.

## **Computer History for macOS**

Computer History is an optional feature in the ChatGPT macOS app that lets ChatGPT and Codex reference selected activity from apps and websites, helping users continue work without re-explaining every detail. It records interaction events—not screenshots, screen recordings, microphone input, or system audio—and private browsing is not included.
Computer History is off by default and available to Pro, Business, and Enterprise users. Business and Enterprise admins must grant access before members can opt in. Users can pause Computer History, choose included apps and sites, and inspect or delete timeline items. Note that Computer History is not currently available in the EEA, UK or Switzerland.

• ChatGPT 官方更新 · web:chatgpt-updates:2026-07-29:442c66de1697bbf05d85d10a9460063aa22f7db902c3a13673f3532156e31019 · help.openai.com ↗ · 发布：2026-07-29T00:00:00+00:00 · 采集：2026-09-17T13:11:16.322077+00:00

## ChatGPT for Academic Researchers

Research rarely happens alone. ChatGPT for Academic Researchers gives faculty and postdoctoral researchers the ability to apply for 12 months of complimentary access to a dedicated ChatGPT workspace for a small, verified team. Each workspace supports up to five members and includes business data protections and ChatGPT Pro-level usage limits, helping researchers collaborate while keeping team research in its own workspace.
To apply, researchers sign in with an institutional email, verify their affiliation through SheerID, and share a qualifying paper and a short description of their intended research use. Approved applicants complete a $0 checkout with a valid card, create the workspace, and can invite up to four verified collaborators from the same institution. Researchers at existing ChatGPT Edu institutions can still apply, but when approved, follow an institution-managed Edu route instead. The program covers ChatGPT access, not API credits.
Read the full eligibility and program details in [ChatGPT for Academic Researchers](https://help.openai.com/articles/20001406).

## Sign in with ChatGPT

We’re beginning to roll out Sign in with ChatGPT across select plugins and partner sites, starting with Airtable, GitLab, HubSpot, Notion, Supabase, and Vercel.
When you connect a supported app from the plugin directory ChatGPT, you can use Sign in with ChatGPT to create or link an account with that service in fewer steps. On participating partner sites, you can also choose ‘Sign in with ChatGPT’ to create or access your account.
This means less time on account setup and an easier way to start working with these tools from ChatGPT and Codex.
When you sign in, only your name, email, and profile picture (if you have one) are shared with the partner. You’ll still separately review and approve the access each plugin receives.

• ChatGPT 官方更新 · web:chatgpt-updates:2026-08-14:17899aefc914aea3a2164af10ff711fdc350fa8775f5494fdb9b115fe08344ae · help.openai.com ↗ · 发布：2026-08-14T00:00:00+00:00 · 采集：2026-09-17T13:11:16.322077+00:00

## **ChatGPT app experience updates**

We’ve made it easier to study with interactive quizzes, manage project memory, and move between typing, dictation, and desktop work.

### **Learning, Projects, and personalized suggestions**

-   **Practice with interactive quizzes.** Ask ChatGPT to quiz you on a topic and answer questions directly in your conversation. Available to all consumer ChatGPT plans and Edu plans on web and mobile.

-   **Change an existing project’s memory setting.** For eligible unshared projects, switch between default and project-only memory without starting a new project. Shared projects remain project-only.

-   **Get suggestions that fit how you use ChatGPT.** Eligible paid users can see homepage suggestions based on their conversation history and connected tools.

### **Web**

-   **Use Think on Free and Go.** Select Think on the web when you want ChatGPT to reason through a harder question.

### **Android**

-   **Switch between typing and dictation.** The dictation button stays available after you add text or attachments, so you can continue building the same message.

-   **See suggestions that stay in sync.** Composer suggestions update as you type or delete text.

-   **Keep typing with a physical keyboard.** Connected keyboards keep the composer focused, preventing interruptions while you write.

### **Linux desktop**

-   **Use ChatGPT and Codex on Linux.** The desktop app is now available globally in public preview on Ubuntu 24.04 LTS and 26.04 LTS, Debian 13, and Fedora 43 and 44.

-   **Work in the built-in browser or Chrome.** Browser actions are available in the Linux app; controlling other desktop apps is not yet supported.

[Download ChatGPT for Linux](https://chatgpt.com/download/).

## **Edit memory settings for existing projects**

You can now change a project’s memory setting after you create it. Open the project, select the three-dot menu, choose Project settings, and select Default memory or Project-only memory under Memory.

With project-only memory, ChatGPT can use conversations from the same project for context, but it won’t reference memories or conversations from outside the project. Information from the project is also kept out of memory used in chats outside the project.Changes may take a few hours to take effect.

This update is available on all ChatGPT plans. Shared projects continue to use project-only memory and cannot be switched to default memory. ChatGPT Work is not available in projects using project-only memory.

Learn more about [Projects in ChatGPT](https://help.openai.com/en/articles/10169521-projects-in-chatgpt).

• Swift.org · rss:a76d8c3e94eba61fc46b5cc05fc51e5f6df1d0f46d901d32e9c27bb251f155ae:swift-blog:17 · swift.org ↗ · 发布：2026-08-20T13:45:00+00:00 · 采集：2026-09-17T07:07:56.208180+00:00

Embedded Swift is a subset of Swift that’s designed for low resource usage, making it capable of running on constrained environments like microcontrollers. Using a special compilation mode, Embedded Swift produces significantly smaller binaries than regular Swift. While a subset of the full language, the vast majority of the Swift language works exactly the same in Embedded Swift. Additional information is described in the Embedded Swift vision document. Embedded Swift is evolving rapidly. Following our updates on Embedded Swift improvements in Swift 6.3 late last year, this post describes a number of additional improvements made in the upcoming Swift 6.4 release. You can try them out today with a Swift development snapshot. Language improvements Embedded Swift continues to expand its subset of the language to include more aspects of “full” Swift, making it easier than ever to bring compatibility with Embedded Swift to existing Swift code bases. Many of these features have some dynamic aspect to them, meaning that they have an impact on runtime performance (for example, due to indirect calls) and code size (due to requiring additional metadata). However, this impact only occurs where these dynamic language features are actually used: code that is highly sensitive to code size and performance can choose to avoid them, for example by enabling warnings in the PerformanceHints diagnostic group. Generalized support for existential (any) types any) types section" href="#generalized-support-for-existential-any-types"> Embedded Swift previously only supported existential (any) types that had an AnyObject constraint, meaning they could only be used with class instances. Now, all any types are available in Embedded Swift, including Any itself. For example: protocol P { func method() } extension Int: P { func method() { print("\(self) is here") } } let a: any P = 17 a.method() // prints "17 is here" The Embedded Swift generics compilation model, which requires that all generic functions and types eventually be specialized, implies some limitations on the use of any types. Specifically, a generic function cannot be called on an any type: extension P { func genericMethod<T: P>(_ other: T) { ... } } let a: any P = 17 a.genericMethod(a) // error: cannot use generic instance method 'genericMethod' on a value of type 'any P' in Embedded Swift Untyped throws Embedded Swift previously only allowed throwing specific error types, like this: func parseRecord() throws(ParsingError) -> Record { ... } “Untyped” throws, which can throw any Error-conforming instance, was previously disallowed in Embedded Swift: func loadImage() throws -> Image { ... } // previously disallowed in Embedded Swift Untyped throws is equivalent to throwing a value of type any Error. With the generalization of any types, Embedded Swift now fully supports untyped throws. Throwing a value of any Error typically requires a heap allocation, so typed throws should still be preferred for code bases that want to avoid heap allocations. Metatypes Embedded Swift has traditionally allowed metatypes (e.g., Int.self) only in very narrow places, for example when using them to specify argument types for generic functions: rawPointer.bindMemory(to: Value.self, capacity: 1) Swift 6.4 introduces complete support for metatypes in Embedded Swift: one can create and use instances of metatypes, including existential types like any (DefaultInitializable.Type). For example, this is now permitted and works in the same way as full Swift: protocol DefaultInitializable { init() } extension Int: DefaultInitializable { } let factory: any (DefaultInitializable.Type) = Int.self let aValue: any DefaultInitializable = factory.init() Library improvements Additional features in the Swift standard library and associated libraries from full Swift are now available in Embedded Swift. Floating point parsing Swift floating point values can be parsed from a string, like this: let inputText: String = getInputText() if let value = Double(inputText) { // value is a Double } As part of a reimplementation of this functionality in Swift, these floating-point parsing APIs are now available in Embedded Swift as well. Concurrency error handling The Embedded Swift concurrency library now supports throwing operations, such as throwing tasks and task groups. For example: let task = Task { if badThing { throw MyError.badThingHappened } return "ok" } print(try await task.value) Try it out! Embedded Swift support is available in the Swift development snapshots. The best way to get started is through the examples in the Swift Embedded Examples repository, which contains a number of sample projects to get Embedded Swift code building and running on various hardware. If you have questions about the improvements described here, or want to discuss your own Embedded Swift work, we encourage you to join the conversation on the Swift forums. You can ask about this post in the associated thread, and share your experiences in the Embedded Swift category.

• Swift.org · rss:a76d8c3e94eba61fc46b5cc05fc51e5f6df1d0f46d901d32e9c27bb251f155ae:swift-blog:2 · swift.org ↗ · 发布：2026-01-30T15:15:00+00:00 · 采集：2026-09-17T07:07:56.208180+00:00

A Reddit thread earlier this month asked about building web apps with Swift. For this edition of “What’s new in Swift,” we invited a developer to share their experience: Hi, I’m Nick Sloan. I’m the head of engineering at Studioworks, a platform that makes it easy and fun to run your creative studio, agency or freelance business. We chose Swift for Studioworks because of how easy it is to write safe and reliable code with great performance. Our Swift project makes use of Hummingbird 2, Soto (for its incredible DynamoDB Codable support), Hummingbird MacroRouting, and Elementary. Studioworks is a big and growing project. We are the largest Elementary codebase, and I suspect we are among the biggest projects using Hummingbird as well. We’ve already processed millions of dollars in invoices for our customers, and after 20 years of shipping web applications, I have never seen fewer crashes and bugs make it to production. Performance has also been excellent, especially after moving our templates to Elementary. Our heaviest pages make it to the browser in less than 100ms. We’ve been deploying web applications with PHP and Python for decades, and getting started with a Swift web project was certainly a bit slower. We had to recreate some of the build, deployment and chat tooling we had been relying on in our Python projects for years, and it took us a bit of experimenting to realize that Elementary was the best choice for templating. Once we got past those hurdles we’ve been able to build about as fast as we ever did with Python, and the quality is much better. Swift on the web has been a resounding success for us, and I hope we’ll see this part of the community continue to grow! — Nick Sloan Learn more about Swift on Linux and server-based use cases by checking out our October edition, which featured highlights from the Server-Side Swift Conference. The swift.org website also has a use case page dedicated to cloud services, including a tutorial to get started. And now for what’s new in Swift this month. Videos to watch On Progressive Disclosure in Swift - Doug Gregor’s must-watch talk explores how Swift lets you progressively use more language features as your experience and codebase evolves. A new episode of NSScreencast dives into livecoding the Billion Row Challenge, with Matt Massicotte as a guest. New package releases Want to build 3D models with code? Cadova is a programmable alternative to traditional CAD tools, with a focus on 3D printing. Feather Database provides a database-agnostic layer that can be shared by multiple database drivers. And it’s designed for modern Swift concurrency. Miguel de Icaza has ported the .NET Foundation mail stack, originally created by Jeffrey Stedfast (MailKit/MimeKit), to Swift. Hello MailFoundation and MimeFoundation! Community highlights Introduction to Building Swift for Yocto - Check out this embedded Linux guide to using meta-swift to build Swift for a Raspberry Pi Zero 2. Registration is open for several Swift community conferences, including SwiftCraft and try! Swift Tokyo. The Call for Proposals (CFP) is still open for try! Swift Tokyo. Swift Evolution The Swift project adds new language features through the Swift Evolution process. These are some of the proposals currently under review or recently accepted for a future Swift release. Under active review: SE-0506 Advanced Observation Tracking - @Observable types automatically track when their properties change. However, advanced use cases like developing middleware infrastructure or widgeting systems require more control and features. This proposal adds options to the existing withObservationTracking to control when/which changes are observed, and a continuous variant that re-observes automatically after coalesced events. SE-0507 Borrow and Mutate Accessors - When you read or write a Swift property, the code that handles that access currently either makes a copy of the value or uses coroutines, which have performance overhead. This proposal adds new borrow and mutate keywords that let properties provide direct access to their stored values and enables properties that hold values that can’t be copied. Recently accepted: SE-0498 Expose demangle function in Runtime module - The Swift compiler uses name mangling, turning Swift symbols into strings like $sSS7cStringSSSPys4Int8VG_tcfC which show up in backtraces and profiling tools. There are times, however, when a human-readable format is preferable. This proposal introduces a new API that allows calling out to the Swift runtime’s demangler, without leaving the process.

• ChatGPT 官方更新 · web:chatgpt-updates:2026-08-04:64bdce9355437874ee246e9d952df64fbf1f4b9666bb2f2bbdd292b42d84122b · help.openai.com ↗ · 发布：2026-08-04T00:00:00+00:00 · 采集：2026-09-17T13:11:16.322077+00:00

## **Large pastes are now handled as attachments for all plans**

Starting today for **ChatGPT Enterprise and Education** users, long pastes are handled differently in ChatGPT.

If you paste **more than 10k characters** into the composer, ChatGPT will automatically convert the content into an **attachment** instead of inserting it directly into the text field. This keeps the composer cleaner and helps prevent large pastes from consuming your entire context window.

You can still move the content back into the message at any time by clicking on **Show in text field** to convert the attachment back to a direct paste.

This is already available to users on all other plans.

## **New education plugins for students and educators**

Three new plugins for ChatGPT Work and Codex help students and educators work with their own course materials, connected tools, and institution-approved apps.

-   **College Student:** Create study guides, practice quizzes, flashcards, and interactive explanations using the materials you’re already studying.

-   **College Educator:** Develop syllabi, course materials, assessments, and interactive learning resources, including content for a learning management system.

-   **K–12 Educator:** Plan lessons, create classroom resources, adapt materials for different learners, and develop interactive visuals.

The plugins are available through ChatGPT Edu and ChatGPT for Teachers district deployments. Schools and institutions control which tools and permissions are available.

[Learn more about the new education plugins](https://openai.com/index/learn-teach-chatgpt-work-codex/).

• Swift.org · rss:a76d8c3e94eba61fc46b5cc05fc51e5f6df1d0f46d901d32e9c27bb251f155ae:swift-blog:7 · swift.org ↗ · 发布：2026-03-31T14:30:00+00:00 · 采集：2026-09-17T07:07:56.208180+00:00

Welcome to “What’s new in Swift,” a curated digest of releases, videos, and discussions in the Swift project and community. Swift 6.3 has been released, expanding Swift into new domains and improving developer ergonomics. A highlight of its release is work to improve cross-platform build tooling. Owen Voorhees shares an update on that effort: Hi! I’m Owen, a lead engineer on the Core Build team at Apple. Last year we shared our goal to bring Swift Build to Swift Package Manager, in an effort to deduplicate build technologies within the Swift ecosystem and deliver a consistent build experience across all platforms that Swift supports. Since the announcement, we’ve been working in the open, landing hundreds of patches to improve Swift Build’s support across various platforms including Linux and Windows, and to integrate it deeply in Swift Package Manager. With Swift 6.3, developers have the option to enable this integration and try it out with their packages. To validate parity with the previous build system, we’ve used the package list from swiftpackageindex.com, testing thousands of open source packages with Swift Build. Most recently, the main branch of Swift started using Swift Build as its default build system, paving the way for Swift Build to be the out-of-the-box option for Swift developers in a future Swift release. Over the coming months, we’ll continue sharing our progress and driving down the remaining bugs to bring the build system to parity. We encourage you to give it a try and file bugs that you encounter. We’re excited by this progress, and look forward to building future tooling improvements across all platforms and project models that will benefit from this build system. Now on to other news about Swift: Videos to watch Interested in Swift for systems programming? The -ization of Containerization, presented at SCaLE, covers the Containerization project and their experience adopting Swift. Swift community meetup #8 featured two talks: real-time computer vision on NVIDIA Jetson, and a production AI data pipeline built with Vapor. A new interview with Matt Massicotte on the Swift Academy podcast goes in-depth on Swift Concurrency. Community highlights How can you gradually deprecate APIs ahead of a major release? Point-Free blogged Hard Deprecations and Soft Landings with SwiftPM Traits, a clever approach to solving a common API deprecation problem. Daniel Jilg shared TelemetryDeck’s adoption story on the Swift blog, including how they use Swift and Vapor for backend services. The March 2026 Swift for Wasm updates are out, highlighting a new JavaScriptKit release with BridgeJS improvements, and continued work in WasmKit. Swift Evolution The Swift project adds new language features through the Swift Evolution process. These are some of the proposals currently under review or recently accepted for a future Swift release. Under active review: SE-0522 Source-Level Control Over Compiler Warnings - Swift lets you configure warning behavior at the module level using compiler flags, but that’s an all-or-nothing approach. This proposal adds fine-grained warning control: a @warn attribute that lets you override warning behavior for a specific diagnostic group within the scope of a single declaration, with support for escalating it to an error, downgrading it to a warning, or suppressing it entirely, without affecting the rest of your module. Recently accepted: SE-0509 Software Bill of Materials (SBOM) Generation for Swift Package Manager - An SBOM is a standardized inventory of all the software components in a project, increasingly required for security auditing and regulatory compliance. This proposal adds native SBOM generation to SwiftPM as both a --sbom-spec flag on swift build, and a separate swift package generate-sbom subcommand, with support for the CycloneDX and SPDX formats. ST-0021 Targeted Interoperability between Swift Testing and XCTest - When migrating from XCTest to Swift Testing, it’s common to call existing XCTest helper functions from new Swift Testing tests. Today, an XCTAssert failure inside a Swift Testing test is silently ignored. This proposal fixes that: XCTest APIs will work as expected when called in Swift Testing, and Swift Testing APIs will work as expected when called in XCTest, if XCTest already provides similar functionality. SE-0515 Allow reduce to produce noncopyable results - Swift’s reduce currently requires its initial value to be copyable, making it incompatible with noncopyable types. This proposal updates reduce to support noncopyable initial values and results, and changes it to consume rather than borrow the initial value, eliminating an unnecessary copy even for copyable types.

• ChatGPT 官方更新 · web:chatgpt-updates:2026-07-06:aaf0f5b244c607503d432fc195b853c814b2e121155e8b2254b50570d6d36796 · help.openai.com ↗ · 发布：2026-07-06T00:00:00+00:00 · 采集：2026-09-17T13:11:16.322077+00:00

## **GPT-5.5 Instant Mini in ChatGPT**

We’re rolling out GPT-5.5 Instant Mini in ChatGPT. It replaces GPT-5.3 Instant Mini as the fallback model users reach after hitting their GPT-5.5 Instant or Auto rate limits. Because it serves as a fallback, it won’t appear in the model picker. This update does not affect the API or Codex.
Compared with GPT-5.3 Instant Mini, GPT-5.5 Instant Mini better tracks evolving user intent, calibrates tone, and avoids repetitive or overly structured responses. Testing also showed stronger personalization and fewer factual issues.

## ChatGPT for PowerPoint is now generally available

ChatGPT for PowerPoint is now generally available globally. From a sidebar in Microsoft PowerPoint, you can create new slides and decks, revise existing presentations, ask questions about a deck’s story and structure, and use Skills and apps to work with repeatable workflows and connected sources. Slides remain editable in PowerPoint, although advanced edits and template matching may still require review.
The add-in is available to Free, Go, Plus, Pro, Business, Enterprise, ChatGPT Edu, and K-12 users, subject to plan limits and workspace settings. Install it from Microsoft Marketplace, sign in with your eligible ChatGPT account, and review important claims, numbers, and slide changes before sharing. [Learn more about ChatGPT for PowerPoint](https://help.openai.com/articles/20001242).

• ChatGPT 官方更新 · web:chatgpt-updates:2026-08-06:3e35db61acd8d685129c65308d2c2b5f59b2db57c62464654916ee99108121cf · help.openai.com ↗ · 发布：2026-08-06T00:00:00+00:00 · 采集：2026-09-17T13:11:16.322077+00:00

## **Updates to GPT‑5.6 in ChatGPT**

Plus and Pro users can now use an updated GPT‑5.6 Sol in ChatGPT with more reliable facts, more focused answers, and a new slider to choose how much thought ChatGPT puts into a response.

GPT‑5.6 Luna will become the default model for Free and Go users this week. Starting next week, they’ll also have unlimited text chats and access to a new Think button for harder questions (subject to abuse guardrails). Limits will still apply for file uploads, images and other tools.

This release applies to the chat experience in ChatGPT. Work and Codex are not changing as part of this release.

• Swift.org · rss:a76d8c3e94eba61fc46b5cc05fc51e5f6df1d0f46d901d32e9c27bb251f155ae:swift-blog:15 · swift.org ↗ · 发布：2026-07-02T16:00:00+00:00 · 采集：2026-09-17T07:07:56.208180+00:00

Welcome to “What’s new in Swift,” a curated digest of releases, videos, and discussions in the Swift project and community. June was an exciting month for Swift, featuring announcements at WWDC and community events around the globe. We invited the organizers of one of them to share with us: Hey, it’s Mikaela and Adrian. We are organizers of CommunityKit, a community-organized conference that takes place the same week as WWDC, and iOSDevHappyHour, a monthly online meetup that keeps the community connected year-round. This is our fifth year coming out to Cupertino, and we love being able to create a place for the community to thrive, no matter where developers live. CommunityKit brought together over 250 developers in real life to geek out over the announcements, stay for the community and vibes, see what everyone is creating, and learn from each other. Some of the highlights from this year’s event were the Indie Fair, where developers showcased their apps; the Watch Party, our annual gathering to watch the keynotes together; and Make Something, Ship Nothing, a hands-on postcard-making hangout to close the week. This year we also introduced workshops, including “Inclusive by Design” by Danielle Lewis, and for the Swift community: “Write Faster, Smarter Swift” by Paul Hudson. We can’t wait to hear about what everyone builds and brings to next year’s Indie Fair, and hope to see you at CommunityKit and iOSDevHappyHour! Now on to other news about Swift: WWDC26 highlights At its WWDC26 conference, Apple provided an update on its adoption of Swift and made a variety of new Swift-related announcements. Some highlights: During the Platforms State of the Union, Apple announced that parts of the core operating system kernel are being written in Swift for upcoming releases. What’s new in Swift featured changes in Swift since last year, including a preview of what’s coming in Swift 6.4, like up to 4x faster URL parsing and support for async code in defer blocks. The QUIC transport layer in Apple’s networking stack was rewritten in Swift. The project has been open sourced and is available for cross-platform use through SwiftNIO integration. A new Swift package, Foundation Models framework utilities, was released with tools for working with LLMs, including custom skills and context management helpers. It runs on Apple platforms and select Linux distributions. The Foundation Models framework itself will be open sourced in the future, meaning the same Swift APIs you use in your app could run on your server. Container Machine is a new tool that provides a lightweight, persistent Linux environment on a Mac. Unlike a container, which is modeled after an application, a container machine is modeled after the environment itself. Container machines share the host environment, including the home directory and configuration. It’s written in Swift and open source. Videos to watch Build real-time apps and services with gRPC and Swift walks through integrating an iOS app and gRPC service using live race data from a go-karting league. See if you can spot where the track is located. 👀 Want to learn about Swift macros with hands-on tutorials? Stewart Lynch published two videos with sample code to follow along: Swift Macros Demystified: Build a Freestanding Expression Macro, and Swift Attached Macros: Build a Real-World Member Macro from Scratch. A new 10-minute Embedded Swift demo uses an accelerometer and the XIAO ESP32-C6 to control a Swift bird that glides across a mini OLED screen. No soldering required! Community highlights Swift Package Index joined Apple and remains open source. The team says they’re working together to build a comprehensive package registry for the Swift community. Yeo Kheng Meng blogged about bringing Swift to the Apple II, complete with a REPL, compiler, file browser, and editor. It’s a subset of Swift and was built with AI assistance. Apple shared an adoption story on the Swift blog: Migrating the TrueType Hinting Interpreter, covering how the TrueType hinting interpreter in macOS and iOS was rewritten in Swift from C. It runs 13% faster on average. The Swift Ecosystem Steering Group announced the creation of the Networking workgroup. This group will work on a unified networking stack for Swift, layered from low-level I/O primitives, through common protocols, to a modern HTTP client and server API. New package releases New Swift bindings for the OkHttp Java library were released. If you’re using Swift on Android and looking for an HTTP client this may be useful. The project was generated with swift-java. Kiln is a new documentation engine written in Swift. Built to replace MkDocs-based documentation sites, it gives more options for the Swift community to render docs, in addition to the DocC project which is used for the official Swift documentation. You can see Kiln in action at the Vapor documentation. Version 0.4.0 of Elementary UI was released, a frontend framework for running Swift applications natively in the browser. Swift Evolution The Swift project adds new language features through the Swift Evolution process. These are some of the proposals currently under review or recently accepted for a future Swift release. Under active review: SE-0526 withDeadline - Asynchronous operations in Swift can run indefinitely, and implementing time limits manually using task groups and clock sleep operations is verbose and error-prone. This proposal adds withDeadline, a function that executes an async operation with a composable absolute time limit specified as a clock instant, canceling the operation if it hasn’t completed in time. It also allows multiple nested operations to share the same deadline, avoiding the drift that accumulates when relative durations are passed through call layers. Recently accepted: SE-0474 Yielding Accessors - When you call a mutating method on a computed property, Swift creates the illusion of in-place mutation by getting a copy, mutating it, then setting it back. This causes unnecessary copy-on-write buffer duplication for types like String, and is impossible for noncopyable types, which can’t be copied out at all. This proposal adds yielding borrow and yielding mutate, two new ways to implement computed properties and subscripts that instead lend the caller direct access to the underlying value without copying it. Recently accepted with modifications: SE-0529 Add FilePath to the Standard Library - FilePath in the swift-system package parses platform-specific path syntax on the developer’s behalf, provides a normalized view of path components, and enables filesystem resolution. However, shipping in an external package means the standard library, Swift runtime, and toolchain libraries such as Foundation cannot depend on it. This proposal adds FilePath and its associated types to the Swift module, alongside essential functionality for construction, decomposition, resolution, and C interoperability. SE-0527 UniqueArray - Swift’s Array can’t store noncopyable elements without compromising its copy-on-write semantics or performance predictability. This proposal adds two new types to a new Containers module: RigidArray, a fixed-capacity array that traps on overflow, and UniqueArray, a dynamically growing array that enforces unique ownership by being noncopyable itself.

• Swift.org · rss:a76d8c3e94eba61fc46b5cc05fc51e5f6df1d0f46d901d32e9c27bb251f155ae:swift-blog:10 · swift.org ↗ · 发布：2026-06-01T12:00:00+00:00 · 采集：2026-09-17T07:07:56.208180+00:00

Goodnotes has been helping millions of users take handwritten notes on iPad for over a decade, earning recognition as Apple’s iPad App of the Year in 2022. Today, the same Swift code that powers our iOS app also runs seamlessly in web browsers through WebAssembly, delivering the exact same ink rendering and note-taking experience users love. This journey demonstrates that Swift excels as a cross-platform language, running high-performance applications while sharing the same codebase. Every bug fix and improvement to Goodnotes benefits all our users simultaneously, regardless of which platform they use. After two years of development and over two years in production at Goodnotes, we’ve shown that Swift on WebAssembly is a viable, powerful approach for building complex, performance-critical web applications. Why we chose Swift for the web When we decided to bring Goodnotes to the web in 2021, we faced a critical decision. After more than 10 years of development, we had accumulated millions of lines of Swift code that implemented countless refinements and optimizations for digital ink rendering, document synchronization, conflict resolution using Conflict-Free Replicated Data Types (CRDTs), and content search and document indexing. We need to maintain more than 60 Frames Per Second (FPS) for real-time ink rendering, which makes performance critical. A JavaScript rewrite, Flutter, or Kotlin Multiplatform would all require rewriting our entire rendering engine from scratch, a substantial undertaking that would have delayed our web launch by years and inevitably introduced behavioral differences between platforms. SwiftWasm emerged as the solution. This community-driven project allows Swift code to compile to WebAssembly, running in browsers with good performance. We started experimenting with SwiftWasm, building prototypes to validate the approach. Our first experiment focused on our handwriting component, a performance-critical part of Goodnotes that would serve as a good indicator of WebAssembly’s capabilities. The results were promising enough that we committed to this path. The most compelling benefit wasn’t just code reuse, but the guarantee of behavioral consistency. When users draw a stroke on their iPad and later open the same document on the web, they see exactly the same curves, the same pressure sensitivity, the same ink flow. This isn’t because we carefully reimplemented the same algorithms twice: it’s because it’s literally the same Swift code running on both platforms. Technical architecture Goodnotes Architecture: Shared Swift code between iOS and Web platforms. Our architecture is built around a clear separation between platform-specific UI components and shared business logic. This design enables us to maintain behavioral consistency while leveraging platform-native capabilities where appropriate. Shared core components The heart of our application consists of three main parts: Content Rendering Engine: This handles the real-time rendering of notebook content and interactive ink strokes. We use a custom rendering engine built on low-level graphics APIs: Metal on iOS and WebGL on the web. The rendering logic is almost entirely shared, with only platform abstraction layers implemented separately for each platform. Business Logic Layer: Document modeling, handwriting recognition, and document indexing are all implemented in shared Swift packages. View Models: Core view models that handle tool interactions and user gestures are shared across platforms. Code sharing metrics Our codebase demonstrates significant code reuse: Total Web Swift codebase: 2.2 million lines of code Shared Swift code: 1.47 million lines (66% of the web app, 34% of the iOS app) While lines of code isn’t the best metric, these numbers reflect the substantial business logic and rendering engine that we successfully share between platforms. Binary size and loading The final WebAssembly binary is approximately 50 MB, which compresses to 12 MB with Brotli compression. We use Service Workers for efficient caching and fast load times for users. JavaScript interoperability We use JavaScriptKit for seamless interoperability between Swift and JavaScript. This allows us to integrate with the existing web ecosystem while keeping our core logic in Swift. Platform compatibility considerations When sharing code between iOS and WebAssembly targets, we encountered several important considerations: Concurrency Model: libdispatch APIs are unavailable on WebAssembly targets. We migrated from direct libdispatch usage to Swift Concurrency’s async/await and actors, for better cross-platform compatibility. Architecture Differences: On wasm32, Swift’s Int has a 32-bit width. Some code assumed Int only held 64-bit values, so it had to be updated to use Int64 explicitly. Dependency Injection: Network access and other I/O operations are abstracted through dependency injection, allowing us to provide platform-specific implementations while keeping the core logic shared. Multithreading with WASI threads One of the most significant technical achievements was implementing true parallelism using WebAssembly System Interface (WASI) Threads with Web Workers and SharedArrayBuffer. This allows us to: Run handwriting recognition in background Web Workers Perform document indexing without blocking the main thread Maintain smooth rendering at more than 60 FPS while processing complex operations Swift Concurrency’s Custom Actor Executors (SE-0392) were crucial for managing the web platform’s constraints. JavaScript objects are isolated to their originating thread, so we needed precise control over where our Swift actors execute. JavaScriptKit provides several APIs to create a SerialExecutor for a dedicated Web Worker, enabling us to pin specific actors to specific Web Workers. This architecture ensures that computationally-intensive tasks like handwriting recognition run in the background while UI operations stay on the main thread, while still allowing access to JavaScript objects inside background threads. Performance Impact: This multithreading approach delivered a greater than 2x improvement in Interaction to Next Paint (INP), significantly enhancing the UI responsiveness during complex operations. Security Considerations: Modern browser security policies require Cross-Origin Isolation to use SharedArrayBuffer. While this adds some complexity to the application, it’s a necessary trade-off for the performance benefits of true parallelism. For applications that can’t meet these requirements, single-threaded cooperative concurrent execution remains a viable option. Development experience One of the most significant aspects of our Swift on WebAssembly experience was the development workflow. The tooling ecosystem is mature and powerful, providing a solid development experience. IDE support We can develop using either Xcode or VS Code with SourceKit-LSP, providing full language server support including autocomplete, error checking, and refactoring capabilities. Xcode doesn’t currently have WebAssembly platform support, so code completion and other features are limited for WebAssembly-specific APIs. SourceKit-LSP, however, has Swift SDK support, so by properly configuring .sourcekit-lsp/config.json, you can get code completion for WebAssembly targets as well. Debugging and development tools You can debug Swift code directly in Chrome DevTools: set breakpoints, inspect variables, and step through your Swift code as naturally as JavaScript. We developed a Chrome DevTools extension library that enables Swift-specific variable reflection and source-level debugging, building upon the existing WebAssembly debugging capabilities. For more details on the enhanced DWARF extension for Swift, see the Swift on WebAssembly debugging guide. Debugging Swift code in Chrome DevTools with full source code visibility and variable inspection. Performance profiling The existing web ecosystem provides powerful performance profiling tools. Chrome’s Performance tab shows exactly where time is spent, down to individual Swift functions, and the Memory tab gives us good insight into memory usage patterns. For most performance optimization tasks, these standard tools are quite effective. For more advanced cases requiring specialized memory profiling capabilities and detailed heap analysis, the growing WebAssembly ecosystem provided the foundation for building custom tools. We developed wasm-memprof for detailed heap profiling when optimizing memory usage. This tool provides insights into memory allocation patterns that aren’t easily visible through standard web profiling tools. Performance profiling with wasm-memprof showing memory allocation patterns and optimization opportunities. Contributing back to the community As part of our journey, we’ve been able to contribute back to the Swift community in meaningful ways. All WebAssembly-related changes have been upstreamed, and the WebAssembly platform has been supported since Swift 6.2! This means that other teams can now benefit from the same tooling and language features that made our project successful. Lessons learned Our experience has shown that Swift on WebAssembly is production-ready for complex applications. The language’s safety features, performance characteristics, and modern concurrency model translate well to the web platform. For teams considering this path, here are our key recommendations: Start with a performance-critical component to validate the approach. Invest in proper platform abstraction layers early. Leverage Swift Concurrency for cross-platform compatibility. Plan for the security requirements of SharedArrayBuffer if multithreading is needed. Consider gradual adoption rather than complete rewrites for existing projects. Consider using Swift for your web projects. The growing WebAssembly ecosystem and improved tooling support make this an increasingly viable option for teams looking to share code across platforms. Get involved with Swift on WebAssembly Swift has fulfilled its promise as a powerful, expressive language that works everywhere. From mobile devices to servers to web browsers, Swift code can run efficiently while maintaining the safety and performance characteristics that developers appreciate. The Swift on WebAssembly ecosystem is more accessible than ever. Here’s how you can get involved: For Developers: Try out the Swift on WebAssembly getting started guide and build your first web application. Explore the JavaScriptKit Hello World tutorial for hands-on learning. Use the WebAssembly debugging guide to set up your development environment. For SwiftWasm Contributors: Participate in discussions with the #webassembly tag on the Swift forums. Open issues and contribute to improve the toolchain. Share your experiences and help build the community. With Swift’s official WebAssembly support, we’ve entered a new era of cross-platform development. The same language that powers iOS applications can now create web experiences that are fast, safe, and maintainable. Swift’s future is increasingly cross-platform, and we’re excited to see what the community builds next.

• Swift.org · rss:a76d8c3e94eba61fc46b5cc05fc51e5f6df1d0f46d901d32e9c27bb251f155ae:swift-blog:16 · swift.org ↗ · 发布：2026-08-07T15:30:00+00:00 · 采集：2026-09-17T07:07:56.208180+00:00

Welcome to “What’s new in Swift,” a curated digest of releases, videos, and discussions in the Swift project and community. Swift is built by contributors working on far more than the language and compiler. Alexander Sandberg reflects on what that kind of contribution looks like, and why there’s a place for you too: Hi, I’m Alex. Four years ago I joined the Swift Website Workgroup to help care for the website you’re reading this on. I was still fairly new to the Swift community then, but I wanted to get involved somehow and make a difference. Contributing to Swift sounds like it should involve working on the compiler or writing and reviewing language evolution proposals, but there’s a lot going on around the language itself too. What I actually spent my time doing was helping improve the site’s content and navigation, taking part in the big redesign, writing release blog posts, reviewing pull requests, and answering issues and forum posts. All of it is work anyone can learn to do, and all of it ends up in front of everyone who comes to Swift.org—which made even the smallest fixes feel worth doing. The Website Workgroup happened to be my way in, but there are groups like it across the project—server, testing, Windows, Android, and more—each a handful of people looking after their part of the ecosystem. Whatever you’re good at or curious about, there’s probably a corner of Swift that could use your help. The community page is a good place to start looking, and the people you’ll meet are some of the friendliest I’ve come across in open source. After four years, I’ve stepped away to focus on other things in life. Thank you to everyone I got to work with over these years! I’m leaving the workgroup, but not the community, and I’ll still be around helping push Swift forward where I can. For more information about workgroups, the How we work page lists all workgroups. And on the Swift forums you’ll often see workgroups sharing open meetings if you’re looking to get involved! Now on to other news about Swift: Videos to watch Balancing High and Low-Level Programming in Swift, presented by John McCall at PLDI 2026, digs into the ongoing work to bring explicit ownership and lifetime features to Swift for systems programming, and the tricky balance of fitting them into a language built around simple, low-friction application development without breaking source compatibility or ABI stability. Speaking of low-level Swift: I Built a Timer for the Playdate in Swift and I Built a 3D Renderer for the Playdate (in Swift) are a fun pair of videos on building for the handheld game console in Swift. Melbourne CocoaHeads has been uploading a great back catalog of past event talks to their YouTube channel, including Christian Mitteldorf’s 2023 talk on running Swift on the server. Recent project changes Google Summer of Code contributions are underway: Ege Kaya is building a Task Registry for the Swift Concurrency runtime (PR #90547), Filip Sakellariou is contributing to swift-syntax (PR #3378, PR #3390), and Padmashree S S is working in sourcekit-lsp (PR #2698, PR #2729). Task Stealers, a new primitive in Swift Concurrency’s runtime, recently merged after generating the liveliest discussion of any PR in July. They change how work gets distributed across the cooperative thread pool, an under-the-hood change worth watching if you write concurrency-heavy code. SwiftPM landed a reference implementation of the package registry service, along with basic auth support, a solid starting point if you’re exploring self-hosting a registry. Small but welcome fixes landed across tooling: the VS Code extension now fails loudly when a Swift installation is broken, and gained a clean build folder command, while Swift Testing’s event stream gained an ABI.Context API that produces human-readable output for CI tooling and dashboards. New package releases Need to tokenize text fast? swift-gigatoken is a zero-copy tokenizer hitting up to 670 MB/s, inspired by Marcel Roed’s gigatoken, with NEON/CRC32 support and compatibility across WebAssembly and Embedded Swift. any-error-swift takes a different approach to error handling, offering a concrete, value-semantic, Hashable type that erases any Swift error, no Foundation required, and works on every platform Swift runs on. Swift Evolution The Swift project adds new features through the Swift Evolution process. These are some of the proposals recently accepted for a future Swift release. Recently accepted: ST-0025 Tag-based Test Execution Filtering - Swift Testing lets you annotate tests with named tags, and separately lets you include or exclude tests by regex using the --filter and --skip options, but there’s no way to combine the two and filter by tag. This proposal adds a tag: prefix to those options, so you can run commands like swift test --skip tag:uiTest to selectively run or skip tests based on their tags rather than relying on a consistent naming scheme. SE-0513 API to get the path to the current executable - There’s currently no portable, reliable way to get the path to the currently running executable in Swift. This proposal adds CommandLine.executablePath, a new property in the standard library that provides a consistent way to get this value across all platforms Swift supports. Recently accepted with modifications: SE-0526 withDeadline - Asynchronous operations in Swift can run indefinitely, and implementing time limits manually using task groups and clock sleep operations is verbose and error-prone. This proposal adds withDeadline, a function that executes an async operation with a composable absolute time limit specified as a clock instant, canceling the operation if it hasn’t completed in time. It also allows multiple nested operations to share the same deadline, avoiding the drift that accumulates when relative durations are passed through call layers.

• Swift.org · rss:a76d8c3e94eba61fc46b5cc05fc51e5f6df1d0f46d901d32e9c27bb251f155ae:swift-blog:14 · swift.org ↗ · 发布：2026-06-23T14:00:00+00:00 · 采集：2026-09-17T07:07:56.208180+00:00

Since its launch, Swift Package Index has become an essential part of the Swift ecosystem, helping developers discover quality packages, evaluate compatibility, and make informed decisions about dependencies. Started by Dave Verwer and Sven A. Schmidt, the project currently indexes thousands of packages and serves as a central hub for package documentation. Recently, Swift Package Index has grown alongside the Swift project’s multi-platform expansion, extending its compatibility testing to encompass Linux, Android, and WebAssembly. Today, Swift Package Index announced that it has joined Apple and continues its mission: Together, we’re building a comprehensive package registry to serve the Swift community’s evolving needs. Learn more by reading the Swift Package Index announcement.

• Swift.org · rss:a76d8c3e94eba61fc46b5cc05fc51e5f6df1d0f46d901d32e9c27bb251f155ae:swift-blog:1 · swift.org ↗ · 发布：2026-01-26T12:00:00+00:00 · 采集：2026-09-17T07:07:56.208180+00:00

We are excited to announce the creation of the Windows workgroup! The primary goal is to ensure ongoing support for Swift on Windows, enabling users to develop Windows applications using the Swift programming language and its associated tools. The new Windows workgroup joins a growing list of Swift workgroups, including the Android workgroup, Build and Packaging workgroup, and Testing workgroup which were all added in the past year. Swift workgroups are community-led efforts, formally recognized by the project, to help move forward key areas. Windows has been an officially supported platform since 2020, and there’s been great progress since then. The VS Code extension supports Windows, and we’ve seen the community take advantage of Swift’s features on Windows, some of which are described in a 2023 blog post about Using Interoperability to Build on Windows. The workgroup will build upon these efforts and focus on work to: Improve and maintain Windows support for the official Swift distribution. Recommend enhancements to core Swift packages such as Foundation and Dispatch to work better with Windows idioms. Make recommendations to the Swift Project about the direction of future Windows support in Swift. Identify and recommend best practices for bridging between Swift and the Windows API, and for shipping Swift libraries with Windows applications. To learn more and get involved: Discuss this announcement on the forums, and share ideas in the Windows category. Learn more about the Windows workgroup by reading its charter. Join the Windows workgroup’s meetings every other Wednesday at 9:00 AM PT (Pacific Time Zone). Workgroup membership and meetings are open to those who wish to participate, and contributors are welcome! To receive an invite, send a message to @windows-workgroup on the Swift forums.

• Swift.org · rss:a76d8c3e94eba61fc46b5cc05fc51e5f6df1d0f46d901d32e9c27bb251f155ae:swift-blog:11 · swift.org ↗ · 发布：2026-06-03T21:45:00+00:00 · 采集：2026-09-17T07:07:56.208180+00:00

Welcome to “What’s new in Swift,” a curated digest of releases, videos, and discussions in the Swift project and community. To start, we’re focusing on some of the many local meetup groups sharing Swift content: Around the world, local meetup groups and conferences bring Swift developers together, and some even predate Swift itself! Many have YouTube channels where they share videos from their events, so you can tune in remotely. Let’s highlight a few. In May the SF Swift meetup hosted a talk by Dan Federman, Agentify Your Swift Repo, which covers building an agent for CI and review feedback. Swift Barcelona has a dedicated YouTube presence, while some groups, like the new MLX India meetup, post event playlists, including a recent talk about using MLX Swift in iOS apps that’s worth checking out. And there’s even a community-organized Swift Community Meetups YouTube channel, which hosts meetups online and is home to a series of cross-platform Swift talks. These groups support developers in a few ways worth calling out: Local community: they’re building local connections among Swift developers and organizations. Room for experimentation: by being less formal and sometimes featuring works in progress, meetups create space to share ideas, get feedback, and learn from one another. I encourage more folks to get involved in their local Swift meetup. And if your group starts publishing content and videos, please share on the Swift Forums in the Community Showcase category. — Dave Lester Now on to other news about Swift: Videos to watch Interested in using Swift for backend server development? Mohammad Azam posted a livestream recording of Introduction to Hummingbird, a walkthrough of the web framework covering installation and development basics. Meet the Temporal Swift SDK, from Replay 2026, introduces the SDK that brings Temporal’s durable workflows (long-running processes that survive crashes, retries, and restarts without losing state) to Swift. The SDK recently reached its 1.0.0 release. Sébastien Stormacq shares a great introduction to what’s possible with AWS Lambda and Swift, presenting Swift, Server-side, & Serverless on the DevStandup YouTube channel. Community highlights Swift and WebAssembly continues to be an exciting part of the project, with regular activity shared on the Swift forums, including the most recent Swift for Wasm May 2026 Updates. And if you missed it, check out the new blog post by Goodnotes about how they brought Goodnotes to the web with Swift and WebAssembly. The 2026 Swift Mentorship Program was announced, and there’s still time until June 15, 2026, to complete the interest survey to be a mentee. Working on contributions is a rewarding way to learn. And on that note, three Swift projects were accepted for GSoC 2026. Training an LLM in Swift, Part 1 - Matt Gallagher takes handwritten matrix multiplication for a Swift port of llm.c from 2.8 Gflop/s to 1.1 Tflop/s, a 382x speedup, using Swift 6.2’s MutableSpan and InlineArray, Relaxed.multiplyAdd from Swift Numerics, DispatchQueue.concurrentPerform, and finally AMX and Metal. A thorough tour of performance optimization in modern Swift. Members of the community using and contributing to the VS Code Swift Extension met in May during the Swift Extension Community Office Hours, featuring demonstrations of using Kiro and more. New package releases Want to build an AI agent in Swift? The Swift Bedrock Library gives you a Swift library for Amazon Bedrock foundation models. SwiftOSC is an Open Sound Control (OSC) toolkit written in Swift which recently became cross-platform, now supporting Apple platforms, Linux, and Android. It’s music to our ears! Steffan Andrews, who developed SwiftOSC, has also released other music-related Swift packages including SwiftMIDI and SwiftTimecode. Ordo One shipped version 1.33.0 of Benchmark, a package for creating sophisticated Swift performance benchmarks for a wide variety of metrics. Swift Evolution The Swift project adds new language features through the Swift Evolution process. These are some of the proposals currently under review or recently accepted for a future Swift release. Under active review: SE-0532 Optional noncopyable improvements and generalizations - Swift’s Optional can wrap noncopyable types, but unwrapping with if let consumes the optional, leaving it unusable afterward. This proposal adds borrow() and mutate() to Optional, returning Ref<Wrapped>? and MutableRef<Wrapped>? to inspect or modify the payload without consuming it, and generalizes map, flatMap, and unsafelyUnwrapped to support noncopyable wrapped types. Recently accepted: SE-0528 Continuation Safe and Performant Async Continuations - When bridging callback-based APIs to Swift’s structured concurrency, developers today must choose between UnsafeContinuation, where misuse is silent and undefined, and CheckedContinuation, which catches mistakes at the cost of allocation and atomic operations. This proposal adds Continuation<Success, Failure>, a ~Copyable type that makes double-resume a compile-time error and a missing resume a runtime trap, with no overhead on the fast path. SE-0519 Ref and MutableRef types for safe, first-class references - Storing a reference to part of a data structure in Swift today requires either a class, which adds heap allocation and reference-counting overhead, or UnsafePointer, which is unsafe and requires extreme care to use properly. This proposal adds Ref<T> and MutableRef<T> to the standard library: safe types that hold shared and exclusive references to a value, usable as local variables, struct members, and generic type parameters. ST-0024 Test case repetition - Swift Testing can repeat tests for a set number of iterations, or until a specific failure/success condition is reached, which is useful for tracking down flaky failures. Currently, if any single test case triggers the condition, all tests in the target are re-run, even ones that passed. This proposal changes repetition to apply per test case and adds --maximum-repetitions and --repeat-until flags to swift test.

• Swift.org · rss:a76d8c3e94eba61fc46b5cc05fc51e5f6df1d0f46d901d32e9c27bb251f155ae:swift-blog:6 · swift.org ↗ · 发布：2026-03-24T10:00:00+00:00 · 采集：2026-09-17T07:07:56.208180+00:00

Swift is designed to be the language you reach for at every layer of the software stack. Whether you’re building embedded firmware, internet-scale services, or full-featured mobile apps, Swift delivers strong safety guarantees, performance control when you need it, and expressive language features and APIs. Swift 6.3 makes these benefits more accessible across the stack. This release expands Swift into new domains and improves developer ergonomics across the board, featuring: More flexible C interoperability Improvements to cross-platform build tooling Improvements for using Swift in embedded environments An official Swift SDK for Android Read on for an overview of the changes and next steps to get started. Language and Standard Library C interoperability Swift 6.3 introduces the @c attribute, which lets you expose Swift functions and enums to C code in your project. Annotating a function or enum with @c prompts Swift to include a corresponding declaration in the generated C header that you can include in your C/C++ files: @c func callFromC() { ... } // Generated C header void callFromC(void); You can provide a custom name to use for the generated C declaration: @c(MyLibrary_callFromC) func callFromC() { ... } // Generated C header void MyLibrary_callFromC(void); @c also works together with @implementation. This lets you provide a Swift implementation for a function declared in a C header: // C header void callFromC(void); // Implementation written in Swift @c @implementation func callFromC() { ... } When using @c together with @implementation, Swift will validate that the Swift function matches a pre-existing declaration in a C header, rather than including a C declaration in the generated header. Module name selectors Swift 6.3 introduces module selectors to specify which imported module Swift should look in for an API used in your code. If you import more than one module that provides API with the same name, module selectors let you disambiguate which API to use: import ModuleA import ModuleB let x = ModuleA::getValue() // Call 'getValue' from ModuleA let y = ModuleB::getValue() // Call 'getValue' from ModuleB Swift 6.3 also enables using the Swift module name to access concurrency and String processing library APIs: let task = Swift::Task { // async work } Performance control for library APIs Swift 6.3 introduces new attributes that give library authors finer-grained control over compiler optimizations for clients of their APIs: Function specialization: Provide pre-specialized implementations of a generic API for common concrete types using @specialize. Inlining: Guarantee inlining — a compiler optimization that expands the body of a function at the call-site — for direct calls to a function with @inline(always). Use this attribute only when you’ve determined that the benefits of inlining outweigh any increase in code size. Function implementation visibility: Expose the implementation of a function in an ABI-stable library to clients with @export(implementation). This allows the function to participate in more compiler optimizations. For a full list of language evolution proposals in Swift 6.3, see the Swift Evolution dashboard. Package and Build Improvements Swift 6.3 includes a preview of Swift Build integrated into Swift Package Manager. This preview brings a unified build engine across all supported platforms for a more consistent cross-platform development experience. To learn more, check out Preview the Swift Build System Integration. We encourage you to try it in your own packages and report any issues you encounter. Swift 6.3 also brings the following Swift Package Manager improvements: Prebuilt Swift Syntax for shared macro libraries: Factor out shared macro implementation code into a library with support for swift-syntax prebuilt binaries in libraries that are only used by macros. Flexible inherited documentation: Control whether inherited documentation is included in command plugins that generate symbol graphs. Discoverable package traits: Discover the traits supported by a package using the new swift package show-traits command. For more information on changes to Swift Package Manager, see the SwiftPM 6.3 Release Notes. Core Library Updates Swift Testing Swift Testing has a number of improvements, including warning issues, test cancellation, and image attachments. Warning issues: Specify the severity of a test issue using the new severity parameter to Issue.record. You can record an issue as a warning using Issue.record("Something suspicious happened", severity: .warning). This is reflected in the test’s results, but doesn’t mark the test as a failure. Test cancellation: Cancel a test (and its task hierarchy) after it starts using try Test.cancel(). This is helpful for skipping individual arguments of a parameterized test, or responding to conditions during a test that indicate it shouldn’t proceed. Image attachments: Attach common image types during a test on Apple and Windows platforms. This is exposed via several new cross-import overlay modules with UI frameworks like UIKit. The list of Swift Testing evolution proposals included in Swift 6.3 are ST-0012, ST-0013, ST-0014, ST-0015, ST-0016, ST-0017, and ST-0020. DocC Swift 6.3 adds three new experimental capabilities to DocC: Markdown output: Generate Markdown versions of your documentation pages alongside the standard rendered JSON covering symbols, articles, and tutorials. Try it out by passing --enable-experimental-markdown-output to docc convert. Per-page static HTML content: Embed a lightweight HTML summary of each page — including title, description, availability, declarations, and discussion — directly into the index.html file within a <noscript> tag. This improves discoverability by search engines and accessibility for screen readers without requiring JavaScript. Try it out by passing --transform-for-static-hosting --experimental-transform-for-static-hosting-with-content to docc convert. Code block annotations: Unlock new formatting annotations for code blocks, including nocopy for disabling copy-to-clipboard, highlight to highlight specific lines by number, showLineNumbers to display line numbers, and wrap to wrap long lines by column width. Specify these options in a comma-separated list after the language name on the opening fence line: ```swift, nocopy let config = loadDefaultConfig() ``` ```swift, highlight=[1, 3] let name = "World" // highlighted let greeting = "Hello" print("\(greeting), \(name)!") // highlighted ``` ```swift, showLineNumbers, wrap=80 func example() { /* ... */ } ``` DocC validates line indices and warns about unrecognized options. Try out the new code block annotations with --enable-experimental-code-block-annotations. Platforms and Environments Embedded Swift Embedded Swift has a wide range of improvements in Swift 6.3, from enhanced C interoperability and better debugging support to meaningful steps toward a complete linkage model. For a detailed look at what’s new in embedded Swift, see Embedded Swift Improvements coming in Swift 6.3. Android Swift 6.3 includes the first official release of the Swift SDK for Android. With this SDK, you can start developing native Android programs in Swift, update your Swift packages to support building for Android, and use Swift Java and Swift Java JNI Core to integrate Swift code into existing Android applications written in Kotlin/Java. This is a significant milestone that opens new opportunities for cross-platform development in Swift. To learn more and try out Swift for Android development in your own projects, see Getting Started with the Swift SDK for Android. Thank You Swift 6.3 reflects the contributions of many people across the Swift community — through code, proposals, forum discussions, and feedback from real-world experience. A special thank you to the Android Workgroup, whose months of effort — building on many years of grassroots community work — brought the Swift SDK for Android from nightly previews to an official release in Swift 6.3. If you’d like to get involved in what comes next, the Swift Forums are a great place to start. Next Steps Try out Swift 6.3 today! You can find instructions for installing a Swift 6.3 toolchain on the Install Swift page.

• ChatGPT 官方更新 · web:chatgpt-updates:2026-06-25:24ad8931cbfcede845a91518a4a80ed1d1f0f07ea9156cd63ea7357f392c14fd · help.openai.com ↗ · 发布：2026-06-25T00:00:00+00:00 · 采集：2026-09-17T13:11:16.322077+00:00

## **Codex Remote GA and DigitalOcean plugin**

Codex Remote is now generally available on all ChatGPT plans. From the ChatGPT mobile app, users can start or continue work on a connected Mac or Windows host, review progress, and approve actions from their phone. Remote Control now uses authenticated one-to-one QR pairing between each supported mobile device and each host. Connections used since June 8 remain paired; older inactive connections need to pair again. Signing out turns off Remote Control without removing existing pairings.
The new [DigitalOcean Droplet Workspace plugin](https://chatgpt.com/plugins/share/5dc672c7116c44ff92595d48e72df522) lets Codex provision a DigitalOcean Droplet, configure SSH access, and connect it to the Codex app as a remote workspace. Users should update the ChatGPT mobile app and Codex app to the latest versions before connecting. [Learn more about using Codex with a ChatGPT plan](https://help.openai.com/articles/11369540).

• Swift.org · rss:a76d8c3e94eba61fc46b5cc05fc51e5f6df1d0f46d901d32e9c27bb251f155ae:swift-blog:12 · swift.org ↗ · 发布：2026-06-04T10:00:00+00:00 · 采集：2026-09-17T07:07:56.208180+00:00

The Swift Ecosystem Steering Group is excited to announce the creation of the Networking workgroup! Workgroups are community-led efforts, formally recognized by the project, to advance key areas of Swift. The primary goal is to guide the evolution of networking libraries, protocols, and APIs in the Swift ecosystem, making networking in Swift excellent everywhere: high-level and safe by default, modular and interoperable, cross-platform, and observable. Networking is one of the most common entry points for Swift developers, and the ecosystem has matured significantly over the years. Foundational libraries like SwiftNIO, AsyncHTTPClient, and swift-http-types, alongside platform stacks such as URLSession and Network.framework, power networking across apps, servers, and beyond. The workgroup will build upon these efforts and pursue the long-term directions outlined in the Networking vision, focusing on work to: Define a unified networking stack with a coherent layered architecture, from shared I/O primitives at the foundation, through common protocol implementations, to ergonomic client and server APIs at the top. Define currency types that let libraries interoperate without coupling to specific implementations, such as IP addresses, hostnames, ports, and HTTP requests and responses. Evolve HTTP APIs by designing and guiding a modern, unified HTTP client and server API built on structured concurrency. Guide the evolution of shared protocol implementations (TLS, HTTP/1.1, HTTP/2, HTTP/3, QUIC, WebSockets) so improvements benefit the entire ecosystem rather than being duplicated across libraries. The new Networking workgroup joins a growing list of Swift workgroups, including the Android workgroup, Windows workgroup, and Build and Packaging workgroup, which were all added in the past year. To learn more and get involved: Discuss this announcement on the forums, and share ideas in the Networking category. Learn more about the Networking workgroup by reading its charter. The workgroup meets biweekly. A regular meeting time is being finalized and will be announced on the forums ahead of the first public meeting. Workgroup membership and meetings are open to those who wish to participate, and contributors are welcome! To receive an invite, send a message to @networking-workgroup on the Swift forums.

• DeepSeek 官方更新 · web:deepseek-updates:2026-08-13:8cf6f4d5069927123b6a624c0bab4d79f9038a0bda81182692f2444f8c47ee20 · api-docs.deepseek.com ↗ · 发布：2026-08-13T00:00:00+00:00 · 采集：2026-09-17T13:11:22.698038+00:00

### DeepSeek-V4-Pro Update[​](#deepseek-v4-pro-update "Direct link to DeepSeek-V4-Pro Update")

The GA release of DeepSeek-V4-Pro has been rolled out on the APP, Web, and API. The API calling method remains unchanged — simply set the model name to `deepseek-v4-pro` to use the latest version.

**Significantly enhanced Agent capabilities**

The GA version of DeepSeek V4 Pro greatly enhances agent capabilities, with particularly significant performance improvements in production environments.

-   HLE (wo / w tools): 42.7/60.0
-   Terminal Bench 2.1: 87.9
-   NL2Repo: 61.5
-   Cybergym: 83.3
-   DeepSWE: 62.7
-   Toolathlon-Verified: 74.1
-   Agents' Last Exam: 25.7
-   AutomationBench (Public): 31.8
-   DSBench-FullStack: 71.1
-   DSBench-Hard: 67.2

**Native support for the Responses API**

The DeepSeek API now natively supports the OpenAI Responses API format and is specifically adapted for Codex. Users can refer to the [official documentation](https://api-docs.deepseek.com/quick_start/agent_integrations/codex) and complete the Codex configuration with a one-click configuration script.

**More flexible thinking effort control**

The thinking modes of V4-Pro and V4-Flash now support three thinking effort levels: low / high / max. In real-world usage, users can flexibly choose based on task complexity: use low for simple tasks, high for daily Agent tasks, and max for more complex scenarios. For setup instructions, please refer to the official API documentation: [Thinking Mode](https://api-docs.deepseek.com/guides/thinking_mode).

**API Pricing Adjustment**

With the official release of the DeepSeek V4 model family, we will [update and adjust API pricing](https://api-docs.deepseek.com/quick_start/pricing). To allocate resources more reasonably, we will adopt peak/off-peak pricing, with off-peak prices set at half of the peak-hour prices, encouraging users to schedule their tasks based on actual usage. The new prices will take effect at 16:00 (UTC Time) on August 16, 2026.

For more details, please refer to [this documentation](https://api-docs.deepseek.com/news/news260813).

* * *

• ChatGPT 官方更新 · web:chatgpt-updates:2026-07-23:04103251be4a026ce99e28bf1102a36133fc4e8ee316014d58c2cd08ce65d7dc · help.openai.com ↗ · 发布：2026-07-23T00:00:00+00:00 · 采集：2026-09-17T13:11:16.322077+00:00

## **Health in ChatGPT**

We’re starting to roll out a new health experience in ChatGPT. Logged-in Free, Go, Plus, and Pro users in the United States who are 18 or older can securely connect supported health records and Apple Health data, see a dashboard of their health information, and ask questions grounded in their personal health context.

Available on web and iOS, Health lets you view lab results, medications, activity, sleep, and other health information in one place. You can ask questions directly in ChatGPT or use the Health tab to explore trends, prepare for appointments, understand test results, track wellness goals, and think through questions to discuss with a healthcare professional. The experience builds on feedback from early testers and gives you control over what you connect and when ChatGPT can use it.

Health is built with layered privacy and security safeguards. Connected medical records and Apple Health information, and conversations that use it, are not used to train our foundation models or target ads.

ChatGPT Health is designed to support, not replace, medical care and is not intended for diagnosis or treatment.

Learn more in our help article [Health in ChatGPT](https://help.openai.com/articles/20001036) and our blog [Launching Health in ChatGPT](https://openai.com/index/health-in-chatgpt/).

## **ChatGPT Voice in Work and Codex on desktop**

ChatGPT Voice is now available in Work and Codex in the ChatGPT desktop app. Start a new task in Voice to speak naturally, interrupt, and ask Voice to start or coordinate work using the tools and permissions available to the selected experience.

Learn more about [ChatGPT Voice](https://learn.chatgpt.com/docs/features/voice) and [pricing and limits](https://learn.chatgpt.com/docs/pricing#chatgpt-voice-in-desktop).

• ChatGPT 官方更新 · web:chatgpt-updates:2026-08-21:4162ce02966fccb4f4a3583583ed95ad5621b91380cbbe819f2cc51bfa8fa943 · help.openai.com ↗ · 发布：2026-08-21T00:00:00+00:00 · 采集：2026-09-17T13:11:16.322077+00:00

## **Improved plugin discovery on web and mobile**

Updated plugin recommendations make it easier to find tools people continue to use.

-   **Find more useful plugins.** Rankings now prioritize plugins people keep using after installation, helping relevant options surface across ChatGPT on web and mobile.

Plugin availability depends on your plan, region, and workspace settings. Desktop is not included in this update.

[Learn more about plugins in ChatGPT and Codex](https://help.openai.com/en/articles/20001256-plugins-in-chatgpt-and-codex).

## **More time-aware answers**

ChatGPT now has a better sense of your local time during a conversation.

-   **Get answers that fit your time of day.** ChatGPT can better account for your local time when responding to time-sensitive questions.

## **Faster long conversations on the web**

Long conversations now load more efficiently in ChatGPT on the web.

-   **Open longer chats without the wait.** ChatGPT loads messages in smaller sections instead of retrieving the entire conversation at once.

## **Interactive content appears sooner on the web**

Interactive content generated by ChatGPT can now begin appearing while it is still being created.

-   **See interactive content sooner.** Interactive experiences now appear progressively in the conversation instead of waiting until the full experience is ready.

## **Updates to ChatGPT on iOS**

Recent photos are easier to reach, and connection-status messages are clearer.

-   **Attach recent photos faster.** Press and hold the + button to open your recent photos, then choose one to add to your conversation.

-   **See clearer connection updates.** ChatGPT now tells you when it is waiting for an internet connection.

Available in the ChatGPT app for iOS. Access to recent photos requires permission to access your photos.

## **Updates to ChatGPT on Android**

Generated images are easier to view, and your recent conversations are easier to browse.

-   **See generated images at full width.** Images you create in ChatGPT now fill the width of the conversation.

-   **Keep recent chats easier to browse.** Your sidebar now shows up to eight recent conversations.

Available in the ChatGPT app for Android, version 1.2026.216 or later.

• Swift.org · rss:a76d8c3e94eba61fc46b5cc05fc51e5f6df1d0f46d901d32e9c27bb251f155ae:swift-blog:13 · swift.org ↗ · 发布：2026-06-12T10:00:00+00:00 · 采集：2026-09-17T07:07:56.208180+00:00

TrueType is a widely used vector font standard for rendering text in web pages, PDFs, operating systems, and applications. Familiar fonts like Helvetica, Garamond, and Monaco are all built on TrueType outlines. The format specifies a hinting interpreter intended to help outlines rasterize faithfully on low-resolution displays. Modern high-resolution displays enable beautiful typography from outlines alone, but TrueType fonts that need hinting to render legibly remain in use and we continue to support them. Font parsers process data from untrusted sources, making the TrueType hinting interpreter a security-critical attack surface. To make the format more resilient on Apple platforms, we rewrote its hinting interpreter from C to memory-safe Swift for the Fall 2025 releases. In addition to memory safety, we also improved performance: on average, our Swift interpreter runs 13% faster than the C interpreter it replaced. To accompany this post, we’ve also published the source code of the Swift TrueType hinting interpreter. We hope sharing our experience helps others doing similar work in Swift. TrueType and the hinting engine Apple developed TrueType in the late 1980s and released it with the launch of System 7 in 1991. TrueType was a major breakthrough for the time: it gave font developers enormous control over how glyphs are displayed, with an advanced grid-fitting algorithm and a sophisticated hinting engine built around a special-purpose bytecode interpreter. TrueType did all this on computers that were vastly less powerful than today’s, so it had to be extremely well-tuned for performance. Then the internet revolutionized how fonts were used. TrueType became embeddable in PDF files in 1994 and in web pages in 2008, and it remains as relevant as ever. However, these new use cases brought additional risk: TrueType could now be exposed to untrusted fonts from anywhere on the internet. TrueType fonts may contain programs the hinting engine runs through a bytecode interpreter. This interpreter involves input-driven control flow, complex data structures, and careful memory management—exactly the kind of code that’s hard to make perfect and where memory errors are easier to exploit. This high inherent complexity also makes correctness especially important. Rewriting in Swift A rewrite required a memory-safe language that could integrate into the existing codebase and provide an equivalent level of performance to the implementation it was replacing. Swift was the obvious choice for the task. Binary compatibility was crucial for this project to succeed: existing programs had to continue to function the same as they did before, effectively unaware that a new implementation was in place. This means not just interface compatibility but pixel-identical glyph rendering as well, relative to the C implementation. Hinting can radically change the on-screen appearance of glyphs, so a small change in the interpreter’s behavior could result in substantial user-visible changes. For this project, we defined correctness to mean exact compatibility with the C implementation’s outputs. Validating correctness To ensure correctness, we developed two test suites. The first was a unit test suite that can target both implementations, providing exhaustive (99.7%) code coverage for both. This suite is included with the open source release of the Swift interpreter. Then, to represent real-world workloads, we used a fuzzer to minimize a corpus of 10 million PDF files down to 4,200 without any loss of code coverage. The documents in the minimized corpus embedded 25,572 fonts with a total of 27 million glyphs that we rendered using four different transformations each, comparing the resulting bitmaps against the reference interpreter. This gave us confidence in the new interpreter’s compatibility. By the end of the project, we wrote nearly four times as many lines of test code as we wrote for the Swift interpreter itself. Achieving high performance Once our new implementation passed all its tests, we turned our attention to performance. We assessed performance at a high level using PDF render time, and then iterated on improvements guided by benchmarks that rendered all of the glyphs from three different fonts. These improvements fell into four main categories. Minimizing runtime overhead Swift uses automatic reference counting for managing the lifetime of shared reference types, and runtime exclusivity checking for preventing overlapping access to data structures. These sources of overhead are often exacerbated by aliasing, an irreducible amount of which existed in the interpreter’s specification by design. These sources of overhead can be eliminated by giving up the convenience of copyability, adopting ~Copyable value types (see also: struct instead of class) throughout the architecture, reserving reference types for high-level abstractions. Span, introduced in Swift 6.2 with back-deployment support all the way back to macOS 10.14.4 and iOS 12.2, allowed us to efficiently operate on sequences of these types. Moving data around Sometimes we want to change the “shape” of structured data when crossing language boundaries to better match the idioms on the other side. In Swift, glyph outlines are represented by a sequence of points, each of which carries a flag for whether it’s ‘on a curve’, flags for whether it’s been ‘touched’ on each axis, and three coordinate pairs: original (in the font’s base units), scaled (to the desired point size), and hinted (the interpreter program’s output). The original C code stored these points in one struct with eight arrays. This is good from a performance perspective because it’s cache-friendly: you can operate on a dimension of many points in long runs, which is fast. But exposing the data in Swift as a collection of point elements resulted in source code that was easier to follow. The initial cross-language bridging code we wrote prioritized expediency, safety, and simplicity by copying the glyph’s data from its C struct into Swift and then back after the program completed. Initially, those copies accounted for around 20% of the new interpreter’s runtime. In the end we wound up using projection types that provide safe access to the underlying C structure. Swift thus gives us readability without copying or otherwise transforming the underlying data structure. Following WebKit’s Safer Swift Guidelines, the example below demonstrates how to wrap a bridged structure from C in a projection type that uses Ref for lifetime safety, brokers bounds-safe access to the underlying data, and returns idiomatic Swift types to its callers. All unsafe expressions bear // SAFETY: comments that document the safety invariants and the reasoning which guarantees that they are true. @safe struct Zone: ~Copyable, ~Escapable { let _element: Ref<fnt_ElementType> @_lifetime(copy element) init(wrapping element: Ref<fnt_ElementType>) { // SAFETY: the `fnt_ElementType` passed by the caller must satisfy: // * `sp`, `ep` point at arrays of length ≥ `maxContourCount`. unsafe _element = element } func readContour(index: Int) -> ClosedRange<Int> { precondition(0..<contourCount ~= index) // SAFETY: `index` is bounds checked above; `_element.pointee.sp` and `_element.pointee.ep` // are live arrays of length `maxContourCount` per `init(wrapping:)` premises. return unsafe Int(_element.value.sp[index])...Int(_element.value.ep[index]) } } Short-lived allocations While runtime overhead imposes its costs in aggregate, and copying data across the language border showed up as a specific hot spot, the costs of short-lived memory allocations can show up in both ways. Operations like filter and map allocate memory, but that allocation is only necessary if the value escapes. The Swift standard library provides .lazy.map and .lazy.filter, but they don’t work in every case. For logic that only iterates over the filter or map, it’s much more efficient to loop with continue (or use for … in … where) and transform elements into local variables as necessary. Allocations that exist exclusively to pass results to a function’s caller can be elided too. For example, we often needed to pop the last n elements from the interpreter’s stack. The obvious implementation of this operation has to allocate space for the elements it returns (because the function simultaneously removes them) and looks something like: mutating func pop( count n: Int ) -> [Element] { defer { items.removeLast(n) } return Array(items.suffix(n)) } Optimizing this operation without compromising safety eventually led us to a continuation-passing approach where the caller passes a block that can operate on a slice of stack elements before they are removed. Swift’s compile-time exclusivity checking ensured that the stack could not be modified from inside the block, and the interface structurally eliminated any need for heap allocations or element copying. mutating func pop<R, E: Error>( count n: Int, _ op: (borrowing Span<Element>) throws(E) -> R ) throws(E) -> R { defer { items.removeLast(n) } return try op(items.span.extracting(last: n)) } Dynamic dispatch Abstraction mechanisms like protocols, generics, and inheritance are very powerful, but they introduce method-call indirection that may manifest at runtime as dynamic dispatch. This overhead can often be eliminated by the optimizer, but that is not possible in all conditions. In our case, not making abstractions more generic than necessary and encouraging the toolchain to inline were sufficient for the optimizer to hoist bounds checks and specialize all of our generic contexts. When you’re profiling your code, if you see unspecialized generics or protocol witness tables in hot paths, that’s a sign that the optimizer does not have sufficient visibility to optimize the call sites and that your implementation may benefit from inlining. Zero-cost abstractions At this point, anyone accustomed to grinding out performance improvements in a project might reasonably expect that our optimizations could come at the cost of readability, but in reality Swift’s type system and optimizer enabled us to employ abstractions that resulted in highly legible code. For example: FixedPoint types provided the same ergonomics as integral types, encapsulating complex rounding and shifting arithmetic. StackElement provided access to 32-bit values with built-in conversions for all eight of its supported numeric types. Our projection types provided safe and natural access to data that was not structured with those considerations in mind. Swift’s type system makes it possible to define powerful and expressive abstractions. When built with optimizations, all of our abstractions added zero cost while substantially improving readability. Memory-safe and faster than C Our goal for this project was to make the TrueType hinting interpreter fully memory-safe, to have the same observable rendering behavior as the C implementation, and to achieve a level of performance that did not regress any user-visible benchmarks. We met these goals. The Swift interpreter includes a small number of thoroughly verified unsafe statements at the language interop boundary; there have been no bugs reported against it since it was enabled; and it’s faster. On average, the Swift interpreter runs 13% faster than the C interpreter it replaced. Here is a chart showing the average CPU megacycles spent per glyph in the Swift implementation versus the C implementation for all of the hinted fonts that ship on macOS, plus a sampling of non-system fonts: ● System font · ● Non-system font · below the line: Swift faster · above: C faster Despite the overall performance improvement, we did not optimize everything! All of the new interpreter’s internal state was written in terms of noncopyable structures that were borrowed by its operations, but the top-level type itself is an @objc class that gets called across a module boundary from an Objective-C++ file. The hot paths are fast, and the cold paths are convenient. Swift in practice The Swift language made this project possible. Swift is memory-safe, has great ergonomics, and can be as fast as carefully written C. This makes it an excellent language for both application and systems development. Code using noncopyable types, value types, and Span is both safe and fast by default, and module-private types can be used to collectively define an architecture at no additional cost. Together with exhaustive test coverage, these well-defined internal interface boundaries make refactoring substantially easier, which in turn accelerates the measure-and-fix optimization loop while minimizing the risk of introducing bugs. This migration effort has deepened our Swift expertise and given us a foundation to build on. After completing the migration, we distilled what we learned into instructions for LLM coding assistants, and have since used them successfully in other projects. LLMs have improved the efficiency of our team’s work converting C/C++ to Swift, and have proven valuable in performing the kind of code transformations used in this effort. To accompany this post, we’ve published the source code for the Swift TrueType hinting interpreter on GitHub. This is production code, intended as a reference implementation rather than an ongoing open source project. We hope seeing how these techniques work in practice helps others achieve similar results.

• ChatGPT 官方更新 · web:chatgpt-updates:2026-06-18:5e264a27cf15d79e316e8b42adb0ac29319a286ee2b0c29257e33950dbd86383 · help.openai.com ↗ · 发布：2026-06-18T00:00:00+00:00 · 采集：2026-09-17T13:11:16.322077+00:00

## **ChatGPT app experience updates**

ChatGPT now has more ways to help with common questions, from pronunciation guidance to World Cup updates. It’s also smoother to use, with easier chat organization and sharing, more control over connected apps, and faster photo uploads on iOS, and more.

### Web, iOS, and Android

-   **Help with pronunciations**: ChatGPT can now include text and audio pronunciation guidance for words in more than 60 languages. Ask how to pronounce a word and ChatGPT can explain it with both audio and text in the conversation.

-   **World Cup updates**: ChatGPT can now help you follow the tournament in a more conversational way. Ask about schedules, matchups, teams and players, country storylines, predictions, and what different match outcomes mean. Get started at [chatgpt.com/football](https://chatgpt.com/football).

-   **App permission controls**: You can now choose when ChatGPT asks before using connected apps. Set ChatGPT to always ask, ask before making changes, or only ask before important changes. Manage preferences in Settings > Apps.

### **Web**

-   **Easier pinning and sidebar organization**: It’s now easier to keep important work close in ChatGPT. You can pin chats and projects directly from the sidebar, and your Pinned section now includes both chats and projects. You can also organize Recents as one combined list, or group chats by project so related work stays together.

-   **Easier conversation sharing**: Sharing a conversation is now faster, with a one-click flow that no longer opens a separate pop-up.

-   **Create a note from a response**: Highlight text in a ChatGPT response and choose ‘Start writing’ to keep working on it, or save it to your Library to come back to later.

### **iOS**

-   **Improved camera and photo uploads**: Taking and uploading photos in ChatGPT on iOS is now faster and smoother, making it easier to add images to your conversations.

### **Android**

-   **Long-press send to choose a model**: Android users on paid plans can now long-press the send button to choose a model for a one-off message without changing their default model.

-   **Updated composer and tools menu**: The composer in Android now has inline mentions, and an updated menu for Plugins.

## **Record & Replay in Codex**

Record & Replay is a new Codex app feature for macOS that lets eligible users demonstrate a workflow once and turn it into a reusable skill. It is useful for repetitive workflows or tasks that are easier to show than describe, such as creating a configured issue, publishing a video, or downloading a recurring report.
Record & Replay requires Computer Use to be available and enabled, and initial availability excludes the European Union, UK and Switzerland. Learn more in the [Record & Replay guide](https://developers.openai.com/codex/record-and-replay).

• Swift.org · rss:a76d8c3e94eba61fc46b5cc05fc51e5f6df1d0f46d901d32e9c27bb251f155ae:swift-blog:8 · swift.org ↗ · 发布：2026-04-08T10:00:00+00:00 · 采集：2026-09-17T07:07:56.208180+00:00

You can now write Swift in a broader range of popular IDEs, including Cursor, VSCodium, AWS’s Kiro, and Google’s Antigravity. By leveraging VS Code extension compatibility, these editors tap directly into the Open VSX Registry, where the official Swift extension is now live. Swift has long supported development using multiple IDEs including VS Code, Xcode, Neovim, and Emacs. Swift is also compatible with editors that implement the Language Server Protocol (LSP). This growing ecosystem of editor support is particularly significant as Swift continues to show its versatility across platforms and development environments, including agentic IDEs. Swift on Open VSX The Swift extension for VS Code is now officially available on the Open VSX Registry, the vendor-neutral, open source extension registry hosted by the Eclipse Foundation. The extension adds first-class language support for projects built with Swift Package Manager, enabling seamless cross-platform development on macOS, Linux, and Windows. This milestone brings Swift support, including code completion, refactoring, full debugging support, a test explorer, as well as DocC support, to a broader ecosystem of compatible editors and allows agentic IDEs like Cursor and Antigravity to automatically install Swift, with no manual download required. Swift in Cursor, powered by the Swift extension on Open VSX. Get Started To start using the Swift extension in any Open VSX-compatible editor, simply open the Extensions panel, search for ‘Swift’ and install the extension. If you’re using Cursor, getting started is easier than ever. Check out our new dedicated guide: Setting up Cursor for Swift Development. It walks you through the setup, features and includes how to configure custom Swift skills for your AI workflows. Swift now has support for a wider range of modern editors and IDEs to meet developers where they are. Download the extension, try it out in your editor of choice, and don’t forget to share your feedback!

• Swift.org · rss:a76d8c3e94eba61fc46b5cc05fc51e5f6df1d0f46d901d32e9c27bb251f155ae:swift-blog:3 · swift.org ↗ · 发布：2026-02-20T10:00:00+00:00 · 采集：2026-09-17T07:07:56.208180+00:00

We are excited to announce the 1.0 release of Swift System Metrics, a Swift package that collects process-level system metrics like CPU utilization time and memory usage. Swift System Metrics runs on both Linux and macOS, providing a common API across platforms. Swift System Metrics visualized in Grafana, demonstrating what's possible with real-time monitoring. Monitoring process metrics enables you to detect performance issues, optimize resource usage, and ensure your service remains reliable and cost-effective under varying loads. You can integrate Swift System Metrics into your service in just a few lines of code, making observability accessible to every developer and ensuring that even the smallest services can have production-grade visibility from day one. Swift System Metrics is part of a larger set of packages that provide an end-to-end solution for integrating metrics into your Swift applications and services. Once system metrics are collected, they’re reported to Swift Metrics, a backend-agnostic metrics API that can work with popular backends like Prometheus and OpenTelemetry. Swift System Metrics also leverages Swift Service Lifecycle to handle process bootstrapping and resource cleanup. With the 1.0 milestone, the API is now stable and ready for use. Note that this package was previously swift-metrics-extras, and renamed to better reflect its purpose. Highlights Collects and reports: CPU utilization time Virtual and resident memory usage Open and maximum available file descriptors Process start time API-stable public interface Support on both Linux and macOS musl libc compatibility The package includes an example Grafana dashboard configuration to start visualizing metrics immediately. Get Started Add the dependency to your Package.swift: .package(url: "https://github.com/apple/swift-system-metrics", from: "1.0.0") Add the library dependency to your target: .product(name: "SystemMetrics", package: "swift-system-metrics") Import and use in your code: import SystemMetrics import ServiceLifecycle import Logging import OTel @main struct Application { static func main() async throws { // Create a logger, or use one of the existing loggers let logger = Logger(label: "Application") // Setup MetricsSystem, for example using swift-otel var otelConfig = OTel.Configuration.default otelConfig.serviceName = "Application" let otelService = try OTel.bootstrap(configuration: otelConfig) // Setup your service let service = FooService() // Create the monitor let systemMetricsMonitor = SystemMetricsMonitor(logger: logger) // Create the service let serviceGroup = ServiceGroup( services: [otelService, service, systemMetricsMonitor], gracefulShutdownSignals: [.sigint], cancellationSignals: [.sigterm], logger: logger ) try await serviceGroup.run() } } The complete documentation is available on Swift Package Index. Get Involved We’re looking for contributions to grow the list of process metrics collected and to expand platform support. PRs are welcome - our contribution guidelines describe how to get started. By reaching 1.0, this project will maintain a backwards-compatible API as it continues to evolve. Thanks to everyone who contributed to this release.

• ChatGPT 官方更新 · web:chatgpt-updates:2026-07-09:cc788486e91f4ce2b9366414d13b705851589abc2a97e82e810f86f6fbec96fb · help.openai.com ↗ · 发布：2026-07-09T00:00:00+00:00 · 采集：2026-09-17T13:11:16.322077+00:00

## **Introducing ChatGPT Work**

ChatGPT Work is an agent for longer, more involved tasks. It can research and analyze information, work across connected apps and files, and create finished documents, spreadsheets, presentations, reports, and Sites. You can follow its progress, answer questions, change direction, and approve important actions as it works.

Work can also keep projects moving through Scheduled Tasks that run once, repeat on a schedule or trigger, or monitor for changes.

On web and mobile, Work is rolling out to paid plans except Free and Go. Pro, Pro Lite, Enterprise, and Edu users receive access first; Plus and Business users will follow over the coming days. Work is available in all supported regions.

On web and mobile, Enterprise and Edu workspaces have a two-week preview period. Work is off by default during the preview, and admins can opt out before it turns on automatically at the end of the preview.

We’re also replacing the App Directory with the Plugin Directory. Existing app connections are unaffected. Plugins can package skills, apps, and app templates for specific workflows, and are available from ChatGPT on web and desktop, including Work and Codex.

[Learn more about ChatGPT Work and Codex](https://help.openai.com/articles/20001275).

[Learn more about plugins and the new plugin directory.](https://help.openai.com/en/articles/20001256-plugins-in-chatgpt-and-codex)

## **Skills are now generally available**

Enterprise and Edu users can find skills under Plugins in the sidebar, on the Skills tab in the Plugin Directory.

## **The new ChatGPT desktop app brings Chat, Work, and Codex together**

The new ChatGPT desktop app is available globally for macOS and Windows. It combines Chat for questions and conversation, Work for research and finished deliverables, and Codex for software development in one app.

On desktop, Work can use local files and desktop apps with your permission. The built-in browser lets ChatGPT gather information from websites and work with supported web-based tools and files. Codex also adds inline editing within diffs, pull-request review in the side panel, faster computer use, and support for multiple repositories in one project.

If you already use the Codex app, update it as usual to move to the new ChatGPT desktop app. If you use the previous ChatGPT desktop app, follow the prompt in the app to download the new version. The previous app may remain installed as ChatGPT Classic. ChatGPT Classic continues to receive model updates, bug fixes, security patches, and support for its existing Enterprise capabilities, while new agent features such as Work and Codex are available in the new app.

[Learn more about moving to the new ChatGPT desktop app.](https://help.openai.com/articles/20001276)

## **Introducing ChatGPT Sites in public beta**

ChatGPT Sites lets you turn your work or ideas into an interactive website or lightweight app without leaving ChatGPT. You can create dashboards, project trackers, launch calendars, prototypes, internal portals, and reports, then preview and refine the result with ChatGPT before sharing it.

Business and Enterprise customers already have access to Sites. Starting today, they can also publish Sites publicly and share them with anyone through a URL, expanding Sites beyond people in their workspace.

To create a Site, start in ChatGPT Work on the web or in Work or Codex in the ChatGPT desktop app. Describe what you want to build, add any relevant content, files, data, links, or constraints, and review the private preview before publishing.

In Enterprise workspaces, public publishing is off by default and must be enabled by an admin. Sites is also rolling out in public beta to Pro, Pro Lite, and Edu users, with Plus users following over the coming days. Sites is not available on Free or Go plans. Public publishing and the expanded beta rollout are not available in the EEA, Switzerland, or the United Kingdom at launch.

[Learn more about creating and managing ChatGPT Sites.](https://help.openai.com/articles/20001339)

## **Retiring group chats in ChatGPT**

Beginning July 9, users on web, iOS, and Android will no longer be able to create new group chats, turn existing conversations into group chats, or join a group chat through an invite link.

Existing group chats will remain available and can continue to be used for now. When an existing group chat becomes read-only, users will retain access to the messages, files, and images already shared there. Individual ChatGPT conversations are unaffected.

## **Retiring Atlas**

We’re deprecating Atlas as we bring browser-based agentic capabilities into ChatGPT and Codex. Atlas is scheduled to stop working on August 9, 2026.
We want to make this transition as easy as possible. Before then, move any Atlas data you want to keep. Atlas browser data, including bookmarks, open tabs, and browser history, will not transfer automatically. You can export your cookies and passwords to the ChatGPT desktop app and your bookmarks to Chrome. Be sure to save any important pages or URLs from your open tabs and browser history before August 9.
Your ChatGPT conversation history is separate and will remain available in ChatGPT, subject to your plan, workspace settings, and account access.
We’re building on what we learned from Atlas to bring a more capable browser experience to ChatGPT, including multiple tabs, downloads, improved navigation, and account login support where available. For deeper browser-based agentic work, try the ChatGPT desktop app. For help while browsing in Chrome, try the ChatGPT Chrome extension or sidebar, where available.
You can learn more about moving off Atlas in our [Help Center article](https://help.openai.com/articles/20001371).

• Swift.org · rss:a76d8c3e94eba61fc46b5cc05fc51e5f6df1d0f46d901d32e9c27bb251f155ae:swift-blog:5 · swift.org ↗ · 发布：2026-03-06T11:00:00+00:00 · 采集：2026-09-17T07:07:56.208180+00:00

TelemetryDeck is an app analytics service specifically for developers, designed to manage usage analytics that are anonymized, privacy-focused, and really easy to use. TelemetryDeck is managing the data of over 16 million people every month, helping thousands of app publishers improve their products, and we’re doing it all with a Swift-based infrastructure. The decision to go with Swift brought a lot of unexpected advantages for us. We come from a world of iOS in the frontend, Python, Node, or Ruby in the backend for server-based applications. Compared to these languages and frameworks, Swift is just as easy to use, and its compiled nature allows us to catch a lot of possible errors at compile time instead of runtime, making it ideal for a hardened, high-performance web service. Let me share our journey and history with you. Adopting Swift for backend services TelemetryDeck is built on Vapor, a web framework written in Swift, for building scalable APIs and services. Similar to other Vapor-based projects like Things, TelemetryDeck runs on containers hosted in Kubernetes. Our metadata lives in Postgres, and our analytics data is stored in Apache Druid. We use Swift-native connectors to access APIs for these services. Some of those come from the Swift community, others we’ve written ourselves and are giving back to the open source ecosystem. In the very beginning, when TelemetryDeck was just a little exploratory hobby project, we wanted to go with Swift on the server simply because we love Swift, and we’re pretty decent at it. Vapor was new, and we figured, let’s try this out and learn a new way to apply and combine our skills. In the end, this decision turned out to be exactly the right one and resulted in a lean architecture that is highly performant, stable, and allows us to develop and iterate quickly. Swift’s performance characteristics are remarkable, especially in multithreading. Python’s Global Interpreter Lock has long constrained true parallelism, though recent versions have improved this. Swift doesn’t face that same challenge. Our infrastructure handles 16 million users per month with resources that would buckle other architectures. The efficiency gains aren’t just theoretical — they translate directly to lower infrastructure costs and a better user experience. Swift’s Codable protocol In an API-based application, you spend an enormous amount of time encoding and decoding JSON. Swift’s Codable protocol turns what’s traditionally error-prone boilerplate into straightforward, type-safe operations. When a request comes in with malformed data, Swift’s type system rejects it immediately with no manual validation required. This isn’t just convenient; it’s a security feature that prevents entire classes of vulnerabilities. struct Notebook: Codable, Content { let appID: UUID var snapshots: [NotebookSnapshot] let createdAt: Date let title: String } With just these few lines, we get parsing, validation, and type safety. Just like the postal system – which only accepts letters addressed in a specific format – the API endpoint described in the above bit of code will reject any data that doesn’t fit into this Codable. We don’t have to check for types, additional properties, or JSON validity, something that lets us trust our APIs more when developing and sleep easy after deploying. “Just lick the stamp and send it,” as a friend of mine once encouraged me. Sharing Data Transfer Objects between Client and Server For a long time, TelemetryDeck offered a SwiftUI-based desktop application for interacting with analytics data, displaying dashboards, and exploring correlations. This meant we could use the same structs for encoding data on the server and decoding it on the client – these are called Data Transfer Objects or DTOs. We kept a separate package “SwiftDTOs” that was shared between both clients and the server code. However, we ran into trouble when we updated our APIs to new versions with slightly different inputs and outputs. We always have the API version in our paths, which helps us continuously develop them while keeping compatibility with clients that take a while to update. This pattern made it hard to update Data Transfer Objects quickly because changing a struct for one API version could break another. For our newest API version, we abandoned this approach. DTOs now live directly in their controllers as inline structs: extension V3QueryController { struct QueryRequest: Codable { let query: QueryDefinition let timeRange: TimeRange } struct QueryResponse: Codable { let results: [QueryResult] let metadata: ResponseMetadata } } The lesson? Keep your DTOs close to your controllers. It’s simpler, more maintainable, and makes changes easier. Sometimes the obvious solution is the right one. Developer experience matters One underappreciated aspect of Swift on the server is the development experience. Swift services can be developed using your favorite IDE, including VS Code, Cursor, Xcode, and editors supporting the Language Server Protocol (LSP). We chose to develop in Xcode, our favorite IDE, with full debugging capabilities. Under the hood, a Swift Vapor project is just a Swift Package with some dependencies. You open it in Xcode, you press Cmd+U to run the tests, you run it and set breakpoints and step through them when you’re following a bug through the stack. Our web API services usually have a database attached to them, so when we’re developing, we can either run PostgreSQL on our machines right next to the code, or we can ask Swift Vapor’s database layer (Fluent) to use a different database for development, such as a local SQLite file. The end result is a very comfortable experience that allows us to use an IDE specifically built for Swift development, without needing to spin up containers. This flexibility matters when you’re building a business. We can onboard new developers quickly because the tooling is familiar. We can iterate rapidly because the feedback loop is tight. And we can maintain high code quality because Swift’s type system catches errors at compile time, not in production. Building the ecosystem we want to use We love the open source ecosystem around Swift on the server. There are Swift packages for external services like payment providers and transactional emails. There are tools like Imperial for Single Sign-On, and Swift Crypto for standards-compliant cryptography. Whenever a piece of the puzzle is missing, we try to give back as much as possible. That is partly because the Swift developer community is fantastic in coming up with SDKs for various services, and we try to contribute to existing projects like StripeKit (where we added support for a lot of APIs that Stripe released) or even release our own new SDKs like BrevoKit. Lessons learned Building a production service teaches you things that tutorials never will. Here are some hard-won lessons from our journey: Use the Swift Package system. A package is a fantastic way to structure and isolate code. Since our services are packages anyway, it’s natural to further subdivide them where it makes sense. Your Swift Vapor service is almost never the bottleneck. Although profiling Swift code is easy, most slowdowns in a web service will arise from a combination of database, storage, and traffic load. You’ll migrate the way you store and access data frequently as you gain experience. Embrace Vapor’s built-in features. Queues and jobs, migrations, and middleware. Vapor has mature solutions for common patterns. We use Fluent almost exclusively over raw SQL, and Vapor’s queue system handles everything from data processing to billing tasks. Version your API URLs from day one. You’ll thank yourself later when you need to evolve your API without breaking existing clients. Always set cache expiration times. It sounds obvious, but that Redis cache that “will never grow too large” absolutely will. Set reasonable TTLs on everything that you cache. Monitor errors and performance. When you build an iPhone app, you automatically get crash logs that help you find out where things are going wrong for your users. On the server, you’ll have to collect those yourself. Before we started doing that, our customers would discover problems before we did, and that never feels good or professional. Now we use a combination of Swift Distributed Tracing and our own analytics-based middleware to find performance bottlenecks and errors that are predictors of service crashes, and we can right the ship before most of our customers notice. Where we are now TelemetryDeck proves that Swift isn’t just for apps — it’s a legitimate choice for building scalable, performant backend services. We’re processing millions of events, serving thousands of developers, and doing it all with Swift and Vapor. The language that powers apps on billions of iOS devices is equally capable of powering the services behind them. The performance gains are real: we can crunch and encode tens of thousands of data points in milliseconds. The development velocity is fantastic: we work in Xcode and write tests for behaviour and not for data parsing. And the type safety prevents entire categories of bugs that dynamically typed languages like Python or JavaScript need additional code to guard against. Perhaps most importantly, the joy of writing Swift makes the long hours of building a business more bearable. If you’re considering Swift for your next server-side project, take it from someone who’s built an entire business on it: Swift delivers on its promise of safe, fast, and expressive code—whether that code runs on an iPhone or in a data center.

• moondustkk · x:x-following:following:moondustkk:2100547659429798319 · x.com ↗ · 发布：2026-09-17T11:29:22+00:00 · 采集：2026-09-17T12:00:39.632171+00:00

A new coalition of tech giants aims to speed AI data center grid connections in exchange for more flexibility. https://t.co/BgKN3Nayp6

• moondustkk · x:x-following:following:moondustkk:2100543615709646901 · x.com ↗ · 发布：2026-09-17T11:13:18+00:00 · 采集：2026-09-17T12:00:53.612960+00:00

Jacob Coxon quit his job—and the world freaked out. How an anonymous math whiz became the face of the AI apocalypse. https://t.co/34IFuVSHT0

• moondustkk · x:x-following:following:moondustkk:2100540547840475447 · x.com ↗ · 发布：2026-09-17T11:01:07+00:00 · 采集：2026-09-17T12:01:16.784167+00:00

Comp AI, a cybersecurity and compliance startup, announced a $34 million Series A round led by Roo Capital and Grand Ventures.
https://t.co/XNbUdHfNaU

• moondustkk · x:x-following:following:moondustkk:2100534351301578774 · x.com ↗ · 发布：2026-09-17T10:36:30+00:00 · 采集：2026-09-17T12:01:44.868345+00:00

ChatGPT, Character. ai, and Snapchat My AI will all be in the crosshairs as the AI Kids Act aims to strip minor-facing chatbots of all the things that make people love them. https://t.co/IRHcUXkdw6

• moondustkk · x:x-following:following:moondustkk:2100530202786812327 · x.com ↗ · 发布：2026-09-17T10:20:00+00:00 · 采集：2026-09-17T12:16:19.925552+00:00

An era-defining mathematics claim raises questions over how AI tools credit earlier work.

https://t.co/wjloYPHhpJ

• moondustkk · x:x-following:following:moondustkk:2100526465020317758 · x.com ↗ · 发布：2026-09-17T10:05:09+00:00 · 采集：2026-09-17T12:16:34.305294+00:00

The Oversight Board is asking Meta to prioritize AI content when it's reported. https://t.co/pa8xkHLMXr

• moondustkk · x:x-following:following:moondustkk:2100516864686244066 · x.com ↗ · 发布：2026-09-17T09:27:00+00:00 · 采集：2026-09-17T10:01:00.482761+00:00

Nature research paper: Reimagining research papers as interactive and reliable AI agents

https://t.co/t1s2K29gU7

• moondustkk · x:x-following:following:moondustkk:2100514258941276503 · x.com ↗ · 发布：2026-09-17T09:16:39+00:00 · 采集：2026-09-17T09:19:59.780031+00:00

RT @perrymetzger: Any time someone tells you that AI has an X% chance of wiping out humanity, ask them for their math that got them the num…

• moondustkk · x:x-following:following:moondustkk:2100489640255869414 · x.com ↗ · 发布：2026-09-17T07:38:50+00:00 · 采集：2026-09-17T09:21:15.311879+00:00

RT @ArthurMacwaters: Just to recap the last week 

- Dario agreed with a ngo-coded physiognomy “whistleblower” that AI has >10% chance of e…

• moondustkk · x:x-following:following:moondustkk:2100459279434027195 · x.com ↗ · 发布：2026-09-17T05:38:11+00:00 · 采集：2026-09-17T12:18:06.972726+00:00

RT @askalphaxiv: Six degrees of Yann LeCun

We mapped the coauthorship graph of every AI researcher on arXiv 🚀

Enter any two names to find…

• moondustkk · x:x-following:following:moondustkk:2100450954923425994 · x.com ↗ · 发布：2026-09-17T05:05:06+00:00 · 采集：2026-09-17T12:18:39.215503+00:00

Treble's voice simulation platform is used by voice AI model developers, AI wearable, and robotics companies
https://t.co/hmrhgd697K

• moondustkk · x:x-following:following:moondustkk:2100448648672993540 · x.com ↗ · 发布：2026-09-17T04:55:56+00:00 · 采集：2026-09-17T12:19:04.841564+00:00

Incredibly exciting that there are entire universes of AI innovation that still exist that weren’t even on most of our radars.

Being able to process information insanely quickly, at crazy low costs, with high levels of capability is huge for a wide number of enterprise tasks.

Data classification tasks, routing decisions inside of a workflow, decision making when handed a particular domain problem, quick judgment calls about safety or security, and more all are the gates in a large number of  processes. 

This model and approach could be quite cool in agentic workflows in the enterprise.

• moondustkk · x:x-following:following:moondustkk:2100392476649390500 · x.com ↗ · 发布：2026-09-17T01:12:44+00:00 · 采集：2026-09-17T14:06:22.494016+00:00

Over the past week, a global reckoning about the risks of AI spread from the frontier labs of Silicon Valley to living rooms across America and the corridors of power in Washington and Beijing. And it all started with a guy nobody had ever heard of. https://t.co/D354fF8mdb

• moondustkk · x:x-following:following:moondustkk:2100387319639375937 · x.com ↗ · 发布：2026-09-17T00:52:14+00:00 · 采集：2026-09-17T14:06:22.494115+00:00

Even as Trump dismisses calls for tighter federal regulation of the AI industry, the quiet freakout among some of his most senior aides over what could be one of the administration’s most consequential policy decisions has begun. https://t.co/8tX3pygxfA

• chatgpt · x:chatgpt:user:chatgpt:2098457932438577407 · x.com ↗ · 发布：2026-09-11T17:05:33+00:00 · 采集：2026-09-17T07:53:54.198192+00:00

An interactive visual story about AI, emotional connection, and whether an encounter can feel meaningful. 
https://t.co/g8gjwUVt1G https://t.co/G6VtbQb1YF

• moondustkk · x:x-following:following:moondustkk:2100276142225973560 · x.com ↗ · 发布：2026-09-16T17:30:28+00:00 · 采集：2026-09-17T14:06:04.864594+00:00

A few thoughts on ZCash (disclosure: Paradigm is an investor in ZEC and ZODL):

- All blockchain ecosystems struggle with long-term funding, especially of public goods, and an inflation-funded developer fund is an elegant mechanism to do so. How this fund is allocated is a fair debate, and we should strive to improve it. But, we believe the dev fund is important, particularly in this age of AI cyber capabilities, rate of quantum progress, etc.

- As ZCash grows in acceptance and adoption as a private complement to Bitcoin, we think pure coinvoting governance introduces unpredictability that could limit long-term trust as a monetary asset, and it is better for the project to combine coinvoting with other forms of governance.

• chatgpt · x:chatgpt:user:chatgpt:2100299088847855819 · x.com ↗ · 发布：2026-09-16T19:01:39+00:00 · 采集：2026-09-17T07:53:18.601309+00:00

RT @simpsoka: What features are we missing in ChatGPT sites? What should we add?

• moondustkk · x:x-following:following:moondustkk:2100586770572177551 · x.com ↗ · 发布：2026-09-17T14:04:47+00:00 · 采集：2026-09-17T14:07:27.795058+00:00

These discoveries are just the beginning.

AlphaGenome Atlas is freely accessible to empower researchers everywhere to decode the genetic causes of disease. Explore the database → https://t.co/j1T7iTF1K6

• moondustkk · x:x-following:following:moondustkk:2100586768248562157 · x.com ↗ · 发布：2026-09-17T14:04:47+00:00 · 采集：2026-09-17T14:07:27.794914+00:00

Every second, millions of genome switches dictate how our cells function and adapt. 🧬

Working with @ScienceStowers, Atlas mapped 2,500+ regulatory patterns uncovering the DNA sequences that act as biological switches and volume dials to turn gene activity up or down across hundreds of cell types.

• moondustkk · x:x-following:following:moondustkk:2100586760036143330 · x.com ↗ · 发布：2026-09-17T14:04:45+00:00 · 采集：2026-09-17T14:07:27.794818+00:00

Researchers at @BroadInstitute, @UniofExeter, and beyond are already using AlphaGenome Atlas to better identify potential disease-causing DNA variants and interpret their role. 🧵 https://t.co/j4OuSz0Djv

• moondustkk · x:x-following:following:moondustkk:2100586685134147730 · x.com ↗ · 发布：2026-09-17T14:04:27+00:00 · 采集：2026-09-17T14:07:27.795094+00:00

RT @FischerKing64: Colleges and universities are so much more expensive because parasites called ‘administrators’ must all get paid. At the…

• moondustkk · x:x-following:following:moondustkk:2100586633154158917 · x.com ↗ · 发布：2026-09-17T14:04:15+00:00 · 采集：2026-09-17T14:07:27.795150+00:00

Tendon cells from soft tissue regained their shape and ability to proliferate after returning from stiff surfaces—though only some gene-expression changes recovered. [Link] @QMUL @advscinews https://t.co/avODT9nU2n

• moondustkk · x:x-following:following:moondustkk:2100585947452567859 · x.com ↗ · 发布：2026-09-17T14:01:31+00:00 · 采集：2026-09-17T14:07:27.795168+00:00

RT @XFreeze: Final day of Grok Bot Galaxy starts in next few few hours 

Day 3 — September 17

• 9:00–10:30 AM — Grok Bot for Marketing Ope…

• moondustkk · x:x-following:following:moondustkk:2100585497886134542 · x.com ↗ · 发布：2026-09-17T13:59:44+00:00 · 采集：2026-09-17T14:00:32.143635+00:00

RT @WallStreetApes: This insane: California and China both started their High Speed Rail projects in 2008

- California hasn’t completed a…

• moondustkk · x:x-following:following:moondustkk:2100585163285532766 · x.com ↗ · 发布：2026-09-17T13:58:24+00:00 · 采集：2026-09-17T14:00:39.538716+00:00

RT @zerohedge: Dragged Off Her Own Porch: York College Students Beaten After Five 911 Calls - And A National Pattern Of Youth Street Mobs h…

• moondustkk · x:x-following:following:moondustkk:2100585063054266616 · x.com ↗ · 发布：2026-09-17T13:58:00+00:00 · 采集：2026-09-17T14:00:49.992795+00:00

La impresionante luz! 

La Basílica de los Santos Ambrosio y Carlos en el Corso es una de las joyas barrocas más imponentes y luminosas de Roma, dedicada a los santos patronos de Milán.

La cúpula de Pietro da Cortona es una de las más grandes y espectaculares de Roma. Cortona la diseñó con un tambor alto calado por grandes ventanales, lo que inunda el crucero con una luz dorada celestial que hace brillar los estucos y frescos interiores a distintas horas del día.

El fresco de la bóveda fue pintado por Giacinto Brandi, representa La caída de los ángeles rebeldes. Es una obra maestra del claroscuro barroco, llena de dinamismo, cuerpos entrelazados y un dramatismo teatral que cambia por completo según cómo entre la luz por las naves.

Es una de las pocas iglesias en Roma que cuenta con un pasillo semicircular detrás del altar mayor (al estilo de las catedrales góticas francesas). 

Es un rincón extraordinario para capturar perspectivas fotográficas únicas, juegos de sombras y el detalle del relicario que custodia el corazón de San Carlos Borromeo.

La fachada de Carlo Borromeo y la riqueza de sus mármoles polícromos hacen que cada rincón sea un festín para los amantes del detalle arquitectónico.

• moondustkk · x:x-following:following:moondustkk:2100584793851224361 · x.com ↗ · 发布：2026-09-17T13:56:56+00:00 · 采集：2026-09-17T14:00:57.975986+00:00

RT @nickshirleyy: 🚨 Welcome to the $15 billion California High Speed Rail to nowhere:
Since 2008 California has spent $15 billion and has n…

• moondustkk · x:x-following:following:moondustkk:2100584071306908118 · x.com ↗ · 发布：2026-09-17T13:54:04+00:00 · 采集：2026-09-17T14:01:05.236956+00:00

RT @SpringerNature: Peer review doesn't have to be a black box. Nature's transparent peer review helps researchers see how papers are revie…

• moondustkk · x:x-following:following:moondustkk:2100578088169959444 · x.com ↗ · 发布：2026-09-17T13:30:17+00:00 · 采集：2026-09-17T14:01:15.584032+00:00

What if the factory belongs closer to public life, not further from it?

👉 Explore collection: https://t.co/8u8SDSfSjM
✍️ By: Eirini Makarouni https://t.co/VPWvCcPRGt

• moondustkk · x:x-following:following:moondustkk:2100575753490358620 · x.com ↗ · 发布：2026-09-17T13:21:01+00:00 · 采集：2026-09-17T14:01:34.768804+00:00

To keep their Washington vacation property under $1 million, the couple spread the square footage across two levels and handled the landscaping themselves. https://t.co/HeGlXlGWhT

• moondustkk · x:x-following:following:moondustkk:2100575297527546170 · x.com ↗ · 发布：2026-09-17T13:19:12+00:00 · 采集：2026-09-17T14:01:43.496560+00:00

It's the most sweeping proposal yet to limit kids' access to social media. https://t.co/2EOzJ88cuU

• moondustkk · x:x-following:following:moondustkk:2100574263950364728 · x.com ↗ · 发布：2026-09-17T13:15:05+00:00 · 采集：2026-09-17T14:02:02.774252+00:00

Patriots owner Robert Kraft made it clear to the singer that he would cancel his tour dates at Gillette Stadium unless Macklemore was dropped from the bill. Other NFL owners were ready to do the same.

Read more: 🔗 https://t.co/RqCrpMfeYU https://t.co/7lBjwTL3iR

• moondustkk · x:x-following:following:moondustkk:2100573989647061247 · x.com ↗ · 发布：2026-09-17T13:14:00+00:00 · 采集：2026-09-17T14:02:11.737278+00:00

Nature research paper: Electroluminescent photoresists extending lithographic scaling to OLEDs

https://t.co/51bC0NFRDs

• moondustkk · x:x-following:following:moondustkk:2100571511635472873 · x.com ↗ · 发布：2026-09-17T13:04:09+00:00 · 采集：2026-09-17T14:02:25.302034+00:00

Both new open-ear products go on sale next month. https://t.co/QqeaGUJ8KE

• moondustkk · x:x-following:following:moondustkk:2100571505386041538 · x.com ↗ · 发布：2026-09-17T13:04:08+00:00 · 采集：2026-09-17T14:02:41.278977+00:00

Trying the apps out also entitles testers to giving feedback that will influence the final versions. https://t.co/hFhylY4Qcs

• moondustkk · x:x-following:following:moondustkk:2100571186857996349 · x.com ↗ · 发布：2026-09-17T13:02:52+00:00 · 采集：2026-09-17T14:02:49.776800+00:00

RT @TechXplore_com: A 0.6-nanometer gallium oxide film retained switchable ferroelectric polarization at 0.8 V, enabled by strain-induced a…

• moondustkk · x:x-following:following:moondustkk:2100570465324384726 · x.com ↗ · 发布：2026-09-17T13:00:00+00:00 · 采集：2026-09-17T14:03:02.939782+00:00

During WWI, journalist Harriet Chalmers Adams traveled to wartime Lorraine, documenting a landscape where beautiful villages and the scars of war existed side by side. Her 1917 account offers a glimpse of France during this turbulent time. 🔗 https://t.co/ZhQiyosqZS

• moondustkk · x:x-following:following:moondustkk:2100563557364437164 · x.com ↗ · 发布：2026-09-17T12:32:33+00:00 · 采集：2026-09-17T14:03:11.591240+00:00

RT @medical_xpress: A 12-week eating-window trial cut a nerve-damage biomarker by 13% in early Huntington’s disease, while participants mai…

• moondustkk · x:x-following:following:moondustkk:2100560103501894063 · x.com ↗ · 发布：2026-09-17T12:18:49+00:00 · 采集：2026-09-17T12:24:34.957022+00:00

Warning against Japan’s military build-up surfaces at Beijing security conference https://t.co/dBYOBrA8ZT

• moondustkk · x:x-following:following:moondustkk:2100557577809543424 · x.com ↗ · 发布：2026-09-17T12:08:47+00:00 · 采集：2026-09-17T12:15:39.109960+00:00

'Tantalizing' data could point towards Nefertiti's burial place

https://t.co/U0e27lW2Un

• moondustkk · x:x-following:following:moondustkk:2100557316697252201 · x.com ↗ · 发布：2026-09-17T12:07:45+00:00 · 采集：2026-09-17T12:15:43.408983+00:00

Tata Sons extends Chandrasekaran's term as chairman but a battle looms https://t.co/sNGTwb1LQB

• moondustkk · x:x-following:following:moondustkk:2100556751447663081 · x.com ↗ · 发布：2026-09-17T12:05:30+00:00 · 采集：2026-09-17T12:15:47.622167+00:00

Swedish PM Kristersson resigns after election loss https://t.co/Jv734WJ1db

• moondustkk · x:x-following:following:moondustkk:2100556442516279658 · x.com ↗ · 发布：2026-09-17T12:04:17+00:00 · 采集：2026-09-17T12:15:51.974605+00:00

An underrated part of DJI's success is in the area of video transmission. https://t.co/eJ88Phs4C9

• moondustkk · x:x-following:following:moondustkk:2100556436375810332 · x.com ↗ · 发布：2026-09-17T12:04:15+00:00 · 采集：2026-09-17T12:15:56.186217+00:00

Flash drives can vary wildly in cost. While a cheap drive will get the job done, a premium one offers extra sped and features that might be worth the money. https://t.co/4M80fG0Z6k

• moondustkk · x:x-following:following:moondustkk:2100556429904036201 · x.com ↗ · 发布：2026-09-17T12:04:13+00:00 · 采集：2026-09-17T12:16:00.391896+00:00

The new wearable's controversial audio features aren't available yet, but there are plenty of upgrades here. https://t.co/btLdjtTak1

• moondustkk · x:x-following:following:moondustkk:2100556274517643462 · x.com ↗ · 发布：2026-09-17T12:03:36+00:00 · 采集：2026-09-17T12:16:04.786489+00:00

Persistent poverty in childhood more than triples the likelihood of being NEET at 17 or 23—and 56% of those NEET at 17 remained so at 23. https://t.co/8qfG2cOav0

• moondustkk · x:x-following:following:moondustkk:2100555459698553003 · x.com ↗ · 发布：2026-09-17T12:00:22+00:00 · 采集：2026-09-17T12:02:38.618482+00:00

#ProjectOfTheDay: Netherlands American Cemetery Visitor Center by KAAN Architecten

The Visitor Center at the Netherlands American Cemetery creates a new place for remembrance without disrupting the cemetery’s monumental landscape. Set below ground level, the building is defined by a floating concrete frame and broad glass walls that bring the surrounding hills into the experience.    

📍Margraten, Netherlands #ArchitizerAwards
Details: https://t.co/zYllz2sORL

• moondustkk · x:x-following:following:moondustkk:2100555047327211980 · x.com ↗ · 发布：2026-09-17T11:58:44+00:00 · 采集：2026-09-17T12:00:20.872161+00:00

RTÉ boycotts Eurovision for second year in a row https://t.co/CsJbyHDhBy

• moondustkk · x:x-following:following:moondustkk:2100554843769258462 · x.com ↗ · 发布：2026-09-17T11:57:55+00:00 · 采集：2026-09-17T12:00:25.845991+00:00

La lección de música, pintada por el maestro holandés Johannes Vermeer entre 1662 y 1665, es una de las cumbres pictóricas del Siglo de Oro neerlandés.

Esta obra maestra en óleo sobre lienzo destaca por su asombroso dominio de la luz, el uso matemático de la perspectiva lineal y sus sutiles dobles sentidos moralizantes y románticos. 

Actualmente forma parte de la Colección Real Británica (Royal Collection Trust) y se custodia en el Palacio de St. James.

Fiel al estilo minimalista y enigmático de Vermeer, la aparente "clase de música" esconde una fuerte tensión psicológica y romántica.

Vermeer utiliza una técnica de cuadrícula geométrica rígida que sitúa al espectador al fondo de la habitación. 

Los objetos del primer plano, como la enorme mesa cubierta por una alfombra turca y la silla de tapicería azul, actúan como barreras físicas que protegen la privacidad de los personajes.

Colgado sobre el instrumento hay un espejo inclinado. En él se revela el rostro oculto de la joven (que de espaldas no podemos ver) mirando fijamente hacia el hombre. 

De manera fascinante, en el reflejo también se alcanzan a ver las patas del caballete del propio Vermeer, lo que sitúa al artista (y a nosotros) compartiendo en secreto la misma habitación.

En la tapa del instrumento musical se lee en latín: “MUSICA LETITIAE CO[ME]S / MEDICINA DOLOR[IS]”. Esto se traduce como "La música es la compañera del placer y el bálsamo para el dolor". 

En el arte holandés de la época, la música era un código visual directo para referirse al cortejo amoroso y la armonía entre géneros.

Detrás del caballero cuelga una pintura real: La Caridad Romana del pintor Dirck van Baburen. 

Esta obra, que pertenecía físicamente a la suegra de Vermeer, retrata a una hija amamantando a su padre encarcelado para salvarlo de morir de hambre; una alegoría sobre el amor ideal, el deber y la piedad que refuerza el tema subyacente de la escena.

Video original realizado para efectos educativos y culturales:

• moondustkk · x:x-following:following:moondustkk:2100550274779644343 · x.com ↗ · 发布：2026-09-17T11:39:46+00:00 · 采集：2026-09-17T12:00:30.097809+00:00

China hits back as US Congress approves tariffs on top buyers of Russian energy https://t.co/Q4EK6TnyNG

• moondustkk · x:x-following:following:moondustkk:2100548069947257215 · x.com ↗ · 发布：2026-09-17T11:31:00+00:00 · 采集：2026-09-17T12:00:35.069697+00:00

Layers of management help bigger bosses get stuff done. Their loss will come back to bite companies sooner than they realize. https://t.co/1zxbzPySxg

• moondustkk · x:x-following:following:moondustkk:2100546309304303717 · x.com ↗ · 发布：2026-09-17T11:24:01+00:00 · 采集：2026-09-17T12:00:44.324537+00:00

Patriots owner Robert Kraft made it clear to the singer that he would cancel his tour dates at Gillette Stadium unless Macklemore was dropped from the bill. Other NFL owners were ready to do the same. https://t.co/tHw48zak0m

• moondustkk · x:x-following:following:moondustkk:2100546307244847492 · x.com ↗ · 发布：2026-09-17T11:24:00+00:00 · 采集：2026-09-17T12:00:49.051938+00:00

Latest from the Nature Podcast 🔊 Transplanted human brain-tissue takes root in mice without a cortex

https://t.co/zv4SJtP5x1

• moondustkk · x:x-following:following:moondustkk:2100542574263980120 · x.com ↗ · 发布：2026-09-17T11:09:10+00:00 · 采集：2026-09-17T12:00:57.878666+00:00

Te gustaría tener este tipo de arte en tu comunidad? 

Este mural forma parte del movimiento de arte callejero en los callejones traseros de Hamilton, Ontario.

Una vibrante expresión de cultura comunitaria que ha transformado los pasajes de servicio y las puertas de garaje de la ciudad en auténticas galerías al aire libre.

El renacer de los callejones industriales
Hamilton posee un profundo pasado industrial y obrero. 

Durante décadas, sus callejones traseros (alleyways) sirvieron como espacios puramente funcionales, destinados al paso de vehículos de servicio y propensos al vandalismo o al descuido visual. 

Sin embargo, en los últimos años, colectivos de artistas locales y asociaciones vecinales impulsaron iniciativas como la transformación de pasajes urbanos (como el conocido Gallery Alley) y proyectos comunitarios de muralismo (visibles en guías de arte urbano de la ciudad).

Esta intervención en los garajes representa una apropiación del espacio público-privado. 

Los dueños de las viviendas ceden voluntariamente sus portones para albergar estas obras, creando una "tradición de pasillo" donde los vecinos caminan y disfrutan de un circuito artístico oculto a las avenidas principales. 

La iconografía de este tipo de murales suele evocar un espíritu lúdico, nostálgico o vacacional que contrasta radicalmente con el clima invernal típico de la provincia de Ontario.

A diferencia del grafiti convencional hecho enteramente con aerosol, este estilo de muralismo en madera utiliza bloques de color plano aplicados con brocha y rodillo, lo que otorga una textura uniforme y mayor durabilidad frente a las duras condiciones climáticas de Canadá.

Video grabado personalmente para efectos culturales y educativos:

• moondustkk · x:x-following:following:moondustkk:2100542489480638944 · x.com ↗ · 发布：2026-09-17T11:08:50+00:00 · 采集：2026-09-17T12:01:02.481690+00:00

「在街上感受城鄉紋理，經常會需要查找相關資料。以下是幾個我最常參考的來源。1. 全國土地分區資料查詢系統。2. 地籍圖資網路便民服務系統。3. 全國建築執照存根查詢。4. 高雄地籍圖資服務網。」（街景背後的城鄉紋理：土地分區、地籍圖、登記建物與使用執照 – Taiwan 2.0 https://t.co/siPgewfU4X ）

• moondustkk · x:x-following:following:moondustkk:2100541462400127262 · x.com ↗ · 发布：2026-09-17T11:04:45+00:00 · 采集：2026-09-17T12:01:07.672949+00:00

A grassroots live tour by YouTuber Louis Rossmann highlights growing resistance to companies that keep people from controlling their devices. https://t.co/KftHsUCxxp

• moondustkk · x:x-following:following:moondustkk:2100540900627644518 · x.com ↗ · 发布：2026-09-17T11:02:31+00:00 · 采集：2026-09-17T12:01:12.470740+00:00

Aztec document goes on display in Mexico after 186 years abroad https://t.co/ftUZ1gHh30

• moondustkk · x:x-following:following:moondustkk:2100538785339490692 · x.com ↗ · 发布：2026-09-17T10:54:07+00:00 · 采集：2026-09-17T12:01:21.301333+00:00

OpenAI says it doesn't believe the industry has solved its problems 'to a sufficient degree to continue responsibly scaling at maximum speed for much longer.' https://t.co/MpRFhuiWsw

• moondustkk · x:x-following:following:moondustkk:2100538506346905655 · x.com ↗ · 发布：2026-09-17T10:53:00+00:00 · 采集：2026-09-17T12:01:25.925622+00:00

Malaysia confronts complexities of schizophrenia after school killing ruling https://t.co/M7MRN7m7Jk

• moondustkk · x:x-following:following:moondustkk:2100536924385509438 · x.com ↗ · 发布：2026-09-17T10:46:43+00:00 · 采集：2026-09-17T12:01:30.813927+00:00

Politician 'will not be deterred' after assault in Dublin https://t.co/7TxIaobo8W

• moondustkk · x:x-following:following:moondustkk:2100536328605893118 · x.com ↗ · 发布：2026-09-17T10:44:21+00:00 · 采集：2026-09-17T12:01:35.495550+00:00

Discipline https://t.co/sDwCEIHxKO

• moondustkk · x:x-following:following:moondustkk:2100536274469745005 · x.com ↗ · 发布：2026-09-17T10:44:08+00:00 · 采集：2026-09-17T12:01:40.413169+00:00

"El Inoportuno", es una de las pinturas tardías más fascinantes del maestro realista brasileño José Ferraz de Almeida Júnior. Fue realizada en 1898 —apenas un año antes de su trágica muerte— y forma parte del acervo permanente de la Pinacoteca del Estado de São Paulo. https://t.co/1JUERCWduz

• moondustkk · x:x-following:following:moondustkk:2100534173656109229 · x.com ↗ · 发布：2026-09-17T10:35:47+00:00 · 采集：2026-09-17T12:01:49.163729+00:00

Vin Diesel, the star of the series, didn’t make a planned appearance at the grand opening of its dynamite new theme park ride. But it didn’t stop his fans from having a good time. https://t.co/j8d63LZqzV

• moondustkk · x:x-following:following:moondustkk:2100532535730008116 · x.com ↗ · 发布：2026-09-17T10:29:17+00:00 · 采集：2026-09-17T12:16:09.410756+00:00

RT @daithaigilbert: EXCLUSIVE:

Data shared with @wired by @statesunited show that 197 election deniers will contest congressional and stat…

• moondustkk · x:x-following:following:moondustkk:2100531160707133604 · x.com ↗ · 发布：2026-09-17T10:23:49+00:00 · 采集：2026-09-17T12:16:14.011022+00:00

RT @BBCBreaking: Canada welcomes EU ambition for it to become "associate member", Prime Minister Mark Carney tells European Parliament http…

• moondustkk · x:x-following:following:moondustkk:2100528435713953901 · x.com ↗ · 发布：2026-09-17T10:12:59+00:00 · 采集：2026-09-17T12:16:24.693443+00:00

These young minds are advancing research and industry across energy, agriculture, and critical materials. https://t.co/AEf1tpjfwY

• moondustkk · x:x-following:following:moondustkk:2100526920530059362 · x.com ↗ · 发布：2026-09-17T10:06:58+00:00 · 采集：2026-09-17T12:16:29.483945+00:00

Exclusive | ‘Provocation’ could accelerate Taiwan reunification: retired PLA major general https://t.co/nhi6QTzosa

• moondustkk · x:x-following:following:moondustkk:2100526377938096323 · x.com ↗ · 发布：2026-09-17T10:04:49+00:00 · 采集：2026-09-17T12:16:38.634400+00:00

197 election deniers are running for Congress or statewide office in 43 states. Many of them are seeking roles that would put them in charge of voting for the 2028 presidential election. https://t.co/QDgeFsMzlA

• moondustkk · x:x-following:following:moondustkk:2100525213326672225 · x.com ↗ · 发布：2026-09-17T10:00:11+00:00 · 采集：2026-09-17T10:00:24.061406+00:00

#ArchitizerAwards

Set against native forest in Porto Feliz, AMS House is organized as three volumes stepping with the terrain. The central social block opens through large glass panels toward the garden and pool, while rammed-earth walls provide privacy and thermal stability.

📍 Porto Feliz, Brazil
Project by: Jacobsen Studio
Details: https://t.co/8VZHtbqD9G

• moondustkk · x:x-following:following:moondustkk:2100525202274738597 · x.com ↗ · 发布：2026-09-17T10:00:08+00:00 · 采集：2026-09-17T10:00:28.777722+00:00

Keep your daily pumpkin spice habit going without overspending on Starbucks. https://t.co/mtusi7KLL1

• moondustkk · x:x-following:following:moondustkk:2100523900194955737 · x.com ↗ · 发布：2026-09-17T09:54:58+00:00 · 采集：2026-09-17T10:00:33.310291+00:00

Ninth woman's body found as South African police investigate string of deaths https://t.co/rIntfNdh6B

• moondustkk · x:x-following:following:moondustkk:2100522808824447372 · x.com ↗ · 发布：2026-09-17T09:50:38+00:00 · 采集：2026-09-17T10:00:37.828489+00:00

Investors have been betting the Federal Reserve is at the start of a series of interest-rate increases. On Wednesday, Chairman Kevin Warsh gave them little reason to think otherwise. https://t.co/mCeYzo7ACY

• moondustkk · x:x-following:following:moondustkk:2100519414974083258 · x.com ↗ · 发布：2026-09-17T09:37:08+00:00 · 采集：2026-09-17T10:00:42.408090+00:00

The Defense Department issued a legal waiver authorizing current and former personnel to share unidentified anomalous phenomena–related classified information, but it doesn’t extend to other agencies. https://t.co/vXBiYBQ9hO

• moondustkk · x:x-following:following:moondustkk:2100518975268393211 · x.com ↗ · 发布：2026-09-17T09:35:24+00:00 · 采集：2026-09-17T10:00:47.084728+00:00

Controversy around OpenAI’s claim to have solved the Navier–Stokes problem highlights how researchers could be inadvertently sharing — and absorbing— ideas through chatbots.

https://t.co/jLBE5nI8P5

• moondustkk · x:x-following:following:moondustkk:2100518972504322232 · x.com ↗ · 发布：2026-09-17T09:35:23+00:00 · 采集：2026-09-17T10:00:51.253984+00:00

Indonesia faces reckoning over maritime safety after deadly ferry disaster https://t.co/E0yripQTK7

• moondustkk · x:x-following:following:moondustkk:2100517655824179594 · x.com ↗ · 发布：2026-09-17T09:30:09+00:00 · 采集：2026-09-17T10:00:55.890701+00:00

Here is an early look at the front page of today's Wall Street Journal https://t.co/Dd6hDFtvQ7 https://t.co/QUeOpdyhFi

• moondustkk · x:x-following:following:moondustkk:2100513984302530822 · x.com ↗ · 发布：2026-09-17T09:15:34+00:00 · 采集：2026-09-17T09:20:04.225754+00:00

RT @Plinz: I wish people would notice that the safety scare is not driven by people outside of the big labs, but directly from their leader…

• moondustkk · x:x-following:following:moondustkk:2100513896050114708 · x.com ↗ · 发布：2026-09-17T09:15:13+00:00 · 采集：2026-09-17T09:20:08.483352+00:00

RT @stevesi: Our framework for reporting model misalignment https://t.co/o1gVtyTWkk // if you scrape away all the anthropomorphic language,…

• moondustkk · x:x-following:following:moondustkk:2100513670719508510 · x.com ↗ · 发布：2026-09-17T09:14:19+00:00 · 采集：2026-09-17T09:20:12.824227+00:00

Hyrox Beijing athletes get full refund, new rule on hygiene risks comes into force https://t.co/4o18CFeFlU

• moondustkk · x:x-following:following:moondustkk:2100513440640962649 · x.com ↗ · 发布：2026-09-17T09:13:24+00:00 · 采集：2026-09-17T09:20:17.233095+00:00

Hyrox Beijing athletes get full refund, new rule on hygiene risks comes into force https://t.co/i8YM2581XM

• moondustkk · x:x-following:following:moondustkk:2100511286979170498 · x.com ↗ · 发布：2026-09-17T09:04:51+00:00 · 采集：2026-09-17T09:20:21.704992+00:00

Flock reveals little about where its license plate readers are assembled, but the answer could have geopolitical and cybersecurity implications. https://t.co/gIx0Tf4Bso

• moondustkk · x:x-following:following:moondustkk:2100511255299489996 · x.com ↗ · 发布：2026-09-17T09:04:43+00:00 · 采集：2026-09-17T09:20:25.590710+00:00

Taipei donates coastguard vessel to Manila amid South China Sea tensions https://t.co/kUMBMqCG1d

• moondustkk · x:x-following:following:moondustkk:2100508056249327648 · x.com ↗ · 发布：2026-09-17T08:52:00+00:00 · 采集：2026-09-17T09:20:29.944510+00:00

This ingenious experiment shows how a virus's protein shell is assembled

https://t.co/K8iJiXJelz

• moondustkk · x:x-following:following:moondustkk:2100505776145997956 · x.com ↗ · 发布：2026-09-17T08:42:57+00:00 · 采集：2026-09-17T09:20:34.376560+00:00

Algae bloom that crippled Israel's desalination plants raises water security concerns https://t.co/thrO8vwUAY

• moondustkk · x:x-following:following:moondustkk:2100505213815673143 · x.com ↗ · 发布：2026-09-17T08:40:43+00:00 · 采集：2026-09-17T09:20:38.777744+00:00

One dead after torrential rain and flash floods hit Barcelona region https://t.co/ItY3ROSX1N

• moondustkk · x:x-following:following:moondustkk:2100502653977407679 · x.com ↗ · 发布：2026-09-17T08:30:32+00:00 · 采集：2026-09-17T09:20:43.109148+00:00

The tragic Qing politician whose legacy lives on in descendant’s new novel https://t.co/a2EKmnGVMe

• moondustkk · x:x-following:following:moondustkk:2100502521487741153 · x.com ↗ · 发布：2026-09-17T08:30:01+00:00 · 采集：2026-09-17T09:20:47.421508+00:00

Nature research paper: Mitochondrial fission factor senses and governs ferroptosis

https://t.co/WsvgjYyRI1

• moondustkk · x:x-following:following:moondustkk:2100502419079586052 · x.com ↗ · 发布：2026-09-17T08:29:36+00:00 · 采集：2026-09-17T09:20:51.563390+00:00

Trump presses Xi on fentanyl precursors ahead of Washington summit https://t.co/PZdOos5gRf

• moondustkk · x:x-following:following:moondustkk:2100501808451821705 · x.com ↗ · 发布：2026-09-17T08:27:11+00:00 · 采集：2026-09-17T09:20:56.013571+00:00

EU announces plan to restrict social media access for children https://t.co/MH2DZ6Xrym

• moondustkk · x:x-following:following:moondustkk:2100501250009686419 · x.com ↗ · 发布：2026-09-17T08:24:58+00:00 · 采集：2026-09-17T09:21:02.584504+00:00

True

• moondustkk · x:x-following:following:moondustkk:2100499084029112669 · x.com ↗ · 发布：2026-09-17T08:16:21+00:00 · 采集：2026-09-17T09:21:06.806685+00:00

‘Depraved’ Singaporean paedophile who abused daughter, granddaughters, jailed https://t.co/cpMX6LCpUc

• moondustkk · x:x-following:following:moondustkk:2100498967670767936 · x.com ↗ · 发布：2026-09-17T08:15:53+00:00 · 采集：2026-09-17T09:21:11.117011+00:00

India's biggest stock exchange launches mega share sale https://t.co/ECX5tcJ8NE

• moondustkk · x:x-following:following:moondustkk:2100488763289821227 · x.com ↗ · 发布：2026-09-17T07:35:21+00:00 · 采集：2026-09-17T12:16:47.617106+00:00

RT @SpaceX: Liftoff! https://t.co/okVd5WzMOT

• moondustkk · x:x-following:following:moondustkk:2100486538777837835 · x.com ↗ · 发布：2026-09-17T07:26:30+00:00 · 采集：2026-09-17T09:21:19.431423+00:00

Philippines’ Teodoro draws 2028 election buzz after China clash https://t.co/Rkt1OKC45h

• moondustkk · x:x-following:following:moondustkk:2100484237929406710 · x.com ↗ · 发布：2026-09-17T07:17:22+00:00 · 采集：2026-09-17T09:21:24.286664+00:00

Trump threatens 'serious tariffs' on EU over proposal to make Canada 'associate member' https://t.co/ztDaUohAvL

• moondustkk · x:x-following:following:moondustkk:2100484207248048547 · x.com ↗ · 发布：2026-09-17T07:17:14+00:00 · 采集：2026-09-17T12:16:52.403269+00:00

RT @aaronburnett: My favorite phrase from this episode “symphony of iteration”.  Also one SpaceX engineer estimated that starship surviving…

• moondustkk · x:x-following:following:moondustkk:2100479880999600616 · x.com ↗ · 发布：2026-09-17T07:00:03+00:00 · 采集：2026-09-17T12:16:56.943976+00:00

Hong Kong police officer left paralysed in 2005 knife attack dies https://t.co/inxniqx2SY

• moondustkk · x:x-following:following:moondustkk:2100479842726674637 · x.com ↗ · 发布：2026-09-17T06:59:54+00:00 · 采集：2026-09-17T12:17:01.304528+00:00

22 months’ jail for Hong Kong driver who fatally struck student at junction https://t.co/gUhjpIydWt

• moondustkk · x:x-following:following:moondustkk:2100477465751671235 · x.com ↗ · 发布：2026-09-17T06:50:27+00:00 · 采集：2026-09-17T12:17:06.131682+00:00

Abducted South African journalist found dead with gunshot wounds https://t.co/nukYvB6oWO

• moondustkk · x:x-following:following:moondustkk:2100476341606871240 · x.com ↗ · 发布：2026-09-17T06:45:59+00:00 · 采集：2026-09-17T12:17:10.748332+00:00

Russia employs North Koreans to build drones for Ukraine war: report https://t.co/1NhTFNBvz2

• moondustkk · x:x-following:following:moondustkk:2100475911044837835 · x.com ↗ · 发布：2026-09-17T06:44:16+00:00 · 采集：2026-09-17T12:17:15.001766+00:00

Indian transgender patients left in limbo following legal amendments https://t.co/hH7kmqXVWy

• moondustkk · x:x-following:following:moondustkk:2100473772591521978 · x.com ↗ · 发布：2026-09-17T06:35:46+00:00 · 采集：2026-09-17T12:17:19.612700+00:00

RT @ArthurMacwaters: I still think about this all the time

• moondustkk · x:x-following:following:moondustkk:2100473700873117994 · x.com ↗ · 发布：2026-09-17T06:35:29+00:00 · 采集：2026-09-17T12:17:24.222748+00:00

RT @cb_doge: That night SpaceX put everything on the line.

December 21, 2015. Nearly 11 years later and it hits just as hard.

Return to f…

• moondustkk · x:x-following:following:moondustkk:2100472354501529808 · x.com ↗ · 发布：2026-09-17T06:30:08+00:00 · 采集：2026-09-17T12:17:29.876812+00:00

India and Pakistan summon diplomats after ships collide in international waters https://t.co/TSB1lhptDQ

• moondustkk · x:x-following:following:moondustkk:2100470884956098872 · x.com ↗ · 发布：2026-09-17T06:24:18+00:00 · 采集：2026-09-17T12:17:34.541491+00:00

RT @spencerpratt: Soon, LA will just be a ghost town with a 100K illegal street vendors cooking rat meat in front of abandoned brick and mo…

• moondustkk · x:x-following:following:moondustkk:2100470006349779066 · x.com ↗ · 发布：2026-09-17T06:20:49+00:00 · 采集：2026-09-17T12:17:39.691444+00:00

Exclusive | BYD, Xiaomi, Innolight likely on guest list for Xi Jinping’s US trip: sources https://t.co/63HZ4ratOK

• moondustkk · x:x-following:following:moondustkk:2100468561017802872 · x.com ↗ · 发布：2026-09-17T06:15:04+00:00 · 采集：2026-09-17T12:17:45.131239+00:00

If your newly updated iPhone suddenly freezes, don't panic. Here's how to fix it. https://t.co/Yn65PBEYj6

• moondustkk · x:x-following:following:moondustkk:2100464625993052185 · x.com ↗ · 发布：2026-09-17T05:59:26+00:00 · 采集：2026-09-17T12:17:54.226495+00:00

Falcon 9 launches USSF-259 to orbit from pad 4E in California https://t.co/eKB269BlYE

• moondustkk · x:x-following:following:moondustkk:2100462189706453093 · x.com ↗ · 发布：2026-09-17T05:49:45+00:00 · 采集：2026-09-17T12:17:58.895487+00:00

Is Japan’s Okinawa finally moving on from anti-base politics? https://t.co/WWX8x7QdYL

• moondustkk · x:x-following:following:moondustkk:2100460995667750948 · x.com ↗ · 发布：2026-09-17T05:45:00+00:00 · 采集：2026-09-17T12:18:02.190508+00:00

MLB teams are using position players on the mound more than ever before, transforming what was once a novelty into a common—and controversial—strategy. https://t.co/bW81r9TPdh

• moondustkk · x:x-following:following:moondustkk:2100458768811049350 · x.com ↗ · 发布：2026-09-17T05:36:09+00:00 · 采集：2026-09-17T12:18:12.724894+00:00

India faces 100% tariff threat over Russian oil after US House vote https://t.co/sX57PkRue2

• moondustkk · x:x-following:following:moondustkk:2100457755853091261 · x.com ↗ · 发布：2026-09-17T05:32:08+00:00 · 采集：2026-09-17T12:18:15.959893+00:00

RT @BBCNews: UK prosecutors file for extradition of the Tate brothers from the US https://t.co/0tmz6ToQdF

• moondustkk · x:x-following:following:moondustkk:2100455356279243157 · x.com ↗ · 发布：2026-09-17T05:22:36+00:00 · 采集：2026-09-17T12:18:18.630692+00:00

Climate change helped trigger Nepal-Tibet disaster, study finds https://t.co/azAvDboNBo

• moondustkk · x:x-following:following:moondustkk:2100453097449402569 · x.com ↗ · 发布：2026-09-17T05:13:37+00:00 · 采集：2026-09-17T12:18:21.909450+00:00

Vigil takes place for Nepal flash flood disaster https://t.co/l3vWSX6W8J

• moondustkk · x:x-following:following:moondustkk:2100452766917284001 · x.com ↗ · 发布：2026-09-17T05:12:18+00:00 · 采集：2026-09-17T12:18:25.807561+00:00

Cheat code unlocked. https://t.co/JJxElf6ICU

• moondustkk · x:x-following:following:moondustkk:2100451967877157220 · x.com ↗ · 发布：2026-09-17T05:09:08+00:00 · 采集：2026-09-17T12:18:28.876535+00:00

Rare tree offspring given lifeline at Scottish botanic garden https://t.co/TkB3RqHh92

• moondustkk · x:x-following:following:moondustkk:2100450848279011367 · x.com ↗ · 发布：2026-09-17T05:04:41+00:00 · 采集：2026-09-17T12:18:42.375251+00:00

Rosenberg: Russia's elections will have few surprises, but for the Kremlin they matter https://t.co/fnFkc2yH7k

• moondustkk · x:x-following:following:moondustkk:2100450607240814856 · x.com ↗ · 发布：2026-09-17T05:03:43+00:00 · 采集：2026-09-17T12:18:51.609503+00:00

Risk of unrest in Asia as Iran war, El Nino threaten food supplies https://t.co/fQ7KmkUS9B

• moondustkk · x:x-following:following:moondustkk:2100450271218352312 · x.com ↗ · 发布：2026-09-17T05:02:23+00:00 · 采集：2026-09-17T12:18:58.957888+00:00

Assad regime planned US journalist's kidnap for weeks, BBC finds https://t.co/sYhPJ6NSB1

• moondustkk · x:x-following:following:moondustkk:2100449679074812113 · x.com ↗ · 发布：2026-09-17T05:00:02+00:00 · 采集：2026-09-17T12:19:01.868024+00:00

When No. 6 Oregon suffered a stunning upset, fans across the country wondered what could have gone wrong. An ex-college passer grinding the tape in his living room had the answer. https://t.co/yfsaAqXZo0

• Claude 官方更新 · web:claude-updates:2026-09-15:30398d689db8166fcd4fe4ff63d54385928d186d67acd000978514a1f2f2da30 · support.claude.com ↗ · 发布：2026-09-15T00:00:00+00:00 · 采集：2026-09-17T13:11:20.521154+00:00

**Launching Salesforce in Claude (beta)**

We've launched Salesforce in Claude, a plugin that brings a seller’s accounts, opportunities, and pipeline into Claude with 37 pre-built sales skills. Prep a call, review a deal, create a pipeline dashboard, and send your forecast all from Claude. It’s now available in beta on all paid plans for organizations Salesforce approves through its beta sign-up. For more information, see our blog post: **[Bringing Salesforce into Claude](http://claude.com/blog/salesforce-in-claude)**.

• moondustkk · x:x-following:following:moondustkk:2100445938128982219 · x.com ↗ · 发布：2026-09-17T04:45:10+00:00 · 采集：2026-09-17T12:19:08.011238+00:00

RT @NASAAdmin: The Moon is not the finish line. It’s the next starting point.

There is a lot we can do with the resources already on the M…

• moondustkk · x:x-following:following:moondustkk:2100442178757488791 · x.com ↗ · 发布：2026-09-17T04:30:14+00:00 · 采集：2026-09-17T12:19:11.020823+00:00

Here is an early look at the front page of The Wall Street Journal https://t.co/AkArFNvhZ7 https://t.co/anWP5DnXaE

• moondustkk · x:x-following:following:moondustkk:2100438959205614062 · x.com ↗ · 发布：2026-09-17T04:17:26+00:00 · 采集：2026-09-17T12:19:13.823280+00:00

Los jugadores de cartas no es una sola obra, sino una de las series pictóricas más trascendentales de la historia del arte moderno, realizada por el maestro postimpresionista francés Paul Cézanne en la década de 1890. 

El año 1892 marca el corazón de este proyecto, periodo en el que consolidó sus piezas principales.

Cézanne utilizó como modelos a los campesinos y trabajadores de las fincas de su natal Aix-en-Provence, logrando convertir una escena cotidiana de taberna en un monumental estudio geométrico y psicológico. 

El artista ejecutó cinco lienzos diferentes sobre este tema, reduciendo progresivamente los elementos decorativos y los personajes para enfocarse estrictamente en la forma, el color y la concentración silenciosa.

Está versión es la de la Barnes Foundation Filadelfia de 1890–1892 y es la más grande (134 x 181 cm) y compleja. Muestra a cinco personajes.

Esta pintura destaca sobre las demás de la serie por ser la más grande y compleja, incluyendo elementos que desaparecieron en las simplificaciones posteriores. 

Tres jugadores principales en primer plano y dos observadores (el hombre de la pipa a la izquierda y el niño con la mirada baja a la derecha).

Se aprecian claramente elementos ambientales como un cuadro colgado en la pared del fondo, cuatro pipas rústicas de arcilla sujetas en un armazón, una repisa de madera en la esquina superior izquierda y un pesado cortinaje amarillo que enmarca el lateral derecho.

La mesa blanca rústica incluye un cajón central que Cézanne decidió omitir en composiciones posteriores para centrar la atención exclusivamente en las manos y los naipes.

• moondustkk · x:x-following:following:moondustkk:2100438883422863814 · x.com ↗ · 发布：2026-09-17T04:17:08+00:00 · 采集：2026-09-17T12:19:16.743334+00:00

India, Pakistan summon diplomats after naval ships collide in Arabian Sea https://t.co/fvMNXkYnZa

• moondustkk · x:x-following:following:moondustkk:2100437812591923267 · x.com ↗ · 发布：2026-09-17T04:12:53+00:00 · 采集：2026-09-17T12:19:19.570397+00:00

US again blocks Palestinian leader from UN meeting in New York https://t.co/MhEvYw4L1t

• moondustkk · x:x-following:following:moondustkk:2100434718357332332 · x.com ↗ · 发布：2026-09-17T04:00:35+00:00 · 采集：2026-09-17T12:19:23.103200+00:00

Set on a hillside overlooking the sea,this home is organized around a central garden and pool. Its four-storey form draws from Pattaya’s urban patterns, mountain terrain and branching trees, shaping terraces and layered openings around the family’s everyday life.

📍 Pattaya City, Thailand
Project by: SPACE | STORY | STUDIO
Details: https://t.co/03FzUefqAq

• moondustkk · x:x-following:following:moondustkk:2100434625986265219 · x.com ↗ · 发布：2026-09-17T04:00:13+00:00 · 采集：2026-09-17T12:19:26.174291+00:00

The StarDial service will also allow phone service in RVs and on boats. https://t.co/4ZxMY7KQ1f

• moondustkk · x:x-following:following:moondustkk:2100434319734981043 · x.com ↗ · 发布：2026-09-17T03:59:00+00:00 · 采集：2026-09-17T14:05:43.306419+00:00

Old-school fashion rules are stale. These are the guidelines to know now. https://t.co/7GJGiAvDtu

• moondustkk · x:x-following:following:moondustkk:2100428707328672088 · x.com ↗ · 发布：2026-09-17T03:36:42+00:00 · 采集：2026-09-17T14:05:48.554391+00:00

RT @judegomila: I've attempted to map everything in oncology in a public website and open source repo for all. 

The site is trying to get…

• moondustkk · x:x-following:following:moondustkk:2100427802550075618 · x.com ↗ · 发布：2026-09-17T03:33:06+00:00 · 采集：2026-09-17T14:06:22.493418+00:00

@matthuang To those who have a governance plan that doesn’t ultimately have coin holder voting underneath- explain why those specific people and organizations get to be in charge, and not others. 

How do you allocate credit and power? 

How do you prevent fake orgs and fake stakeholders?

• moondustkk · x:x-following:following:moondustkk:2100426280483057791 · x.com ↗ · 发布：2026-09-17T03:27:03+00:00 · 采集：2026-09-17T14:06:22.493674+00:00

Safest place to do it. https://t.co/OtLoVH5yAI

• moondustkk · x:x-following:following:moondustkk:2100425310894469447 · x.com ↗ · 发布：2026-09-17T03:23:12+00:00 · 采集：2026-09-17T14:06:14.195569+00:00

@matthuang Point two is akin to saying that shareholders shouldn’t control a corporation. 

It anoints trusted third parties off-chain using pleasant-sounding words like “community.” It creates sybil attacks and unending politics.

Blockchains are markets, not democracies or aristocracies.

• moondustkk · x:x-following:following:moondustkk:2100424527553462612 · x.com ↗ · 发布：2026-09-17T03:20:06+00:00 · 采集：2026-09-17T14:06:22.493719+00:00

From a casual LinkedIn message to a police raid at an upscale home: Inside the bizarre paternity case that sparked the “Justice for Clayton” movement. 🔗 https://t.co/SRXjtWzlmr https://t.co/CCuvrOHZL7

• moondustkk · x:x-following:following:moondustkk:2100420755016085708 · x.com ↗ · 发布：2026-09-17T03:05:06+00:00 · 采集：2026-09-17T14:06:22.493747+00:00

When No. 6 Oregon suffered a stunning upset, fans across the country wondered what could have gone wrong. An ex-college passer grinding the tape in his living room had the answer. 🏈 https://t.co/cTkEvoZM2l https://t.co/fqWc2KguoH

• moondustkk · x:x-following:following:moondustkk:2100419220236386754 · x.com ↗ · 发布：2026-09-17T02:59:00+00:00 · 采集：2026-09-17T14:06:22.493792+00:00

Yemen’s Houthi rebels had sat out the Iran war, until the militant group saw an opening. Now, it holds one of the world’s most important energy corridors. https://t.co/IkDkXz3Muy

• moondustkk · x:x-following:following:moondustkk:2100413051224588423 · x.com ↗ · 发布：2026-09-17T02:34:29+00:00 · 采集：2026-09-17T14:06:22.493819+00:00

RT @medical_xpress: All 10 healthy dietary patterns examined were linked to slower biological aging, measured through DNA changes—yet no si…

• moondustkk · x:x-following:following:moondustkk:2100411675522334914 · x.com ↗ · 发布：2026-09-17T02:29:01+00:00 · 采集：2026-09-17T14:06:22.493846+00:00

We have a better idea of what the future of the Fed looks like and they are hiking the rates.
https://t.co/blirFvONKD

• moondustkk · x:x-following:following:moondustkk:2100407887747272869 · x.com ↗ · 发布：2026-09-17T02:13:58+00:00 · 采集：2026-09-17T14:06:22.493859+00:00

RT @JTLonsdale: Wow. Got them dead to rights.

A main fundraising apparatus of the left was consistently and knowingly breaking the law, an…

• moondustkk · x:x-following:following:moondustkk:2100405653101560048 · x.com ↗ · 发布：2026-09-17T02:05:06+00:00 · 采集：2026-09-17T14:06:22.493881+00:00

Flights and hotels for weddings are now customary expenses, in addition to gifts. But invitations aren’t obligations.

Here, how to set expectations—or say no—graciously: 🔗 https://t.co/69FtQN9A74 https://t.co/jKt3I4Ab0b

• moondustkk · x:x-following:following:moondustkk:2100404660557226213 · x.com ↗ · 发布：2026-09-17T02:01:09+00:00 · 采集：2026-09-17T14:06:22.493917+00:00

Starship’s 14th test flight is scheduled to reach orbit next week, carrying 26 Starlink satellites as the rocket moves beyond passively safe suborbital trajectories. https://t.co/Znixo2exRj

• moondustkk · x:x-following:following:moondustkk:2100404268003938577 · x.com ↗ · 发布：2026-09-17T01:59:35+00:00 · 采集：2026-09-17T14:06:22.493937+00:00

Cuando Derek Jeter y Andy Pettitte fueron al montículo para sacar a Mariano Rivera por última vez…

Si te gusta el béisbol y no sabes quién es esta leyenda, realmente no te gusta el béisbol… 

Mariano "Mo" Rivera es considerado de manera unánime el mejor cerrador en la historia de las Grandes Ligas (MLB).

Más juegos salvados de por vida con 652 salvamentos. Superó a Trevor Hoffman (601) para convertirse en el líder absoluto de todos los tiempos.

Más juegos finalizados con 952 partidos en los que entró a lanzar para sacar el último out del juego. 

Mejor ERA+ de la historia ya que logró un ERA+ de 205. Esta estadística ajustada indica que su efectividad fue un 105% mejor que el promedio de la liga durante su época (mínimo de 1,000 entradas lanzadas). 

Mejor efectividad de la era moderna ya que su porcentaje de carreras limpias admitidas (ERA) de por vida fue de 2.21, el más bajo para cualquier lanzador que haya iniciado su carrera después de 1919. 

 https://t.co/cgeJOcNRa1

• moondustkk · x:x-following:following:moondustkk:2100404142044844100 · x.com ↗ · 发布：2026-09-17T01:59:05+00:00 · 采集：2026-09-17T14:06:22.493962+00:00

The Pentagon only wants troops who meet waist-to-height requirements and have "impeccable" grooming to attend Defense Secretary Pete Hegseth’s upcoming "State of the Force" speech. https://t.co/AIa97E8cIr https://t.co/IDWS76BFkp

• moondustkk · x:x-following:following:moondustkk:2100393582418911731 · x.com ↗ · 发布：2026-09-17T01:17:08+00:00 · 采集：2026-09-17T14:06:22.493979+00:00

Falcon 9’s first stage lands on the Of Course I Still Love You droneship https://t.co/Rwc9UkZVkp

• moondustkk · x:x-following:following:moondustkk:2100494603354784185 · x.com ↗ · 发布：2026-09-17T07:58:33+00:00 · 采集：2026-09-17T12:16:43.338860+00:00

RT @SpaceX: Falcon 9’s first stage lands on the Of Course I Still Love You droneship https://t.co/Rwc9UkZVkp

• moondustkk · x:x-following:following:moondustkk:2100392984240160839 · x.com ↗ · 发布：2026-09-17T01:14:45+00:00 · 采集：2026-09-17T14:06:22.493994+00:00

💯

• moondustkk · x:x-following:following:moondustkk:2100391523338588268 · x.com ↗ · 发布：2026-09-17T01:08:57+00:00 · 采集：2026-09-17T14:06:22.494032+00:00

Liftoff! https://t.co/okVd5WzMOT

• moondustkk · x:x-following:following:moondustkk:2100390465967137202 · x.com ↗ · 发布：2026-09-17T01:04:45+00:00 · 采集：2026-09-17T14:06:22.494043+00:00

RT @TechXplore_com: A cellulose-derived solvent makes 1–5 mm wood sheets temporarily flexible, then locks in a new shape while preserving m…

• moondustkk · x:x-following:following:moondustkk:2100389042315481360 · x.com ↗ · 发布：2026-09-17T00:59:05+00:00 · 采集：2026-09-17T14:06:22.494070+00:00

People who live through the harrowing experience join a little-known fraternity where members get silk ties and $5,000 watches. https://t.co/NjuOXf0bdX https://t.co/jIDnciN3Xz

• ChatGPT 官方更新 · web:chatgpt-updates:2026-09-14:a775b6c5bfe404e91950f139b00226e22f5fee68414f159e12ddaf16c8b646bc · help.openai.com ↗ · 发布：2026-09-14T00:00:00+00:00 · 采集：2026-09-17T13:11:16.322077+00:00

## Updated permissions for new Health connections

For new Health plugin connections, ChatGPT will now default to each user’s global plugins permissions setting.

If you make no changes to your Plugins settings, the default permission setting is “[Allow low-risk actions](https://help.openai.com/en/articles/11487775-apps-in-chatgpt#permission-options).” This means that once you’ve connected your information in Health, ChatGPT can use it when relevant without asking for permission each time, unless you are attempting a sensitive action, such as asking ChatGPT to email a training plan based on your data to a running partner. This change reflects what we’ve seen from users: more than 70% of Health users already choose to allow ChatGPT to use data connected in Health without asking for permission each time.

You can change your Health permissions anytime in **Settings → Plugins → Health**. [Learn more](https://help.openai.com/en/articles/20001036-health-in-chatgpt#privacy-and-data-controls) about how we keep Health data private and secure.

## Changes to automatic switching to thinking in ChatGPT (Plus and Pro)

We’re retiring automatic switching from **Instant** to **Thinking** (reasoning) for ChatGPT Plus and Pro users globally. You can still select an available option in the model picker to give ChatGPT more time to think or reason.
We’re also removing the **Higher intelligence** setting from ChatGPT on the web for these plans. ChatGPT can still switch automatically for safety purposes.
For more information, see: [GPT-5.6 and GPT-6 Pro in ChatGPT](https://help.openai.com/articles/20001354).

• moondustkk · x:x-following:following:moondustkk:2100388085481578744 · x.com ↗ · 发布：2026-09-17T00:55:17+00:00 · 采集：2026-09-17T14:06:22.494091+00:00

Watch Falcon 9 launch the USSF-259 mission from pad 4E in California  https://t.co/w7OMJQCpxR

• moondustkk · x:x-following:following:moondustkk:2100386414345236662 · x.com ↗ · 发布：2026-09-17T00:48:39+00:00 · 采集：2026-09-17T14:06:22.494137+00:00

My Mom’s book is coming out soon @mayemusk 

https://t.co/oWLuqi4O1K

• moondustkk · x:x-following:following:moondustkk:2100385184642367942 · x.com ↗ · 发布：2026-09-17T00:43:45+00:00 · 采集：2026-09-17T14:06:22.494149+00:00

RT @KatieMiller: On the rise of transgenderism with @GadSaad:

“When I’m a Hollywood superstar, as you said earlier—the stats on having one…

• moondustkk · x:x-following:following:moondustkk:2100384977645093051 · x.com ↗ · 发布：2026-09-17T00:42:56+00:00 · 采集：2026-09-17T14:06:22.494167+00:00

RT @stevenmarkryan: • Flashback: Elon Musk Explaining SpaceX In 2008

So long ago he needed a name tag 😂 https://t.co/bNFQF0fxxR

• moondustkk · x:x-following:following:moondustkk:2100384403495293125 · x.com ↗ · 发布：2026-09-17T00:40:39+00:00 · 采集：2026-09-17T14:06:22.494183+00:00

New episode of @SpaceX documentary

• moondustkk · x:x-following:following:moondustkk:2100383652731585004 · x.com ↗ · 发布：2026-09-17T00:37:40+00:00 · 采集：2026-09-17T14:06:22.494225+00:00

@SciGuySpace (which is still small potatoes compared to airflight tonnage)

• moondustkk · x:x-following:following:moondustkk:2100383440105624052 · x.com ↗ · 发布：2026-09-17T00:36:50+00:00 · 采集：2026-09-17T14:06:22.494214+00:00

@SciGuySpace Starship will deliver somewhere between 1M and 10M tons per year to orbit.

Even if non-SpaceX tonnage to orbit were to triple, it would still be less than 0.1% of what SpaceX plans to do.

• moondustkk · x:x-following:following:moondustkk:2100383340821950635 · x.com ↗ · 发布：2026-09-17T00:36:26+00:00 · 采集：2026-09-17T14:06:22.494249+00:00

How Ed Sheeran, a musician known for his crowd-pleasing, inoffensive everyman appeal, landed in this situation illustrates the hazards of planning a lucrative tour at a time of intense geopolitical unrest and cultural division over violent conflicts. https://t.co/WoQ6yfNxdJ

• moondustkk · x:x-following:following:moondustkk:2100383084709282012 · x.com ↗ · 发布：2026-09-17T00:35:25+00:00 · 采集：2026-09-17T14:06:22.494265+00:00

RT @medical_xpress: Aficamten improved health status and exercise capacity versus placebo in 517 people with symptomatic nonobstructive hyp…

• moondustkk · x:x-following:following:moondustkk:2100381725402570839 · x.com ↗ · 发布：2026-09-17T00:30:01+00:00 · 采集：2026-09-17T14:06:22.494285+00:00

All systems are looking good, and weather is 95% favorable for liftoff. Propellant load is about to begin for Falcon 9’s launch of the USSF-259 mission from California → https://t.co/L2HTgohXdK https://t.co/UxLouHForc

• moondustkk · x:x-following:following:moondustkk:2100377922075050383 · x.com ↗ · 发布：2026-09-17T00:14:54+00:00 · 采集：2026-09-17T14:06:22.494300+00:00

Keep going. https://t.co/Ct064L5uak

• moondustkk · x:x-following:following:moondustkk:2100375477361758389 · x.com ↗ · 发布：2026-09-17T00:05:11+00:00 · 采集：2026-09-17T14:06:22.494320+00:00

MLB teams are using position players on the mound more than ever before, transforming what was once a novelty into a common—and controversial—strategy.

Read more: ⚾️ https://t.co/2VrlEX8oOz https://t.co/pQvheaM4S2

• chatgpt · x:chatgpt:user:chatgpt:2098535078573875646 · x.com ↗ · 发布：2026-09-11T22:12:06+00:00 · 采集：2026-09-17T07:53:30.962545+00:00

RT @adamhfry: This week's ChatGPT feature drop - Sept 11:

1/ We upgraded ChatGPT Images: it's much faster, more precise at edits, and bett…

• chatgpt · x:chatgpt:user:chatgpt:2098533675050672134 · x.com ↗ · 发布：2026-09-11T22:06:31+00:00 · 采集：2026-09-17T07:53:37.794355+00:00

Your pet has a job now.

Pets help you keep track of your chats while you’re away from the desktop app. Now, you can start a new chat right from your pet, too.

Not a pet person? Meet Mini: the same shortcuts and updates, taking up less space on your screen.

Pick your pet—or go Mini—in the ChatGPT desktop app settings.

• chatgpt · x:chatgpt:user:chatgpt:2098519342530920488 · x.com ↗ · 发布：2026-09-11T21:09:34+00:00 · 采集：2026-09-17T07:53:44.185733+00:00

GPT-VI ASTRA https://t.co/qtCb31sUhc

• chatgpt · x:chatgpt:user:chatgpt:2098457925459349913 · x.com ↗ · 发布：2026-09-11T17:05:31+00:00 · 采集：2026-09-17T07:54:28.692562+00:00

A browser-based 3D coloring experience where you can rotate, paint, and personalize models directly on their surfaces.

https://t.co/BHpTLtJWcs https://t.co/nUJbXmb3Cc

• chatgpt · x:chatgpt:user:chatgpt:2098457921676083585 · x.com ↗ · 发布：2026-09-11T17:05:30+00:00 · 采集：2026-09-17T07:54:34.013986+00:00

Here are some ways to get started with Sites if you’re  looking for inspiration:

- Launch dashboards to track and share results

- Personal portfolios to help your work stand out

- Onboarding hubs to help new teammates get up to speed

- Games (like trivia!) for your friend’s birthday party

- Maps of your favorite local spots for out-of-town visitors

- Trip itineraries to get everyone on the same page

• chatgpt · x:chatgpt:user:chatgpt:2098457920291946894 · x.com ↗ · 发布：2026-09-11T17:05:30+00:00 · 采集：2026-09-17T07:54:40.439714+00:00

Three months ago, we launched ChatGPT Sites – an easy way for anyone to build and host fully functional, interactive web apps. Since then, people have created over 5M sites.

We’ve been listening to your feedback and ICYMI, we’ve launched a few Sites updates:

🫂 Build together: Invite teammates to edit, save, and publish to your shared Site

🔐 Share privately: Invite specific people to a private Site without making it public

🚀 Launch faster: Go from prompt to deployment in half the time

👀 Explore your data: Ask ChatGPT to inspect your Site’s database. Editors can view it, too

🌐 Add your own custom domain

• chatgpt · x:chatgpt:user:chatgpt:2098457929091580020 · x.com ↗ · 发布：2026-09-11T17:05:32+00:00 · 采集：2026-09-17T07:54:04.600190+00:00

A theatrical calculator that turns arithmetic into a spectacle of dramatic effects and sound. 

https://t.co/9C0XjWB1kL https://t.co/XWaCs5trGi

• chatgpt · x:chatgpt:user:chatgpt:2098457927195758950 · x.com ↗ · 发布：2026-09-11T17:05:31+00:00 · 采集：2026-09-17T07:53:59.413974+00:00

A live-style change log that brings together earthquakes, space weather, and regional alerts. 
https://t.co/t6xx1qBAjD

• chatgpt · x:chatgpt:user:chatgpt:2098457930807091480 · x.com ↗ · 发布：2026-09-11T17:05:32+00:00 · 采集：2026-09-17T07:53:49.379620+00:00

A shared Japanese word-chain game where each new word continues the previous word’s ending.

https://t.co/pv1XjGyhss

• chatgpt · x:chatgpt:user:chatgpt:2098457923475411241 · x.com ↗ · 发布：2026-09-11T17:05:31+00:00 · 采集：2026-09-17T07:54:23.410304+00:00

We love what you've been building, here a few you've shared with us:

A cozy pixel-art game for exploring small worlds, collecting creatures, and caring for virtual pets. 
https://t.co/r3AC4XpcHk https://t.co/GBLHSeJS73

• ChatGPT 官方更新 · web:chatgpt-updates:2026-09-11:9eafa8396bd8597d5c75ce80853e460d0fa3c8ac6680c512639e384c114fca38 · help.openai.com ↗ · 发布：2026-09-11T00:00:00+00:00 · 采集：2026-09-17T13:11:16.322077+00:00

## Planned custom GPT retirement and migration to plugins

We’re planning to retire custom GPTs across ChatGPT plans and provide a migration path to [plugins](https://help.openai.com/articles/20001256), which can combine reusable instructions with connected apps. You can continue using existing GPTs until the retirement date that applies to you, subject to your permissions.
To prepare, review the GPTs you rely on and identify their creators and anyone who needs access. When migration becomes available, test migrated plugins and review sharing. Timing and availability may vary by plan and workspace; follow applicable notices and [learn more about the planned retirement](https://help.openai.com/articles/20001519).

• chatgpt · x:chatgpt:user:chatgpt:2098141242911785452 · x.com ↗ · 发布：2026-09-10T20:07:08+00:00 · 采集：2026-09-17T07:54:45.265254+00:00

RT @ebloch: Stock watchlist rolling out in ChatGPT + Finances for US Plus/Pro! 

Tell Chat what stocks to follow. It’ll keep an eye on what…

• chatgpt · x:chatgpt:user:chatgpt:2098135108092076424 · x.com ↗ · 发布：2026-09-10T19:42:45+00:00 · 采集：2026-09-17T07:54:50.388166+00:00

RT @thsottiaux: To make sure our current users have an incredible experience and continued access to Astra, we are going to pause subscript…

• chatgpt · x:chatgpt:user:chatgpt:2098127115791294703 · x.com ↗ · 发布：2026-09-10T19:11:00+00:00 · 采集：2026-09-17T07:54:56.348305+00:00

And starting today, @Dropbox, @Box, and Sharepoint are now natively integrated into the ChatGPT Library across both ChatGPT and ChatGPT Work.

Rolling out this week to paid users. https://t.co/qeWFv9PSwG

• chatgpt · x:chatgpt:user:chatgpt:2098065296968011853 · x.com ↗ · 发布：2026-09-10T15:05:21+00:00 · 采集：2026-09-17T07:55:03.220333+00:00

Now everyone can put data to work.

We’re introducing a new Data agent in ChatGPT Work so you can turn your company’s data into answers, interactive dashboards, and action—just by asking.

Just add the Data Plugin in ChatGPT Work, connect to the data sources and context you already use, and start the conversation. https://t.co/EG1nQFZfEU

• Claude 官方更新 · web:claude-updates:2026-09-10:13723a9ebb2f373a10bc588e7733989b5f9a4e8d8b36076735d09a6c7e2d3d70 · support.claude.com ↗ · 发布：2026-09-10T00:00:00+00:00 · 采集：2026-09-17T13:11:20.521154+00:00

**Smart reports (beta)**

We’ve launched smart reports, which analyze how a team uses Claude and report on the work getting done, what it costs, where sessions run into friction, and which repeated patterns are worth packaging as shared skills. Smart reports are available in beta on Claude Enterprise plans. Learn more in **[Get started with smart reports](https://support.claude.com/en/articles/16893491-get-started-with-smart-reports)**.

• DeepSeek 官方更新 · web:deepseek-updates:2026-09-10:99cd890badb2ff58e102cfa6eef1237dca9460a143feedc40e819f58a0ab2919 · api-docs.deepseek.com ↗ · 发布：2026-09-10T00:00:00+00:00 · 采集：2026-09-17T13:11:22.698038+00:00

### DeepSeek-V4.1-Flash Release[​](#deepseek-v41-flash-release "Direct link to DeepSeek-V4.1-Flash Release")

Today, we officially release the DeepSeek-V4.1-Flash model. It is the smallest model in our new architecture family, with native multimodal visual understanding. The new architecture is designed for a higher capability ceiling, faster inference, higher throughput, and scaling to larger models.

-   GPQA Diamond: 90.9
-   HLE: 36.8 (39.1\*)
-   Codeforces (Rating): 3471
-   MathArena Apex: 65.6
-   Terminal-Bench 2.1: 90.6
-   Terminal-Bench 3.0: 30.0
-   Terminal-Bench 4.0: 31.2
-   DeepSWE v1.1: 74.2
-   ProgramBench: 20.3
-   NL2Repo-Bench: 65.4
-   CyberGym: 88.1
-   SEC-Bench Pro: 62.8
-   ExploitGym: 15.3
-   HLE (w/tools): 63.9
-   Automation-Bench: 54.8
-   Agents' Last Exam: 31.8
-   Chartography (w/tools): 78.9
-   BabyVision (w/tools): 89.6
-   ZeroBench-main (w/tools): 49.0

\* Tested only on the pure-text subset of the HLE benchmark set.

**API changes**

DeepSeek V4.1 Flash is now available on the DeepSeek API with native multimodal support. Change the model name to `deepseek-flash` to call the latest V4.1 Flash model. The previous-generation models V4 Flash and V4 Flash Vision Exp have been retired; for compatibility, the model names `deepseek-v4-flash` and `deepseek-v4-flash-vision-exp` are temporarily routed to V4.1 Flash.

In response to user demand, we have decided to continue providing API services for DeepSeek V4 Pro after September 14, 2026, with the billing method remaining unchanged. We will provide further notice should there be any changes. Thank you for your understanding and support!

**API pricing adjustment**

With the release of DeepSeek-V4.1-Flash, API prices have been reduced accordingly. For details, please refer to [Models & Pricing](https://api-docs.deepseek.com/quick_start/pricing).

For more details, please refer to [this documentation](https://api-docs.deepseek.com/news/news260910).

* * *

• chatgpt · x:chatgpt:user:chatgpt:2097791572909687068 · x.com ↗ · 发布：2026-09-09T20:57:40+00:00 · 采集：2026-09-17T07:55:08.450399+00:00

RT @athyuttamre: 🌌 ChatGPT Voice can now use GPT-5.6 Sol and GPT-6 Astra

We’re updating how intelligence works in voice. Now, you can sele…

• chatgpt · x:chatgpt:user:chatgpt:2097776711853109252 · x.com ↗ · 发布：2026-09-09T19:58:37+00:00 · 采集：2026-09-17T07:55:14.074690+00:00

RT @nickbaumann_: Images 2.5 & Astra are a wonderful combination

(sound on) https://t.co/VFK7wqXgH9

• chatgpt · x:chatgpt:user:chatgpt:2097726297996705977 · x.com ↗ · 发布：2026-09-09T16:38:17+00:00 · 采集：2026-09-17T07:55:19.339552+00:00

RT @michpokrass: following up after astra launch week and the navier-stokes announcement i wanted to share a little bit about how we think…

• ChatGPT 官方更新 · web:chatgpt-updates:2026-09-08:b939b9a551e7c72c95c49ace1a6a9630944bc48d3e5f5360eafc071c57972eb0 · help.openai.com ↗ · 发布：2026-09-08T00:00:00+00:00 · 采集：2026-09-17T13:11:16.322077+00:00

## **ChatGPT Images 2.5: new ways to create and edit**

ChatGPT Images 2.5 brings sharper details, more precise editing, and faster image generation. We’re also introducing new ways to create and share:

-   **Start from a template.** Open Images → Templates, choose a template, and customize the details. ChatGPT may ask follow-up questions to help refine your request.

-   **Turn a sketch into an image.** On mobile, type **@** in the message box and select **Sketch**. Draw your idea, then describe the image you want to create.

-   **Edit and comment.** On mobile, open a generated image full-screen to edit it or add comments to tell ChatGPT what you’d like to change..

-   **Share a prompt.** Share the prompt for a generated image so others can create their own version.

Some users will also see animated dots and a Snake game while images are generated. The loading experience varies by platform and rollout; Snake isn’t available in Work mode.

Templates are not yet available in Work mode. Existing image-generation limits are unchanged.

[Learn more about Images in ChatGPT](https://help.openai.com/en/articles/11084440-images-in-chatgpt).

• Claude 官方更新 · web:claude-updates:2026-09-01:16d2c5e3f663a9f5da12a2a4e09b71c7b6e7c9d732f3ac2e100386ed40749190 · support.claude.com ↗ · 发布：2026-09-01T00:00:00+00:00 · 采集：2026-09-17T13:11:20.521154+00:00

**Claude Fable 5.1 and Claude Mythos 5.1 launch**

We just launched Claude Fable 5.1 and Claude Mythos 5.1, the world’s most advanced models for coding and knowledge work. For more information, see our blog post: **[Claude Fable 5.1 and Mythos 5.1](https://www.anthropic.com/claude-fable-and-mythos-5-1)**.

• 豆包 官方更新 · web:doubao-updates:2026-09-01:f0f3b867b8600fc1ba49de83f466ed40b870b2c84d0c14ce3ca6a6fd8fec625f · www.volcengine.com ↗ · 发布：2026-09（仅月份，具体日期未知） · 采集：2026-09-17T13:17:10.321005+00:00

## doubao-seed-evolving · 更新

公告月份：2026-09（官方仅标注月份，具体发布日期未知）。
适用范围：火山方舟模型服务；不代表豆包 App/Web 已开放。
模型类型：深度思考模型

Seed-Evolving 版本更新：Agent 长程任务执行与异步子任务验收更可靠，多模态理解与处理效率提升，深度调研和工具执行的可信度增强，Office 办公跨应用协同及 Coding 工程实践能力进一步提升。版本升级自动生效，通过统一 Model ID doubao-seed-evolving 调用即可获取最新版本的模型服务。模型详情请参考Seed-Evolving。

• 豆包 官方更新 · web:doubao-updates:2026-07-01:2d08841309b2523ae1a86916773553e83557afdf5f4b70cf3c2ee90dafa69a7d · www.volcengine.com ↗ · 发布：2026-07（仅月份，具体日期未知） · 采集：2026-09-17T13:17:10.321005+00:00

## doubao-seed-evolving · 更新

公告月份：2026-07（官方仅标注月份，具体发布日期未知）。
适用范围：火山方舟模型服务；不代表豆包 App/Web 已开放。
模型类型：深度思考模型

Seed-Evolving 版本更新：支持 1M 超长上下文，长程任务稳定性增强；tokens 消耗更少，工具调用更简洁，整体效率提升。版本升级自动生效，通过统一 Model ID doubao-seed-evolving 调用即可获取最新版本的模型服务。调用指南请参考深度思考。

• ChatGPT 官方更新 · web:chatgpt-updates:2026-07-13:a6faf6f4cee4f10f209b9d64bee95ee9e25f05bfc3b08a859b3ef823eb1932b0 · help.openai.com ↗ · 发布：2026-07-13T00:00:00+00:00 · 采集：2026-09-17T13:11:16.322077+00:00

## ChatGPT returns to WhatsApp in the EEA

ChatGPT is available again on WhatsApp in the European Economic Area (EEA). You can get started without a ChatGPT account by messaging the verified 1-800-CHATGPT contact at +1-800-242-8478.
On WhatsApp, you can message ChatGPT, upload images, send voice notes, create images, and use ChatGPT in many languages. Linking your ChatGPT account is optional and provides higher usage limits.
Availability is based on the country code associated with your WhatsApp number and may roll out gradually. Usage limits apply.
ChatGPT is also available on Kakao in South Korea and Viber in supported markets.

• ChatGPT 官方更新 · web:chatgpt-updates:2026-07-08:61f25773c269a98b714d03442b0424c832e33302ca3435833a3e94693b1ae9e6 · help.openai.com ↗ · 发布：2026-07-08T00:00:00+00:00 · 采集：2026-09-17T13:11:16.322077+00:00

## **Introducing GPT-Live-1 in ChatGPT Voice**

We’re rolling out a new ChatGPT Voice experience. ChatGPT Voice is now powered by GPT-Live-1 for paid users and GPT-Live-1 mini for Free users. Both models can listen and speak at the same time, making turn-taking and interruptions feel more natural.
GPT-Live-1 works inside a ChatGPT chat, with spoken responses appearing alongside streamed text. It can use web search and memory, show visual results through supported widgets, and work with text and images in the same conversation.
GPT-Live-1 is rolling out across consumer plans, including Free, on chatgpt.com and the ChatGPT iOS and Android apps in supported regions. It is not available in ChatGPT Business, Enterprise, or Edu workspaces at launch. GPT-Live-1 does not support video or screen sharing at this time; eligible subscribers can continue using those capabilities with Advanced Voice Mode.

• ChatGPT 官方更新 · web:chatgpt-updates:2026-08-25:f0750702b78dc86742ff2a4ed3d35000aadb36f66255ae2180d948d3f795b9e5 · help.openai.com ↗ · 发布：2026-08-25T00:00:00+00:00 · 采集：2026-09-17T13:11:16.322077+00:00

## Scheduled tasks can respond to app updates and be shared

Scheduled tasks in ChatGPT Work can use webhooks to respond when something changes in a supported app, helping you keep work moving without repeatedly checking for updates. You can also share a task so someone else can review and customize it, connect their own apps, and schedule an independent copy.
Supported webhooks include new Gmail messages, Slack channel messages, and GitHub pull request activity.

### Create webhook-triggered tasks

Plus and Pro users can ask Work on web, iOS, and Android to respond when a new Gmail message arrives, a Slack channel receives a message, or a GitHub pull request changes. For example, ask ChatGPT to summarize an email announcing that registration has opened, or draft next steps after a client shares feedback in Slack. Add @ChatGPT to each monitored Slack channel. Actions that require approval pause until you review them.

### Share scheduled tasks

Free, Go, Plus, and Pro users can share scheduled tasks. Recipients can review and customize the instructions and, when applicable, the schedule, connect their own apps, and create an independent copy. Running a shared webhook-triggered task requires access to Work and the recipient’s own connected app.

### Scheduled tasks for Free users

Free users can create up to three active scheduled tasks. Tasks can run once or recur no more than once per day using flexible scheduling windows. Free and Go users cannot create webhook-triggered tasks.
For more information, see: [Scheduled tasks in ChatGPT](https://help.openai.com/articles/10291617).

## **ChatGPT Work can now complete tasks on signed-in websites**

ChatGPT Work’s browser on web and mobile can also help you get tasks done on some websites that require you to sign in—so you can hand off a task and step away while it keeps working.

To get started, just ask ChatGPT Work to complete your task. If a website allows authentication, ChatGPT will surface the login screen so you can securely enter your credentials. You may also need to enter a security code. Once you’re signed in, ChatGPT can continue working. Your session may remain signed in for future tasks, so you don’t have to log in again every time. ChatGPT Work’s browser supports the use of password managers for simple, safe login. ChatGPT cannot see your username or password, and they are never seen by the model or used in model training. ChatGPT does not store your username or passwords. You can delete your browsing history from all sites or one site individually at any time from Settings > Cloud browser > Browser data.

You can now ask ChatGPT Work to do things like:

-   Figure out utilities for a new apartment and sign up for the right plan

-   Book a DMV appointment or fill out your passport renewal forms

-   Check X-ray and bloodwork costs through your insurance portal

-   Find profiles of people that fit your job description and are open to work

You stay in control of which websites ChatGPT Work can access, and it will always ask for confirmation before consequential actions, such as completing a reservation or payment.

Authentication in ChatGPT Work’s browser is available for Plus and Pro plans.

• ChatGPT 官方更新 · web:chatgpt-updates:2026-07-14:62bba95d8b85f77236277618d759dc6a82557aee5830d31b5d1f80205e0eb8ea · help.openai.com ↗ · 发布：2026-07-14T00:00:00+00:00 · 采集：2026-09-17T13:11:16.322077+00:00

## Search across chats, projects, and files is now supercharged.

We’ve rolled out a new way to search your chats, projects, images, and documents in ChatGPT on web, iOS, and Android. You can quickly find past chats, projects, images, and documents from one place.
Start searching from the ChatGPT sidebar to find what you need across everything, or use filters to narrow results by content type. Select a result to open the chat, project, or file directly in ChatGPT.
Available on all ChatGPT plans globally.

• Claude 官方更新 · web:claude-updates:2026-08-06:df3ea5025c9dcbe97b72e0156416797b1362a08a95a81b0d5006ee8799edb260 · support.claude.com ↗ · 发布：2026-08-06T00:00:00+00:00 · 采集：2026-09-17T13:11:20.521154+00:00

**Skill and plugin security scanning (beta)**

Enterprise plans can now turn on skill and plugin security scanning to automatically check third-party skills and plugins for malicious content when someone uploads or edits them. For more information, see **[Get started with skill and plugin scanning](https://support.claude.com/en/articles/15927065)**.

* * *

• ChatGPT 官方更新 · web:chatgpt-updates:2026-06-26:715afbd3946c97aa42098655dae5f71383ed3d88998f80d31ac91459600f322b · help.openai.com ↗ · 发布：2026-06-26T00:00:00+00:00 · 采集：2026-09-17T13:11:16.322077+00:00

## **Personal finance in ChatGPT expands to Plus users and Android**

The personal finance experience in ChatGPT is rolling out to Plus users in the U.S. on web and iOS. It is also now available on Android for Pro and Plus users in the U.S.
Eligible users can securely connect supported financial accounts, view a dashboard of their finances, and ask ChatGPT questions grounded in their financial context

## **Improved dictation in ChatGPT**

We’re rolling out a new speech-to-text model for dictation in ChatGPT across all plans. This is a behind-the-scenes update, so there are no changes to how you use dictation.
The update improves transcription accuracy across languages and accents, including when multilingual speakers mix or switch between languages. Evaluations showed gains in Japanese, Korean, Chinese, Urdu, Vietnamese, accented English, and long-form Spanish, French, Italian, and Portuguese.
Dictation is also more robust in noisy public spaces and workplace environments, with quiet or whispered speech, and when transcribing combinations of letters and numbers.
In our ChatGPT dictation evaluations, word error rate—a standard measure of transcription accuracy—was at least 10% lower for top languages tested than with the previous production model.

## **Retiring GPT-4.5**

As of June 26, 2026, GPT-4.5 is no longer available in ChatGPT, including for custom GPTs. Existing conversations that used GPT-4.5 can continue with GPT-5.5.
This retirement was previously announced on May 28, 2026. It applies to ChatGPT only; there are no changes to the API.

• ChatGPT 官方更新 · web:chatgpt-updates:2026-06-24:c4ba0a4c885658ec6a4799c4e8c53b31f6fe305f26b1ead68a8841474dad5d44 · help.openai.com ↗ · 发布：2026-06-24T00:00:00+00:00 · 采集：2026-09-17T13:11:16.322077+00:00

## **GPT-5.5 Instant Update**

We’re updating GPT-5.5 Instant, our most-used model in ChatGPT, to improve conversational quality—especially when users are making a decision, asking for advice, planning something, researching options, or shopping.
The model is better at identifying the underlying goal behind a question and carrying context across multiple turns. It also follows complex instructions more reliably: when a request includes several constraints or requirements, responses are more likely to address all of them and clearly explain why a recommendation is a good fit.
When users add constraints, clarify what they mean, or push back on an answer, the model should adapt more effectively instead of repeating its original approach.
Shopping and local business queries should also feel more useful. The model makes better use of location context to surface relevant nearby options and can bring together product recommendations, business information, and images more coherently when those elements are helpful.
Overall, responses should feel more tasteful and cohesive. Formatting is less templated and more intentionally designed, with greater creativity alongside better restraint.

• Claude 官方更新 · web:claude-updates:2026-07-07:3e39b70bb486decba8568a615e830a627ea770490bcc4968108adecb6bcb382b · support.claude.com ↗ · 发布：2026-07-07T00:00:00+00:00 · 采集：2026-09-17T13:11:20.521154+00:00

**Claude Cowork on web and mobile**

Claude Cowork is now available on web and mobile in addition to desktop. We are rolling this capability out over the next several weeks starting with the Max plan, with more plans to follow. Cowork runs your sessions remotely (in beta), so your sessions and files are saved to your Claude account and go where you go, on any device. Work continues when you close your laptop, and scheduled tasks run with no device online. Chat and Cowork also share one home now, with one place for your projects and artifacts across both. For more information, see **[Use Claude Cowork on web, desktop, and mobile](https://support.claude.com/en/articles/15520349)**.

**Write tools for the Microsoft 365 connector**

The Microsoft 365 connector now goes beyond search. With write tools enabled, Claude can draft, send, and organize email, manage calendar events, update mailbox settings, and create and update files in OneDrive and SharePoint. Read and search tools work as before, and Teams remains read-only. Before members can use write tools, a Microsoft Entra administrator needs to consent to the updated permission set and an admin needs to enable them for the organization. For more information, see **[Set up the Microsoft 365 connector](https://support.claude.com/en/articles/12542951-set-up-the-microsoft-365-connector#h_a51d877afd)** and **[Connect to Microsoft 365](https://support.claude.com/en/articles/15183774)**.

• ChatGPT 官方更新 · web:chatgpt-updates:2026-08-10:d7a3e9a5ad35f868a4dc4c8e83ed4efda71a8a6e1982df24db834e8c5f69bb28 · help.openai.com ↗ · 发布：2026-08-10T00:00:00+00:00 · 采集：2026-09-17T13:11:16.322077+00:00

## **Restaurant reservations in ChatGPT**

ChatGPT can now help you find available restaurant reservations with OpenTable, Resy, and Yelp.
Ask in the chat and include where and when you’d like to eat, your party size, and any preferences, such as cuisine, budget, dietary needs, or atmosphere. You’ll see available reservation times directly in the response. You can also ask follow-up questions to narrow your options or check a specific restaurant. When you find a time that works, select it to book it.
Rolling out across all ChatGPT plans on mobile, web, and desktop. Reservations are available globally with OpenTable and in the US with Resy and in the US and Canada with Yelp. ChatGPT Work does not include restaurant reservation search.

• Claude 官方更新 · web:claude-updates:2026-06-30:0ee1c7f95972bcd5ef7c92c7531bff1c23b578d04038fb18f3a26df63e8cedb6 · support.claude.com ↗ · 发布：2026-06-30T00:00:00+00:00 · 采集：2026-09-17T13:11:20.521154+00:00

**Claude Sonnet 5 launch**

We launched Claude Sonnet 5, our most agentic Sonnet model yet, with substantial improvements over Sonnet 4.6 in reasoning, tool use, coding, and knowledge work. For more information, see our blog post: **[Introducing Claude Sonnet 5](https://www.anthropic.com/news/claude-sonnet-5)**.

• ChatGPT 官方更新 · web:chatgpt-updates:2026-06-22:d4e7279de514d3a0f40d495a32ff37680b7e1801917cdbee2bc32779ce9071b5 · help.openai.com ↗ · 发布：2026-06-22T00:00:00+00:00 · 采集：2026-09-17T13:11:16.322077+00:00

## **Large pastes are now handled as attachments for more plans**

Starting today for **ChatGPT Free and Go** users, long pastes are handled differently in ChatGPT.
If you paste **more than 10k characters** into the composer, ChatGPT will automatically convert the content into an **attachment** instead of inserting it directly into the text field. This keeps the composer cleaner and helps prevent large pastes from consuming your entire context window.
You can still move the content back into the message at any time by clicking on **Show in text field** to convert the attachment back to a direct paste.
This is already available to **ChatGPT Plus, Pro,** and **Business** users. Previously if you sent more than 5k characters it was sent as an attachment. The threshold has now also been raised to 10k characters.

• Claude 官方更新 · web:claude-updates:2026-07-09:658e13c800c3a4b578f37bb6906da0d9fb499a3510f8e879ca2ccb1f3137d7e7 · support.claude.com ↗ · 发布：2026-07-09T00:00:00+00:00 · 采集：2026-09-17T13:11:20.521154+00:00

**A new way to reflect with Claude**

Your monthly recap is a new feature at **Settings > Reflect** that shows you the topics you spent time on, your most active day and peak hour, and observations about how you work with Claude. It's in beta on Free, Pro, and Max plans on the web and Claude Desktop, and requires memory to be on. Alongside it, **Settings > Time and focus** lets you set optional break reminders and quiet hours. For more information, refer to **[See your monthly recap](https://support.claude.com/en/articles/15672559)** and **[Set break reminders and quiet hours](https://support.claude.com/en/articles/15672868)**.

• ChatGPT 官方更新 · web:chatgpt-updates:2026-07-15:39e410bd70f532f22a02c0b420b18c39bfc527dfd47b9af8dbcb0f701c966241 · help.openai.com ↗ · 发布：2026-07-15T00:00:00+00:00 · 采集：2026-09-17T13:11:16.322077+00:00

## **Increased custom instructions limit**

We’re increasing the character limit for custom instructions in ChatGPT. Plus, Pro, Enterprise, Business, and Education users can now save up to 5,000 characters, up from 1,500, giving them more room to customize ChatGPT’s response style and behavior.

• Claude 官方更新 · web:claude-updates:2026-07-01:fe00bb85ace03b6dd11e2b132df1a3a1ef533162a7ddc5d5ec00f5672825b69e · support.claude.com ↗ · 发布：2026-07-01T00:00:00+00:00 · 采集：2026-09-17T13:11:20.521154+00:00

**Access to Claude Fable 5 and Claude Mythos 5 restored**
We are restoring access to Claude Fable 5 and Claude Mythos 5. See **[our statement](https://www.anthropic.com/news/redeploying-fable-5)** for more information.

**Model entitlements for Enterprise plans (beta)**

Admins on Enterprise plans can now control what models their users have access to, and what effort level settings they can use. For more information, see **[Manage model access for your organization](https://support.claude.com/en/articles/15694740)**.

* * *

• ChatGPT 官方更新 · web:chatgpt-updates:2026-08-28:d25eaf0ecfe153bac0943fcd7fa928ebd3ba4f2fd7b93edeace705ce68a5e7cf · help.openai.com ↗ · 发布：2026-08-28T00:00:00+00:00 · 采集：2026-09-17T13:11:16.322077+00:00

## **Connect multiple Google accounts to ChatGPT**

You can now connect multiple accounts for Gmail, Google Calendar, and Google Contacts plugins in ChatGPT, bringing your personal and work accounts into the same conversation. Ask ChatGPT to check your calendars together when planning your day or find emails across your connected inboxes.

Available globally on supported Plus, Pro, Business, and Enterprise plans across web, desktop, iOS, and Android.

• ChatGPT 官方更新 · web:chatgpt-updates:2026-08-27:a8a66ed74163341a337cbe7162ff3ad87a46b78880cd33b43240a3c205bd6a5d · help.openai.com ↗ · 发布：2026-08-27T00:00:00+00:00 · 采集：2026-09-17T13:11:16.322077+00:00

## More controls in temporary chat

We’re rolling out new controls in temporary chat. You can now choose to personalize responses in a temporary chat with memory, plugins, and custom instructions from your regular chat settings. You can also choose to save a temporary chat.

### How it works

-   **Personalized**: When you create a temporary chat, you can choose to have personalized responses.  Personalized temporary chats can use your memories, custom instructions, and plugins to tailor responses. They do not create new memories and remain out of your chat history unless you save them. Personalization can be set only when you start a temporary chat and cannot be changed after the conversation begins.

-   **Non-personalized:** By default, temporary chats are not personalized. Non-personalized temporary chats do not use memory, custom instructions, or plugins, and they do not create new memories.

-   **Saving a temporary chat:** When a conversation becomes useful for the future, you can add it to your chat history. Saving converts it into a regular chat, so it then follows your account-level personalization settings and model-improvement preferences. For example, if you have memory enabled and save a temporary chat, this will become part of your chat history and can be referenced to personalized future responses.

• DeepSeek 官方更新 · web:deepseek-updates:2026-08-21:aacf1ba217be2a965e6c4fb5320cf6aadb6797fdb33cfaf097eccbd3469ebfca · api-docs.deepseek.com ↗ · 发布：2026-08-21T00:00:00+00:00 · 采集：2026-09-17T13:11:22.698038+00:00

### DeepSeek-V4-Flash-Vision-Exp Release[​](#deepseek-v4-flash-vision-exp-release "Direct link to DeepSeek-V4-Flash-Vision-Exp Release")

Today, the new multimodal vision understanding model DeepSeek-V4-Flash-Vision-Exp is now available on the DeepSeek API platform. This is an experimental model that can be accessed by setting `model='deepseek-v4-flash-vision-exp'`.

-   Terminal Bench 2.1: 83.9
-   NL2Repo: 57.7
-   DeepSWE: 59.3
-   DSBench-Hard: 63.6
-   AutomationBench (Public): 25.7
-   ApexBench (Pass@1): 36.5
-   Agents' Last Exam: 27.3
-   Chartography: 64.3
-   ZeroBench (Pass@5): 35.0

\* For the Code Agent text tasks in the public benchmark sets, the DeepSeek family models were tested using the DeepSeek Harness minimal mode as the framework, with the max effort level, topp=0.95, and temperature=1.0; in the ApexBench and Agents' Last Exam evaluations, the text model DeepSeek-V4-Flash ignores the multimodal elements within them.

In terms of pure-text capabilities (agent, reasoning, world knowledge, etc.), DeepSeek-V4-Flash-Vision-Exp is on par with the official DeepSeek-V4-Flash.

On agent benchmarks that require visual understanding, DeepSeek-V4-Flash-Vision-Exp delivers a significant leap over DeepSeek-V4-Flash, bringing its multimodal agent capabilities close to Opus-4.8.

For usage details, please refer to the [Vision guide](https://api-docs.deepseek.com/guides/vision).

* * *

• ChatGPT 官方更新 · web:chatgpt-updates:2026-07-31:6233cf9ae39b2b1b1ce77e19b3adf1d4d88974f829643b9070ae1202a6bde762 · help.openai.com ↗ · 发布：2026-07-31T00:00:00+00:00 · 采集：2026-09-17T13:11:16.322077+00:00

## **Retiring the DALL·E GPT**

On August 30, 2026, we’re retiring the official DALL·E GPT in ChatGPT. We encourage you to download any images you want to keep before then.
To continue creating or editing images, use ChatGPT Images. User-created GPTs with image generation enabled are not affected.
Learn more in [Images in ChatGPT](https://help.openai.com/en/articles/11084440-images-in-chatgpt).

• Claude 官方更新 · web:claude-updates:2026-06-25:b1eb4f8dd5c08a3ed85cf15dde10d8e2ead479089d1ce878e1ba5cf32ccdee85 · support.claude.com ↗ · 发布：2026-06-25T00:00:00+00:00 · 采集：2026-09-17T13:11:20.521154+00:00

**Enable Trusted Devices for Remote Control**

Admins for Team and Enterprise plans can now require members to verify their device before viewing or steering local Claude Code sessions remotely. For more information, see **[Trusted Devices](https://code.claude.com/docs/en/remote-control#trusted-devices)**.

• 豆包 官方更新 · web:doubao-updates:2026-08-01:19f1d1d50a38f0560774372beb351a7bea13974afd9ef9b6669e8c91a86223b5 · www.volcengine.com ↗ · 发布：2026-08（仅月份，具体日期未知） · 采集：2026-09-17T13:17:10.321005+00:00

## doubao-seed-evolving · 更新

公告月份：2026-08（官方仅标注月份，具体发布日期未知）。
适用范围：火山方舟模型服务；不代表豆包 App/Web 已开放。
模型类型：深度思考模型

Seed-Evolving 版本更新：通用综合能力全面提升，复杂代码任务、长程 Agent 执行表现优化，搜索与工具调用可靠性增强；显著改善搜索幻觉、工具调用幻觉、状态幻觉等问题，更好支撑真实开发、深度调研与自动化工作流等复杂场景。版本升级自动生效，通过统一 Model ID doubao-seed-evolving 调用即可获取最新版本的模型服务。调用指南请参考深度思考。

• 豆包 官方更新 · web:doubao-updates:2026-08-01:6d1f9710066ca6739f553d0f14743c82971abffcffced4a6f11de890f81f8b5a · www.volcengine.com ↗ · 发布：2026-08（仅月份，具体日期未知） · 采集：2026-09-17T13:17:10.321005+00:00

## doubao-seed-evolving · 更新

公告月份：2026-08（官方仅标注月份，具体发布日期未知）。
适用范围：火山方舟模型服务；不代表豆包 App/Web 已开放。
模型类型：深度思考模型

Seed-Evolving 版本更新：进一步强化从需求理解、信息检索到工具执行和结果输出的完整 Agent 任务链路；通用 Agent 与 Coding、深度搜索与真实性、工具调用与复杂任务、视觉理解与长程规划能力持续增强，在真实开发、深度调研、流程自动化与多模态任务中表现更稳定，输出内容更真实可靠。版本升级自动生效，通过统一 Model ID doubao-seed-evolving 调用即可获取最新版本的模型服务。调用指南请参考深度思考。

• Claude 官方更新 · web:claude-updates:2026-07-10:d2602282888bfa7095a96d29f8d03b2d345c3d7f4a47f94b5b0a527b18c545e7 · support.claude.com ↗ · 发布：2026-07-10T00:00:00+00:00 · 采集：2026-09-17T13:11:20.521154+00:00

**Updated memory for Claude**

Memory on Claude now works as a set of individual, categorized entries that Claude reads and updates during your conversations, replacing the previous daily memory summary. See **[Use Claude’s chat search and memory to build on previous context](https://support.claude.com/en/articles/11817273-use-claude-s-chat-search-and-memory-to-build-on-previous-context)**.

• Claude 官方更新 · web:claude-updates:2026-06-23:ffe926d6b51391ea34527f9cc4b6f7a33a0f60bd8078a9ee1c4320b0b926e4f4 · support.claude.com ↗ · 发布：2026-06-23T00:00:00+00:00 · 采集：2026-09-17T13:11:20.521154+00:00

**Delegate work to Claude in Slack with Claude Tag**

Team and Enterprise plans can now tag Claude directly in Slack conversations and delegate tasks to it while they focus on other work. For more information, see our blog post: **[Introducing Claude Tag](https://www.anthropic.com/news/introducing-claude-tag)**.

• Claude 官方更新 · web:claude-updates:2026-07-24:57f65090dac2145428fa6a349486e101dc98d219708dca8706945be3befb099d · support.claude.com ↗ · 发布：2026-07-24T00:00:00+00:00 · 采集：2026-09-17T13:11:20.521154+00:00

**Claude Opus 5 launch**

We launched Claude Opus 5, a thoughtful and proactive model that comes close to the frontier intelligence of Claude Fable 5 at half the price. For more information, see our blog post: **[Claude Opus 5](https://www.anthropic.com/news/claude-opus-5)**.

• Claude 官方更新 · web:claude-updates:2026-07-14:f632d3a140c71a322b061a5d7fcbf8535e89e25dccb732dd6cf2d1c114d87113 · support.claude.com ↗ · 发布：2026-07-14T00:00:00+00:00 · 采集：2026-09-17T13:11:20.521154+00:00

**HIPAA configuration for your Claude organizations is now self-serve**

You can now manage HIPAA readiness for your Claude organizations yourself. This applies to both Claude Enterprise and the Claude Platform (API). In each product, an eligible admin can review the Business Associate Agreement (BAA), download the implementation guide, and enable the HIPAA configuration in a single flow. For more information, refer to **[HIPAA-ready Enterprise plans](https://support.claude.com/en/articles/13296973-hipaa-ready-enterprise-plans)** and **[HIPAA readiness for Claude API](https://platform.claude.com/docs/en/manage-claude/api-and-data-retention#hipaa-readiness)**.

• ChatGPT 官方更新 · web:chatgpt-updates:2026-08-07:8f61c633ae28dab6c55e6adcc42509a0e2445d35e14f1f77864d88bbd87ee08e · help.openai.com ↗ · 发布：2026-08-07T00:00:00+00:00 · 采集：2026-09-17T13:11:16.322077+00:00

## **ChatGPT app experience updates**

We’ve made it easier to bring documents and saved files into your conversations, find what you need, and pick up where you left off.

### Writing, files, and search

-   **Keep your formatting when you paste.** On the web, text copied from Google Docs or another ChatGPT conversation keeps its headings, bold text, links, and lists, both while you’re writing and after you send your message.

-   **Add saved files without uploading them again.** Open the composer menu on the web, select **Add from library**, and choose a file you’ve already saved.

-   **Select several Library items at once.** On the web, drag across multiple files or folders to select them together.

-   **Find folders, files, and past conversations faster.** Search now includes folders and matches conversation titles across web, iOS, and Android. On the web, you can also search within Library to find your saved files.

-   **Get answers that stay closer to your documents.** ChatGPT makes better use of information in attached files when you ask it to summarize a document, find specific details, review material, or draft from a source.

-   **Use an updated composer when signed out.** The signed-out experience on the web now includes a cleaner, unified composer menu.

### **iOS and Android**

-   **Safeguards for background Voice conversations**. On mobile, ChatGPT can recognize when a background Voice conversation no longer seems active, check whether you’re still there, and end the session if you don’t respond. You can turn off Background conversations at any time in **Settings → Voice**.

### **iOS**

-   **Dictate with fewer interruptions.** If dictation fails because of a connection issue or timeout, ChatGPT automatically retries once, helping you finish recording without starting over.

-   **Pick up completed answers without the replay.** When you return to a conversation, completed answers now appear without slowly replaying as though they’re still being generated.

### **Android**

-   **Start Voice conversations faster.** Improved connection times make it quicker to begin a Voice conversation with ChatGPT.

### **ChatGPT Work on the web**

-   **See your Work usage at a glance.** Eligible users on personal plans and ChatGPT Business can see their remaining Work usage in the sidebar. Available credit options depend on your account and workspace permissions.

## Files and Projects in ChatGPT Voice

GPT-Live in ChatGPT Voice now supports file uploads and Projects. You can now upload files in a voice conversation and analyze its contents or ask questions. You can also use voice in Projects, referencing recent project chats, sources, and project instructions.
[Learn more about ChatGPT Voice](https://help.openai.com/en/articles/20001274).

• moondustkk · x:x-following:following:moondustkk:2100259401730834936 · x.com ↗ · 发布：2026-09-16T16:23:56+00:00 · 采集：2026-09-17T14:06:22.494203+00:00

Europe's response to Falcon 9 exiting the market should not be to shovel contracts to the dead-end, solid rocket technology Avio is flying. Rather it needs to invest heavily in rapid reuse, because what's coming down the line is even more of a steamroller than Falcon 9. https://t.co/3aQO4F6AAh

{
  "event": "Swift 6.4将优化调试模块追踪并缩减构建产物体积",
  "facts": [
    "该内容为Swift官方博客2026年9月11日发布的调试能力优化说明",
    "显式模块追踪相关改动始于Swift 6.3，将被纳入即将推出的Swift 6.4版本",
    "使用SwiftPM、Xcode的开发者无需修改项目配置即可自动获得优化收益",
    "自行维护Bazel、Buck、CMake等自定义构建系统的开发者可能需要调整适配",
    "Swift 6.4起，Linux/Windows平台可移除-modulewrap，Darwin平台可移除-add_ast_path参数",
    "Swift 6.4起dSYM包将不再存储二进制Swift模块，体积会大幅缩小",
    "Swift 6.4将优化LLDB桥接头文件导入速度，提升调试效率",
    "移除dSYM内二进制Swift模块不影响本地新构建程序的变量查看等常规调试操作"
  ],
  "new_information": [
    "本次披露Swift 6.3已落地显式模块依赖追踪、调试信息携带模块路径等基础改动",
    "本次预告Swift 6.4将推出旧构建参数废弃、dSYM瘦身、桥接头导入提速等优化"
  ],
  "importance": 90,
  "personal_relevance": "Swift及苹果平台开发者可关注本次更新的调试提效效果与构建系统适配要求",
  "worth_monitoring": true
}

按冻结快照重建的来源文本，包含实际字符截断；不代表请求已成功送达。系统指令不在此展示，也没有逐句引用标注。

rss:a76d8c3e94eba61fc46b5cc05fc51e5f6df1d0f46d901d32e9c27bb251f155ae:swift-blog:19
发布时间：2026-09-11T15:30:00+00:00
When your Swift program hits a breakpoint and stops so you can inspect it, the debugger’s expression evaluator has to find the exact Swift module your code was built from. Until now, that lookup wasn’t always precise. The upcoming Swift 6.4 release will include changes, begun in Swift 6.3, that address this by updating how the Swift compiler references explicitly-built Swift modules in debug info. The majority of developers will automatically benefit from faster, more reliable debugging and smaller build products, without any modifications to their SwiftPM or Xcode projects. For developers who maintain their own build systems using, for example, Bazel, Buck, or CMake, some adjustments may be necessary to take advantage of these changes. This article explains how the debugger uses Swift modules. Next, it describes how Swift 6.3 changes the way modules are tracked in debug info to solve several problems with the previous representation. Finally, it shows how to adjust build systems to take advantage of the new representation and eliminate some build steps that are no longer necessary. Swift modules and expression evaluation LLDB’s standout feature is its powerful expression evaluator. Because LLDB embeds the Clang and Swift compilers, it can JIT-compile any valid source code and run it in the context of your application while stopped at a breakpoint. This includes not just calling code in your application, but also defining new data types, functions, and closures. Debugging features that are usually reserved for interpreted or JIT-compiled languages like JavaScript become available to ahead-of-time-compiled languages like C++ and, of course, Swift! In order to JIT-compile user expressions that make use of data types defined in the debugged program, LLDB’s embedded Swift compiler needs to import the Swift modules defining those types. In a world before explicitly-built modules, LLDB would find the base name of the main module at the current breakpoint in the debug info and then kick off an implicit import of a module with that name. With a cold module cache this would launch an expensive compilation of that module and all its dependencies. To illustrate this, let’s walk through a simple example: (lldb) p myObj Here myObj is just a local variable: LLDB can find its location in the debug info and resolve its type via reflection metadata. No need to bother the Swift compiler. Let’s make it more complex: (lldb) p myObj.myComputedProperty In this case, myComputedProperty is really a function call; in order to evaluate this, LLDB needs the expression evaluator to run code in the target. In order to initialize a Swift compiler instance with the state of the current module, LLDB finds the name of the current function’s Swift module in debug info. We can visualize what LLDB does using the dwarfdump utility: $ dwarfdump Foo.o ... DW_TAG_module DW_AT_name ("Foo") Conceptually, LLDB then wraps the expression in a function that can be compiled: (lldb) log enable lldb expr (lldb) p myObj.myComputedProperty ... import Foo func lldb_expr(_ $__lldb_arg : UnsafeMutablePointer<Any>) { let myObj: MyObject = /* some LLDB magic */ // Expression begins here: myObj.myComputedProperty ... One problem with this is that import Foo is quite imprecise: Even though the Swift language doesn’t allow multiple modules to have the same name, even the most stringently engineered application may have more than one copy of the same module. For example, there might be a private version of a module containing all of its private declarations (which would be great for LLDB) and also a Swift interface file that only contains the public interface for the module. Or there might be macOS and Mac Catalyst variants of the same module in the same process. Swift modules, debug info, and the build system Let’s look at where those modules are found next. In order to communicate the location of Foo.swiftmodule to LLDB, Swift build systems rely on some cooperation from the linker. On Darwin the system linker accepts an option called -add_ast_path and build systems are expected to specify this option to list every binary Swift module when linking. # Linker invocation on macOS ld -add_ast_path /path/to/Foo.swiftmodule Foo.o -o MyApplication The linker translates these options into symbol table entries. The debug info linker dsymutil then collects all Swift modules and stores them in a special __swift_ast section in the dSYM bundle, where LLDB can find them by name. Alternatively, when debugging without dSYM bundles, LLDB reads the symbol table entries in the binary to collect a list of all binary Swift modules. Such an approach would not work on platforms where the linker isn’t aware of Swift. For these platforms, which include Windows, Linux, and FreeBSD, the Swift compiler provides a -modulewrap action that takes a binary Swift module and outputs an object file with a .swift_ast section holding the contents of the module. This object file can then be passed to any linker to get added to the binary, where LLDB can find it. # Modulewrap and linker invocation on Linux swift-frontend -modulewrap Foo.swiftmodule -o Foo.swiftmodule.o lld Foo.o Foo.swiftmodule.o -o MyApplication This can create scalability issues, especially for large applications: Module files can get large and for an entire application you can often end up with a large portion of the SDK in the resulting binary. That can be quite problematic for the binary size. As mentioned above, the chances of LLDB finding the right module in a Swift AST section or symbol table just by its base name diminish as the application gets more complex. Binary Swift modules are version-locked to the precise compiler that created them. This is at odds with the intent of dSYM bundles, which are meant for long-term archival serialization of debug info. If a matching explicit module cannot be found, LLDB falls back to an implicit module import which may involve recompiling parts of the SDK from source. This can be very slow. Precise module tracking To evaluate expressions, the debugger needs to be able to find and import Swift modules. Until now, this relied either on special linker support or additional compilation steps, with a high cost for binary size. On top of that the debugger was imprecisely locating Swift modules by name. Starting in Swift 6.3 and continuing since, we have been making changes to the Swift compiler, the Swift driver, and LLDB that improve performance, reliability, and scalability. These changes are built on top of explicitly-built modules. What’s new Explicitly-built modules track their explicit Swift dependencies: Explicitly-built binary Swift modules have always kept track of their explicitly-built Clang module dependencies. This is why LLDB can import explicit modules so much faster than implicit modules, which may need to recompile their dependencies from source. In Swift 6.3, explicitly-built binary Swift modules also keep track of their Swift module dependencies. This makes importing an explicitly-built module fast and unambiguous because no module needs to be looked up by name. This happens automatically. Users don’t need to make any changes. Users with distributed build systems will already be familiar with the Swift frontend’s path remapping options, which now also affect Swift module paths. Debug info stores path of object file’s own Swift module: Once LLDB finds the top-level module it can precisely import it and all of its dependencies. But how can LLDB find precisely the module that belongs to the Swift file at the current breakpoint? In Swift 6.3, the Swift compiler can store the path to it in the debug info. Because a Swift file’s own Swift module is not an input to an object file compilation, there is a new -debug-module-path compiler option to communicate the path to each object file compilation action. This path is also subject to the standard path remapping options used by users with distributed build systems. Swift driver passes module path to compile jobs: Users of swiftpm or Xcode do not need to think about this, because the Swift driver also knows about the new -debug-module-path option and automatically passes the path to the object file’s own Swift module to the compiler. However, users maintaining their own third-party build system to orchestrate Swift compilations with explicitly-built modules that are calling the Swift frontend directly and bypassing the Swift driver need to make sure to communicate the path to the top-level module to each object file compilation job. What’s deprecated Beginning in Swift 6.4, you can safely make the following changes. swiftc -modulewrap and ld -add_ast_path: Because the module paths are now communicated via debug info and the module headers themselves, third-party build systems doing explicit module builds can now remove all -modulewrap actions on Linux and Windows; and remove the use of the -add_ast_path linker option on Darwin (macOS, iOS, etc…). Binary Swift modules in dSYM bundles: As a consequence, dsymutil will no longer process binary Swift modules. This is a good thing, because binary Swift modules—which can only be parsed by the exact toolchain that produced them—were always at odds with dSYM bundles being a long-term archival format. Moreover, Swift modules often depend on Clang modules, and these Clang modules also were never included in dSYM bundles. By removing the binary Swift modules, dSYM bundles will get smaller. But don’t we need them for debugging? Since Swift 1.0, binary Swift modules were included in dSYM bundles because they were needed to resolve the types of local variables. However, starting with Swift 5.6, LLDB could perform this operation by reading the reflection metadata in the binary. The absence of binary Swift modules in dSYM bundles does not affect LLDB’s ability to inspect the contents of variables or dump object descriptions with po. Binary Swift modules are still needed to evaluate complex expressions like function calls or computed getters. Expression evaluation continues to work as long as LLDB finds all binary modules in their original (or remapped) location. This is always the case when debugging a just-built binary on the same machine. If the absence of binary Swift modules in dSYM bundles creates an unforeseen problem with your workflow, please let us know, either on the Swift LLDB forum or by creating an issue on the bug tracker. When compiling with caching enabled, all paths pointing to Swift modules and module debug info are content-addressable storage references, identified by content rather than file location, so everything described here also works transparently with compilation caching. Coming in Swift 6.4: Faster bridging header import in LLDB Beyond more reliable path tracking, Swift 6.4 will also speed up importing bridging headers, a step common enough across Swift projects that most developers will feel the difference. Up to and including Swift 6.3, LLDB always compiles a bridging header from source, a step that can add noticeable time to debugging sessions that use one. In recent nightly development toolchains, LLDB can use the new precise explicit module information to import precompiled bridging headers and their explicit module dependencies directly. This makes debugging explicitly-built projects with bridging headers as fast and reliable as debugging fully modularized projects. Summary With these changes for explicitly-built modules: Binaries built with debug info on Windows and Linux, and dSYM bundles on Darwin will get dramatically smaller, since they no longer contain any binary Swift modules (6.4+) Contextual module imports in LLDB become more reliable due to precise tracking instead of by-name lookups Certain performance cliffs around module importing in LLDB are eliminated (such as SDK module dependencies in dSYMs triggering imp

兴趣配置

{
  "interests": {
    "very_high": [
      "AI coding agents",
      "Codex",
      "SwiftUI",
      "Swift",
      "Apple developer platform",
      "developer tools"
    ],
    "high": [
      "AI",
      "independent development",
      "productivity",
      "cognitive science"
    ],
    "low": [
      "generic AI hype",
      "cryptocurrency",
      "celebrity news"
    ]
  }
}

• Swift.org · rss:a76d8c3e94eba61fc46b5cc05fc51e5f6df1d0f46d901d32e9c27bb251f155ae:swift-blog:19 · swift.org ↗ · 发布：2026-09-11T15:30:00+00:00 · 采集：2026-09-17T07:07:56.208180+00:00

When your Swift program hits a breakpoint and stops so you can inspect it, the debugger’s expression evaluator has to find the exact Swift module your code was built from. Until now, that lookup wasn’t always precise. The upcoming Swift 6.4 release will include changes, begun in Swift 6.3, that address this by updating how the Swift compiler references explicitly-built Swift modules in debug info. The majority of developers will automatically benefit from faster, more reliable debugging and smaller build products, without any modifications to their SwiftPM or Xcode projects. For developers who maintain their own build systems using, for example, Bazel, Buck, or CMake, some adjustments may be necessary to take advantage of these changes. This article explains how the debugger uses Swift modules. Next, it describes how Swift 6.3 changes the way modules are tracked in debug info to solve several problems with the previous representation. Finally, it shows how to adjust build systems to take advantage of the new representation and eliminate some build steps that are no longer necessary. Swift modules and expression evaluation LLDB’s standout feature is its powerful expression evaluator. Because LLDB embeds the Clang and Swift compilers, it can JIT-compile any valid source code and run it in the context of your application while stopped at a breakpoint. This includes not just calling code in your application, but also defining new data types, functions, and closures. Debugging features that are usually reserved for interpreted or JIT-compiled languages like JavaScript become available to ahead-of-time-compiled languages like C++ and, of course, Swift! In order to JIT-compile user expressions that make use of data types defined in the debugged program, LLDB’s embedded Swift compiler needs to import the Swift modules defining those types. In a world before explicitly-built modules, LLDB would find the base name of the main module at the current breakpoint in the debug info and then kick off an implicit import of a module with that name. With a cold module cache this would launch an expensive compilation of that module and all its dependencies. To illustrate this, let’s walk through a simple example: (lldb) p myObj Here myObj is just a local variable: LLDB can find its location in the debug info and resolve its type via reflection metadata. No need to bother the Swift compiler. Let’s make it more complex: (lldb) p myObj.myComputedProperty In this case, myComputedProperty is really a function call; in order to evaluate this, LLDB needs the expression evaluator to run code in the target. In order to initialize a Swift compiler instance with the state of the current module, LLDB finds the name of the current function’s Swift module in debug info. We can visualize what LLDB does using the dwarfdump utility: $ dwarfdump Foo.o ... DW_TAG_module DW_AT_name ("Foo") Conceptually, LLDB then wraps the expression in a function that can be compiled: (lldb) log enable lldb expr (lldb) p myObj.myComputedProperty ... import Foo func lldb_expr(_ $__lldb_arg : UnsafeMutablePointer<Any>) { let myObj: MyObject = /* some LLDB magic */ // Expression begins here: myObj.myComputedProperty ... One problem with this is that import Foo is quite imprecise: Even though the Swift language doesn’t allow multiple modules to have the same name, even the most stringently engineered application may have more than one copy of the same module. For example, there might be a private version of a module containing all of its private declarations (which would be great for LLDB) and also a Swift interface file that only contains the public interface for the module. Or there might be macOS and Mac Catalyst variants of the same module in the same process. Swift modules, debug info, and the build system Let’s look at where those modules are found next. In order to communicate the location of Foo.swiftmodule to LLDB, Swift build systems rely on some cooperation from the linker. On Darwin the system linker accepts an option called -add_ast_path and build systems are expected to specify this option to list every binary Swift module when linking. # Linker invocation on macOS ld -add_ast_path /path/to/Foo.swiftmodule Foo.o -o MyApplication The linker translates these options into symbol table entries. The debug info linker dsymutil then collects all Swift modules and stores them in a special __swift_ast section in the dSYM bundle, where LLDB can find them by name. Alternatively, when debugging without dSYM bundles, LLDB reads the symbol table entries in the binary to collect a list of all binary Swift modules. Such an approach would not work on platforms where the linker isn’t aware of Swift. For these platforms, which include Windows, Linux, and FreeBSD, the Swift compiler provides a -modulewrap action that takes a binary Swift module and outputs an object file with a .swift_ast section holding the contents of the module. This object file can then be passed to any linker to get added to the binary, where LLDB can find it. # Modulewrap and linker invocation on Linux swift-frontend -modulewrap Foo.swiftmodule -o Foo.swiftmodule.o lld Foo.o Foo.swiftmodule.o -o MyApplication This can create scalability issues, especially for large applications: Module files can get large and for an entire application you can often end up with a large portion of the SDK in the resulting binary. That can be quite problematic for the binary size. As mentioned above, the chances of LLDB finding the right module in a Swift AST section or symbol table just by its base name diminish as the application gets more complex. Binary Swift modules are version-locked to the precise compiler that created them. This is at odds with the intent of dSYM bundles, which are meant for long-term archival serialization of debug info. If a matching explicit module cannot be found, LLDB falls back to an implicit module import which may involve recompiling parts of the SDK from source. This can be very slow. Precise module tracking To evaluate expressions, the debugger needs to be able to find and import Swift modules. Until now, this relied either on special linker support or additional compilation steps, with a high cost for binary size. On top of that the debugger was imprecisely locating Swift modules by name. Starting in Swift 6.3 and continuing since, we have been making changes to the Swift compiler, the Swift driver, and LLDB that improve performance, reliability, and scalability. These changes are built on top of explicitly-built modules. What’s new Explicitly-built modules track their explicit Swift dependencies: Explicitly-built binary Swift modules have always kept track of their explicitly-built Clang module dependencies. This is why LLDB can import explicit modules so much faster than implicit modules, which may need to recompile their dependencies from source. In Swift 6.3, explicitly-built binary Swift modules also keep track of their Swift module dependencies. This makes importing an explicitly-built module fast and unambiguous because no module needs to be looked up by name. This happens automatically. Users don’t need to make any changes. Users with distributed build systems will already be familiar with the Swift frontend’s path remapping options, which now also affect Swift module paths. Debug info stores path of object file’s own Swift module: Once LLDB finds the top-level module it can precisely import it and all of its dependencies. But how can LLDB find precisely the module that belongs to the Swift file at the current breakpoint? In Swift 6.3, the Swift compiler can store the path to it in the debug info. Because a Swift file’s own Swift module is not an input to an object file compilation, there is a new -debug-module-path compiler option to communicate the path to each object file compilation action. This path is also subject to the standard path remapping options used by users with distributed build systems. Swift driver passes module path to compile jobs: Users of swiftpm or Xcode do not need to think about this, because the Swift driver also knows about the new -debug-module-path option and automatically passes the path to the object file’s own Swift module to the compiler. However, users maintaining their own third-party build system to orchestrate Swift compilations with explicitly-built modules that are calling the Swift frontend directly and bypassing the Swift driver need to make sure to communicate the path to the top-level module to each object file compilation job. What’s deprecated Beginning in Swift 6.4, you can safely make the following changes. swiftc -modulewrap and ld -add_ast_path: Because the module paths are now communicated via debug info and the module headers themselves, third-party build systems doing explicit module builds can now remove all -modulewrap actions on Linux and Windows; and remove the use of the -add_ast_path linker option on Darwin (macOS, iOS, etc…). Binary Swift modules in dSYM bundles: As a consequence, dsymutil will no longer process binary Swift modules. This is a good thing, because binary Swift modules—which can only be parsed by the exact toolchain that produced them—were always at odds with dSYM bundles being a long-term archival format. Moreover, Swift modules often depend on Clang modules, and these Clang modules also were never included in dSYM bundles. By removing the binary Swift modules, dSYM bundles will get smaller. But don’t we need them for debugging? Since Swift 1.0, binary Swift modules were included in dSYM bundles because they were needed to resolve the types of local variables. However, starting with Swift 5.6, LLDB could perform this operation by reading the reflection metadata in the binary. The absence of binary Swift modules in dSYM bundles does not affect LLDB’s ability to inspect the contents of variables or dump object descriptions with po. Binary Swift modules are still needed to evaluate complex expressions like function calls or computed getters. Expression evaluation continues to work as long as LLDB finds all binary modules in their original (or remapped) location. This is always the case when debugging a just-built binary on the same machine. If the absence of binary Swift modules in dSYM bundles creates an unforeseen problem with your workflow, please let us know, either on the Swift LLDB forum or by creating an issue on the bug tracker. When compiling with caching enabled, all paths pointing to Swift modules and module debug info are content-addressable storage references, identified by content rather than file location, so everything described here also works transparently with compilation caching. Coming in Swift 6.4: Faster bridging header import in LLDB Beyond more reliable path tracking, Swift 6.4 will also speed up importing bridging headers, a step common enough across Swift projects that most developers will feel the difference. Up to and including Swift 6.3, LLDB always compiles a bridging header from source, a step that can add noticeable time to debugging sessions that use one. In recent nightly development toolchains, LLDB can use the new precise explicit module information to import precompiled bridging headers and their explicit module dependencies directly. This makes debugging explicitly-built projects with bridging headers as fast and reliable as debugging fully modularized projects. Summary With these changes for explicitly-built modules: Binaries built with debug info on Windows and Linux, and dSYM bundles on Darwin will get dramatically smaller, since they no longer contain any binary Swift modules (6.4+) Contextual module imports in LLDB become more reliable due to precise tracking instead of by-name lookups Certain performance cliffs around module importing in LLDB are eliminated (such as SDK module dependencies in dSYMs triggering implicit imports) Developers maintaining their own build systems can remove support for -modulewrap actions and remove -add_ast_path from the linker flags, but may need to pass -debug-module-path to the compiler if they are not letting the Swift driver handle the frontend options Finally, static archives were easy to overlook: projects that didn’t use -add_ast_path when linking them often had confusing debugging issues inside those archives as a result. This entire class of issues has been designed away. tl;dr: -modulewrap and -add_ast_path are replaced by -debug-module-path. Debug info gets smaller and more precise.

{
  "event": "DeepSeek-V4-Flash API启动公开测试",
  "facts": [
    "该消息2026年7月31日发布，DeepSeek-V4-Flash API官方版本进入公测，调用方式不变，设置模型名为`deepseek-v4-flash`即可使用",
    "官方称其Agent能力显著增强，多项基准测试成绩远超V4-Pro-Preview版本",
    "公开基准中的代码Agent任务采用即将发布的DeepSeek Harness极简模式，配合最高力度、topp=0.95、temperature=1.0参数测试",
    "本次公布的DSBench-FullStack、DSBench-Hard分别为DeepSeek内部全栈开发、编码Agent难题测试集",
    "该版本原生支持Responses API格式，已专门适配Codex，配置方法可查阅官方文档",
    "DeepSeek-V4-Flash-0731与此前Preview版模型架构、大小一致，仅经过重新后训练",
    "本次更新仅涉及DeepSeek-V4-Flash API，V4-Pro API、DeepSeek APP/WEB端模型均无变化，V4-Pro官方版本即将推出"
  ],
  "new_information": [
    "原生支持Responses API格式，已针对Codex完成专门适配，相关接入配置可参考官方文档",
    "本次仅更新对应API产品，APP/网页端模型、V4-Pro API均无调整，V4-Pro正式版即将发布"
  ],
  "importance": 90,
  "personal_relevance": "你开发AI编码代理或使用Codex时，可评估接入该公测API验证编码任务表现。",
  "worth_monitoring": true
}

按冻结快照重建的来源文本，包含实际字符截断；不代表请求已成功送达。系统指令不在此展示，也没有逐句引用标注。

web:deepseek-updates:2026-07-31:43e75c17729d1992505709366337182aa8003529ba6ddfdbc8dde4e2b57433ea
发布时间：2026-07-31T00:00:00+00:00
### DeepSeek-V4-Flash Update[​](#deepseek-v4-flash-update "Direct link to DeepSeek-V4-Flash Update")

The official release of the DeepSeek-V4-Flash API is now in public beta. The API calling method remains unchanged — simply set the model name to `deepseek-v4-flash` to use the latest version.

**Significantly enhanced agent capabilities, with benchmark results far exceeding V4-Pro-Preview:**

-   Terminal Bench 2.1: 82.7
-   NL2Repo: 54.2
-   Cybergym: 76.7
-   DeepSWE: 54.4
-   Toolathlon verified: 70.3
-   Agent Last Exam: 25.2
-   Automation Bench (Public): 25.1
-   DSBench-FullStack: 68.7
-   DSBench-Hard: 59.6

Note 1: For the Code Agent tasks in the public benchmark sets, the official DeepSeek-V4-Flash was tested using the DeepSeek Harness minimal mode (to be released soon) as the framework, with the max effort level, topp=0.95, and temperature=1.0
Note 2: DSBench-FullStack is an internal full-stack development test set, and DSBench-Hard is an internal Coding Agent hard-problem test set

**The official V4-Flash natively supports the Responses API format and is specifically adapted for Codex. For the specific configuration, please refer to the [documentation](https://api-docs.deepseek.com/quick_start/agent_integrations/codex).**
**DeepSeek-V4-Flash-0731 keeps the same model architecture and size as DeepSeek-V4-Flash-Preview, and was only re-post-trained.**
**Note: This update only upgrades the DeepSeek-V4-Flash API. The DeepSeek-V4-Pro API and the APP/WEB models are unchanged.**
**The official release of DeepSeek-V4-Pro will follow soon.**

* * *

兴趣配置

{
  "interests": {
    "very_high": [
      "AI coding agents",
      "Codex",
      "SwiftUI",
      "Swift",
      "Apple developer platform",
      "developer tools"
    ],
    "high": [
      "AI",
      "independent development",
      "productivity",
      "cognitive science"
    ],
    "low": [
      "generic AI hype",
      "cryptocurrency",
      "celebrity news"
    ]
  }
}

• DeepSeek 官方更新 · web:deepseek-updates:2026-07-31:43e75c17729d1992505709366337182aa8003529ba6ddfdbc8dde4e2b57433ea · api-docs.deepseek.com ↗ · 发布：2026-07-31T00:00:00+00:00 · 采集：2026-09-17T13:11:22.698038+00:00

### DeepSeek-V4-Flash Update[​](#deepseek-v4-flash-update "Direct link to DeepSeek-V4-Flash Update")

The official release of the DeepSeek-V4-Flash API is now in public beta. The API calling method remains unchanged — simply set the model name to `deepseek-v4-flash` to use the latest version.

**Significantly enhanced agent capabilities, with benchmark results far exceeding V4-Pro-Preview:**

-   Terminal Bench 2.1: 82.7
-   NL2Repo: 54.2
-   Cybergym: 76.7
-   DeepSWE: 54.4
-   Toolathlon verified: 70.3
-   Agent Last Exam: 25.2
-   Automation Bench (Public): 25.1
-   DSBench-FullStack: 68.7
-   DSBench-Hard: 59.6

Note 1: For the Code Agent tasks in the public benchmark sets, the official DeepSeek-V4-Flash was tested using the DeepSeek Harness minimal mode (to be released soon) as the framework, with the max effort level, topp=0.95, and temperature=1.0
Note 2: DSBench-FullStack is an internal full-stack development test set, and DSBench-Hard is an internal Coding Agent hard-problem test set

**The official V4-Flash natively supports the Responses API format and is specifically adapted for Codex. For the specific configuration, please refer to the [documentation](https://api-docs.deepseek.com/quick_start/agent_integrations/codex).**
**DeepSeek-V4-Flash-0731 keeps the same model architecture and size as DeepSeek-V4-Flash-Preview, and was only re-post-trained.**
**Note: This update only upgrades the DeepSeek-V4-Flash API. The DeepSeek-V4-Pro API and the APP/WEB models are unchanged.**
**The official release of DeepSeek-V4-Pro will follow soon.**

* * *

{
  "event": "火山方舟发布豆包深度思考模型260915",
  "facts": [
    "官方公告月份为2026年9月，具体发布日期未知。",
    "模型名为 doubao-seed-2-1-pro-260915，类型为深度思考模型。",
    "适用范围为火山方舟模型服务，不代表豆包 App/Web 已开放。",
    "官方标称支持1M超长上下文，可用于Agent长程任务、工具执行、Coding工程实践等场景。",
    "来源未说明价格、额度、地区、版本限制及实际效果。"
  ],
  "new_information": [
    "火山方舟推出新深度思考模型 doubao-seed-2-1-pro-260915，官方公告时间为2026年9月。",
    "该模型标称1M上下文，面向长程Agent、工具执行和Coding；不能据此认为App/Web已开放。"
  ],
  "importance": 72,
  "personal_relevance": "对AI coding agents和开发者工具相关，可关注其API实际代码与工具调用表现。",
  "worth_monitoring": true
}

按冻结快照重建的来源文本，包含实际字符截断；不代表请求已成功送达。系统指令不在此展示，也没有逐句引用标注。

web:doubao-updates:2026-09-01:0c10f5ca6b4d0136587509c4451cf8f529d6d7b1c67ea7089c77790c6aab1c45
发布时间：2026-09（仅月份，具体日期未知）
## doubao-seed-2-1-pro-260915 · 新发布

公告月份：2026-09（官方仅标注月份，具体发布日期未知）。
适用范围：火山方舟模型服务；不代表豆包 App/Web 已开放。
模型类型：深度思考模型

支持 1M 超长上下文，可用于 Agent 长程任务执行、异步子任务验收、多模态理解、深度调研、工具执行、Office 办公跨应用协同及 Coding 工程实践等场景。调用指南请参考深度思考。

兴趣配置

{
  "interests": {
    "very_high": [
      "AI coding agents",
      "Codex",
      "SwiftUI",
      "Swift",
      "Apple developer platform",
      "developer tools"
    ],
    "high": [
      "AI",
      "independent development",
      "productivity",
      "cognitive science"
    ],
    "low": [
      "generic AI hype",
      "cryptocurrency",
      "celebrity news"
    ]
  }
}

• 豆包 官方更新 · web:doubao-updates:2026-09-01:0c10f5ca6b4d0136587509c4451cf8f529d6d7b1c67ea7089c77790c6aab1c45 · www.volcengine.com ↗ · 发布：2026-09（仅月份，具体日期未知） · 采集：2026-09-17T13:17:10.321005+00:00

## doubao-seed-2-1-pro-260915 · 新发布

公告月份：2026-09（官方仅标注月份，具体发布日期未知）。
适用范围：火山方舟模型服务；不代表豆包 App/Web 已开放。
模型类型：深度思考模型

支持 1M 超长上下文，可用于 Agent 长程任务执行、异步子任务验收、多模态理解、深度调研、工具执行、Office 办公跨应用协同及 Coding 工程实践等场景。调用指南请参考深度思考。