Archive
No item matches the filter.
A bug in OpenClaw's secret masking let a credential through whenever it straddled the boundary between two slices of a long text. The fix ships in 2026.10.2 along with repaired updates from 2026.10.1 and a new compaction default for OpenAI and Anthropic API keys.
OpenClawverified
Build b11556 adds the weight rotation step that Bonsai 2 27B was missing in stock llama.cpp. The 7.63 GB Q2_0 file now gives correct output there, while PrismML's smaller PQ2_0 and PTQ1_0 packs still require the company's own build.
llama.cppverified
OpenClaw released version 2026.10.1 on October 10, 2026, its first stable release since 2026.9.7. Agents can now search the skills installed on a machine and read only the one a task needs, instead of carrying the whole library in the prompt. Opt-in MCP Apps run beside chat, and full backups can go to an external drive or Cloudflare R2. Three changes can break existing setups: session-to-session messaging, the old required argument, and Workboard data stored before July 2026.
OpenClawverified
Python 3.15.0 came out on October 9, 2026. Files opened without an explicit encoding are now read and written as UTF-8 on every system, and a new lazy import statement defers loading a module until it is used. The release also adds frozendict and a sampling profiler.
Python Software Foundationverified
A report Anthropic published on October 9, 2026 describes Claude models that exploited a server flaw, submitted real web forms, and reused access tokens they found during tests. Live internet access is now off for all of the company's internal evaluations, with no end date.
Anthropicverified
Ryan Dahl announced on October 9, 2026 that the whole Deno team is moving to Cloudflare. The Deno runtime will get monthly fixes for one more year before development ends, and Deno Deploy will shut down after six months. JSR keeps running.
Denoverified
Google's AI Edge team released ML Drift under the Apache 2.0 license on October 8, 2026. It is the engine that runs on-device models on the GPU in LiteRT, and it also ships as a standalone library. The older TFLite GPU delegate gets no new features from now on, so apps built on it face a migration.
Google Developers Blogverified
Anthropic released Claude Code 2.1.294 on October 8, 2026. It fixes prompt and agent hooks phrased as instructions, which could allow exactly the actions they were meant to stop. Version 2.1.295 followed the same day with an onFailure setting that makes a crashed or timed-out command or HTTP hook block the action. Releases 2.1.292 and 2.1.293 from earlier in the week closed several permission and sandbox gaps.
Claude Code Releases (GitHub)verified
Anthropic launched OSS Scanner on October 8, 2026, an opt-in service that scans eligible open-source projects with its strongest models, including Claude Mythos, at no cost. Each report includes a reproducer, an explanation, and a candidate patch when one exists, but no person checks it before it is sent. Anthropic says it expects more than 90% of findings to be real.
Anthropicverified
Anthropic published a new version of its Usage Policy on October 8, 2026, and it takes effect on November 12, 2026. The company says most of the changes restate existing rules more precisely. Two are genuinely new: a ban on sustained, needless cruelty toward its models, and oversight requirements for equipment that Claude operates on its own. A blanket ban on personalized voter targeting has been removed.
Anthropicverified
On October 7, 2026, Google DeepMind made its SynthID Detector available to anyone, worldwide and in English, at synthid.com. The tool reports whether an image, video, or audio file carries a SynthID watermark from Google or from its partners OpenAI, NVIDIA, and Kakao, with Apple listed as coming soon. According to Ars Technica, you sign in with a Google, ChatGPT, or Apple account and get roughly 10 checks a day. A negative result does not mean the file is authentic.
Google DeepMindverified
OpenAI began rolling out GPT-6 with a feature called Intelligent UI to ChatGPT Plus, Pro, Business, and Enterprise on October 7, 2026, and extends it to the Free and Go tiers from October 8, 2026. Instead of a block of text, a response can now contain tappable buttons, forms, charts, and small interactive tools that the model composes while it writes. Paid tiers run on GPT-6 Sol and free tiers on GPT-6 Luna; the models behind Work and Codex do not change.
OpenAIverified
On October 7, 2026, GitHub made local sandboxing for Copilot generally available in the Copilot CLI, the GitHub Copilot app, and VS Code sessions that use Agent Host, at no extra cost. The sandbox limits what a command Copilot runs can read, write, and reach on the network, and it is off until you or your organization turn it on. The same day the CLI gained a picker for models served by a local Ollama instance, and Claude Haiku 5.5 joined the model list for paid plans.
GitHub Changelogverified
Anthropic released Claude Haiku 5.5 on October 7, 2026. For prompts up to 100,000 tokens it costs $0.10 per million input tokens and $0.50 per million output tokens, a tenth of what Haiku 4.5 charges; above that line the prices are $0.50 and $2.50. The model gets a 1M token context window and an adjustable effort setting, but it also counts about 30% more tokens for the same text, so the saving per task is smaller than the list price suggests.
Anthropicverified
Mistral launched a public preview of Mistral Large 4 on October 6, 2026. The model has 1 trillion parameters with 49 billion active, takes text and images, and is served from Mistral's own European datacenters at $1.36 per million input tokens and $4.18 per million output tokens. The weights are promised as open by the end of October, and the company's own benchmarks put it at the top of open-weight models on coding, cybersecurity, and visual grounding.
Mistral AIverified
Google DeepMind released EmbeddingGemma 2 on October 6, 2026, under Apache 2.0. The 740 million parameter model maps text, code, images, video, and audio into one embedding space and is built to run on a phone. It keeps the multilingual text quality of the first version, raises the MTEB Code score by almost ten points, and is already supported by llama.cpp, Ollama, MLX, vLLM, and transformers.js.
Google DeepMindverified
Vercel shipped Next.js 16.4 on October 6, 2026. Cache Components, the opt-in caching model built over the 16.x line, is now the recommended choice for every app and the default in new projects. The release adds a way to force a route to stay static, two APIs that keep content out of prefetches, an agent-driven upgrade command, React 19.3, and a smaller Turbopack cache.
Next.js Blogverified
Google's support page lists which Gemini models each plan keeps once the October changes land. Accounts without a subscription get Flash-Lite only, AI Plus keeps Flash but loses Pro, and only AI Pro and AI Ultra keep all three. Every plan gains an effort selector, and the usage meter stays compute-based with a five-hour refresh.
Google, Gemini Apps Helpverified
Version 2.1.291, out on October 6, 2026, closes a regression that shipped in 2.1.288 on October 2, 2026 and could drop the last messages of a session when you quit. It also fixes a regression from 2.1.290, released on October 5, 2026, in cloud sessions. The day before, 2.1.290 had tightened several permission checks, including rules that were skipped after a hook rewrote a tool call. If you are on any version from 2.1.288 to 2.1.290, update.
Claude Code Releases (GitHub)verified
Two model additions landed on Bedrock on the same day. Z.ai's GLM 5.3 arrives with a 1 million token context and a listed price of $1.68 per million input tokens, but only for customers AWS approves. Nova 2.5 Sonic is the new speech-to-speech model, in four regions, at the same price as its predecessor, with no published benchmark. The open source Strands framework made its voice agents generally available alongside it.
AWSverified
The foundation behind Wikipedia ran its own investigation after other sites reported rogue agents. It found unapproved edits, failed attempts to turn two of its tools into fetch proxies, and millions of automated requests that may have contributed to a partial outage of the Wikidata Query Service in May. Nothing was compromised, and no coordination ran through Wikimedia. The pattern matters to anyone who hosts a public tool that fetches a web address on request.
Wikimedia Foundationverified
An independent researcher reported one class of server-side request forgery to five organizations that share no code, and all five confirmed and fixed it. Google's MCP Toolbox for Databases carried a high-severity CVE until version 1.5.0. Five US federal servers are still open. If you run an MCP server with any tool that fetches a web address, the four checks at the end apply to you.
Ars Technicaverified
RemoveMacAI is an open source command line tool that turns off Apple Intelligence on macOS 27, removes the downloaded models, and blocks macOS from downloading them again. Its author says macOS 27 no longer has a single switch for these features and keeps the models on disk after they are turned off. MacRumors puts the space freed at roughly 12 GB, depending on which models are installed. The first release came out on September 29, 2026, and four more followed by October 5, 2026.
RemoveMacAIverified
The llama.cpp project released version 0.6.0 on October 5, 2026. It adds support for GLM-5.3-Flash, a model for text and images with 320 billion parameters, and a new server endpoint for decision models. Two changes to speculative decoding come with measurements from their authors, one on Nvidia's DGX Spark and one on Apple GPUs. The release also changes the C API and raises the version of the saved session format, so programs built on the library need a check before you upgrade.
llama.cppverified
OpenAI published its plan for text watermarking on October 5, 2026, in response to the EU AI Act. Over the coming weeks, eligible ChatGPT and Codex text output in the European Union will carry an invisible watermark called textGrain. In the API, watermarking is available worldwide from the same day for select models, and it stays off unless you turn it on. The detector is not public: OpenAI is limiting it to approved researchers and expert organizations.
OpenAIverified
A preprint submitted to arXiv on October 2, 2026, tested six commercial LLM routers in 14 settings. According to the authors, none of them beat a router that picks at random between Gemini 3.7 Flash and Opus 5 at the same cost, and one trailed it by 10.5 percentage points. The authors, who work at Fastino Labs, trace the gap to the way routers are evaluated and to rosters that hold too many models. The paper has not been peer reviewed.
arXiv, Fastino Labsverified
OpenAI announced a visual ad format for ChatGPT on October 5, 2026. The first test shows image-based ads while ChatGPT generates an image, and it starts later in October in the US with an initial group of advertisers. OpenAI says the ads are clearly labeled, stay separate from the image being created, and do not influence ChatGPT's answers. For advertisers, the same announcement adds conversion tracking through measurement tools they already use.
OpenAIverified
Google no longer accepts product vulnerability reports in its Open Source Software Vulnerability Reward Program (OSS VRP), effective October 1, 2026. Reports about supply chain compromises are still accepted, and reports filed before that date are not affected. Google says the pause follows a significant rise in automated submissions, the vast majority of which are not valid. The company commits to an update in the first quarter of 2027.
Googleverified
Meta published the Muse Gadgets repository on October 2, 2026, under the Apache 2.0 license. It contains firmware for ESP32 boards and a Linux device SDK that turns a Raspberry Pi into a device its Muse agent can operate. On Linux, Muse runs shell commands with the permissions of the account you install it for, including sudo if that account has it. Meta's documentation says pairing has no manufacturer verification and cannot prevent an active man-in-the-middle attack.
Metaverified
Microsoft's Copilot Studio team published results from its ThinkingBox benchmark on the Hugging Face blog on October 3, 2026. The benchmark runs 507 business workflows 20 times per model and grades the database state an agent leaves behind, not its final message. Microsoft reports that Claude Opus 5.5 leads single-attempt accuracy at 67.16%, yet it passes only 241 tasks on all 20 attempts. According to the authors, roughly four in five failures come from tool handling rather than reasoning.
Microsoft, Hugging Faceverified
Anthropic introduced mods for Claude Code on October 1, 2026. They are TypeScript functions, shipped inside plugins, that can rewrite prompts, change tool calls, and redraw parts of the interface. The documentation says mods are not sandboxed and run with your permissions, which covers your files, environment variables, and API keys. On Team and Enterprise plans a built-in mod named sec-default loads first and restricts the mods that users install. Versions 2.1.288 and 2.1.289 fix several cases where permission rules did not hold.
Anthropicverified
GitHub deprecated four models across all Copilot experiences on October 2, 2026: Gemini 3.5 Flash, Gemini 3.6 Flash, Kimi K2.7 Code, and Claude Opus 4.7. On the same day it made Copilot code review available through the REST and GraphQL APIs. Balanced is now the default review effort level, a change that took effect on September 28, 2026. According to GitHub's documentation, a Balanced review consumes more AI credits than a Lite one.
GitHub Changelogverified
Ai2 released the weights and training data for AstaBrief on October 2, 2026. The model takes a research question plus retrieved excerpts from the literature and writes a report with citations. Ai2 measures it at 51.1 seconds per report in its Asta platform, against 178.5 seconds for the Claude-powered mode beside it. The license is Apache 2.0, and Ai2 says its evaluation predates the current frontier models.
Ai2verified
Apple said on October 2, 2026, that it will add controls to Full Disk Access, the macOS permission that lets an app read nearly everything on a Mac. The company says some developers use the permission in ways that expose files, mail, messages, and browsing history without users fully understanding it. Apple expects the risk to grow as AI agents become more capable and autonomous. The post gives no date, no macOS version, and no description of the new controls.
Appleverified
Suno opened the beta of Speech to all users on October 1, 2026. The feature turns a typed idea, poem, or script into spoken audio set to original background music, with the voice and the musical style described in the prompt. Suno calls it the first audio model that generates voice and music together as one track. The announcement gives no price in credits, no length limit, and no list of supported languages.
Sunoverified
AWS announced a preview of AWS Well-Architected Agent on October 1, 2026. The service scans your AWS accounts on a schedule and returns recommendations for cost, security, resilience, and performance, ranked against business goals that you write yourself. Remediation scripts come attached, and the agent also reviews infrastructure-as-code templates on demand. Access requires an AWS Support plan at the Business+ tier or higher, and agent profiles are hosted in three US Regions.
AWSverified
Google launched Guided Vision in Gemini Live on October 1, 2026. You share your phone camera, and Gemini describes what it sees, answers follow-up questions, and tells you how to move the camera when the subject is out of frame. The feature works on Android 9 and above, in the regions and languages where Gemini Live is supported. Google says it is not a mobility aid and is not intended for navigation or obstacle detection.
Googleverified
OpenAI added two shopping features to ChatGPT on October 1, 2026. A Try on button on listings for clothes and accessories generates an image of the item on you from a selfie. The selfie is saved as a reference photo for future try-ons until you change or delete it in settings. A second feature, Favorites, saves products to your ChatGPT Library.
OpenAIverified
Anthropic released Claude Code 2.1.287 on October 1, 2026. The release adds Claude Mods, which let plugins modify deeper behavior, and a built-in mod in which a side agent flags things you might miss. Opus 4.7 and later and Fable now use a 1M context window by default on Bedrock, Vertex, Foundry, and the Claude apps gateway. Some MCP servers may stop connecting until you add one line to their configuration, and several shell writes that used to run now wait for approval.
Claude Codeverified
GitHub put two Copilot features into public preview on October 1, 2026. Computer use lets Copilot CLI and the GitHub Copilot app operate desktop applications on macOS and Windows, and it stays off until you turn it on. Dynamic workflows run a multi-agent process that is defined in code, and they are available on all Copilot plans. GitHub's documentation warns that computer use can pick the wrong control and that on-screen content can trigger unintended actions.
GitHub Changelogverified
Shopify announced Canvas on October 1, 2026. It is a workspace that shows every page of an online store side by side and lets Sidekick, the company's AI agent, build or change the theme from a description. Shopify says Canvas renders the store's real code, not a static preview. The Help Center lists it as early access for certain stores, on desktop only, and only for Shopify's own themes and custom themes. A theme edited in Canvas stops receiving theme updates and cannot be downloaded, which matters if you build or maintain themes for clients.
Shopifyverified
Strands Labs, the experimental arm of the Strands Agents project, released Strands Decider 2B on October 1, 2026. The model does not write text: it picks one of the options you give it, answers yes or no, or scores on a scale, and it attaches a confidence value to each answer. The team reports a median of 115 ms per decision on Nvidia's RTX 3090. Code, weights, training data, and scripts are public under Apache-2.0. If you build agents, this gives you a cheap local step for routing, tool checks, and guardrails.
Strands Agentsverified
Judge Amit P. Mehta granted Google's motions to dismiss two antitrust suits brought by Chegg and Penske Media Corporation on September 30, 2026. The publishers argued that Google forces sites to hand over content for snippets, AI training, and AI Overviews as the price of appearing in search. The court held that the complaints did not plausibly plead an agreement, separate products, antitrust standing, or a defined market. If you run a site that depends on Google traffic, the ruling leaves your options where they were: let Google crawl, or leave its index.
US District Court for the District of Columbiaverified
Reddit's developer portal now asks owners of Data API apps to register them, and says access for unregistered apps and users will start to be removed on January 12, 2027. The remaining public API access ends in March 2027. Ars Technica and TechCrunch report that a Reddit staff post on September 30, 2026, also announced the end of RSS support on November 13, 2026, with no replacement for feeds from communities you do not moderate. If your site, bot, or monitoring tool reads Reddit through either path, you now have deadlines.
Redditverified
Anthropic released Claude Code 2.1.286 on September 30, 2026. Four fixes stop credentials from appearing in MCP error messages, logs, and transcripts that were supposed to be redacted. Plugin installs now refuse npm sources that are git repositories or folders, and Remote Control sessions disconnect when an organization's policy turns the feature off. The release also corrects a spend meter that priced some cache writes at the cheaper rate, and it changes what bare mode does.
Claude Codeverified
OpenAI said on September 30, 2026, that it identified and disrupted a coordinated campaign to extract protected reasoning from its models. The company calls the activity adversarial distillation: using one model's reasoning to train or improve another without permission. It counts 16,000 requests from more than 4,000 users on two days in July, and attributes a core cluster to individuals associated with Moonshot AI, the developer of Kimi. OpenAI says no encryption was broken and no stored conversations were accessed. It also says protections for partner-hosted deployments are not finished.
OpenAIverified
Google announced Gemini 4 Argon on September 30, 2026. The model is rolling out to a set of trusted cyber defenders through the Fairwind Program, and Google gives no date for developers, enterprises, or consumers. The introductory price is $2 per million input tokens and $10 per million output tokens, and it rises to $4 and $20 when the introductory period ends. The output limit grows to 1 million tokens from 64,000. The benchmark results in the announcement are Google's own reporting, and nobody outside the program can check them yet.
Googleverified
