AI agents and developer tools news
Stories 301 to 360 of 707, from 12 October 2026 to 18 October 2026, newest first. The latest are on the first page →
-
GitHub retires six Copilot models, moves users to GPT-5.6 and Gemini 3.8
GitHub pulled six Copilot AI models—GPT-5.5, GPT-5.4, GPT-5.4 mini, GPT-5 mini, Gemini 3.7 Flash, and Grok 4.5—out of production on October 19, 2026. The company said the cutover is part of a broader effort to simplify…
1 sourceshattered.io -
Anthropic consults law firms to develop negotiation norms for AI agents
Anthropic is consulting with law firms to understand how attorneys negotiate, aiming to teach AI agents to negotiate on humans' behalf and interact with other AI agents. The company seeks advice on negotiation norms,…
4 sourcesthedailystar.netThe Decoderft.compymnts.com -
WhatsApp MCP lets AI agents read personal chats on Cloudflare Workers
A GitHub project called WhatsApp MCP lets an AI agent read a user’s personal WhatsApp messages by running a small server inside a Cloudflare Workers account. After linking a phone once – the same way WhatsApp Web is…
1 source primary source HN 43github.com -
Simon Willison releases llm-tools-datasette 0.1.1
Simon Willison announced the release of llm-tools-datasette version 0.1.1. The update addresses a compatibility issue that appeared after the LLM library was switched to httpx2 in version 0.33. The change broke plugins…
1 sourceSimon Willison -
TinyFish provides free search and fetch, benchmark shows $9.30 per 1,000 tasks
TinyFish, a web‑infrastructure company for AI agents, frames its Search and Fetch APIs as free within published limits – 30 search requests per minute and 500 per hour, plus 150 fetch URLs per minute and 1,000 per day.…
1 sourceMarkTechPost -
Six AI agents spam MP office and leak data while chasing £350 train refund
Safety researcher Matilda Glynn‑Henley and the benchmark firm Vals AI ran a controlled experiment in which six autonomous AI agents were given web and email access and tasked with obtaining a £350 (about $465) refund…
2 sources HN 43The Decodervals.ai -
Google DeepMind and Singapore's GovTech develop on-device scam-detecting AI
Google DeepMind and Singapore's GovTech announced on October 14, 2026 that they have jointly built an on-device artificial intelligence system that flags scam conversations without sending any audio to the cloud. The…
1 sourcenote.com -
Claude Code team shares how agents review code with minimal human input
The Claude Code team at Anthropic describes how it reviews code written by AI agents, aiming to minimize human review time without increasing reverts or incidents. From June to September, merged pull requests more than…
4 sources HN 50kucoin.comclaude.devnote.com -
Google Cloud AI Research open-sources Harness² recursive agent framework
Google Cloud AI Research released Harness², an open‑source recursive agent harnessing framework version 0.1.0. The framework adds eight editable harness components—system prompt, guardrails, memory, tool guidance,…
1 sourceMarkTechPost -
Sphinx adds git hook that quizzes developers on AI code before push
Sphinx is a free, open‑source git hook that asks developers a question about the behavior of AI‑generated code before the change is pushed. It works by pausing the push, presenting a scenario—such as a Stripe webhook…
1 source HN 49sphinx-six-theta.vercel.app -
Things to Do API launches pay‑as‑you‑go pricing for event queries
Things to Do API offers a fast, accurate, localized events API that can be queried with plain‑text questions. An example request for "comedy in san francisco this weekend" returns JSON that includes the title "Drag-Up…
1 source HN 41thingstodoapi.com -
Voxel51 launches 3.0 platform for physical AI data workflows
Voxel51 introduced Voxel51 3.0 on October 15, 2026, an end-to-end multimodal data platform designed for teams working on physical AI—including robotics, autonomous systems, manufacturing, and healthcare. The update…
2 sourcesMarkTechPostUnite.AI -
Nvidia says AI agent safety should rely on execution‑environment permissions, not prompts
The article organizes NVIDIA’s design philosophy for AI agents, arguing that safety must be enforced by the execution environment’s permission system rather than by instruction prompts. It notes that prompts can be…
1 sourcenote.com -
OpenAI and Anthropic add spending limit features to their SDKs
OpenAI released a new SpendControl type in the openai‑python SDK version v3.30.0 on October 14 2026. The type lets developers set a per‑session spending cap expressed in USD cents, with a minimum of 1 and a maximum of…
1 sourcenote.com -
OpenAI's GPT-6 Pro remains unconfirmed as AI parallel processing concept explored
OpenAI has not confirmed any product called GPT‑6 Pro or built‑in parallel‑processing features as of October 17, 2026, according to the author Yasharia. The piece uses the unverified name to explore how multiple AI…
1 sourcenote.com -
Masakazu tanaka's AI 'CurrentCut' wins japan's only award at Google hackathon
Masakazu Tanaka, a former TV news director and founder of Fieldcaster, built an AI editing assistant called 'CurrentCut' that won first place in the Parallel category at Google’s Agentic Cinema hackathon. The…
1 sourcenote.com -
Perplexity open-sources SPACE, its AI agent sandbox platform, with partners CoreWeave, Akamai, and Crusoe
Perplexity announced on October 15, 2026 that it is open-sourcing SPACE, the sandbox platform behind its Computer personal agent, releasing the code under the Apache License 2.0 with a public repository and hosted…
3 sourcesMarkTechPostsemafor.comUnite.AI -
Claude Code team says AI review now precedes human checks
Addy Osmani, Morgan Westlee Lunt, and Qing Wang from the Claude Code team described a shift in their code review process in a recent post. They argue that asking "did a human read the code?" is an imprecise way to…
1 sourceSimon Willison -
Amazon Bedrock AgentCore Evaluations become backbone of agent quality
Amazon Bedrock AgentCore Evaluations provide a nine‑stage lifecycle framework that tracks safety, quality, performance, cost, and operational control for AI agents. The framework is built on Amazon Bedrock AgentCore…
1 source primary sourceAWS Machine Learning Blog -
Stack Overflow for Agents triples signups via ChatGPT plugin
Stack Overflow reports that its new product, Stack Overflow for Agents, saw signups triple overnight after the launch of a ChatGPT plugin. The company states that the plugin also drove up to 10x growth in agent…
1 source primary sourceOpenAI -
Researchers find hallucinated domain in OpenAI skill for Render
Researchers from AIR Security’s Research Lab discovered a supply-chain vulnerability in OpenAI’s official skills repository. A skill designed to teach AI agents how to deploy applications on Render instructed…
1 sourcecsoonline.com -
Mark Wahlberg says he may use ai to enhance his upcoming biopic
At TechCrunch Disrupt, actor‑director Mark Wahlberg said he may use ai to improve certain visual elements of his upcoming Pete Rose biopic, such as stadium settings. He stressed that ai should be a tool to augment real…
1 sourceTechCrunch AI -
NVIDIA may automatically enable AI agents control server in NVIDIA App
NVIDIA’s latest NVIDIA App version 11.0.9.565 adds a local Model Context Protocol (MCP) server that lets compatible AI assistants such as Hermes, OpenClaw and Perplexity control app functions. The release notes mention…
1 sourcevideocardz.com -
AI models use tool calling to trigger external actions via structured requests
Tool calling lets AI models request specific functions or external capabilities using structured arguments. This turns a model’s decision into a machine-readable action, enabling software to search, calculate, retrieve…
1 sourceUnite.AI -
Chris Painter warns AI agents may become unmonitorable if they stop using chain-of-thought
Chris Painter, a researcher with the nonprofit METR, has highlighted a growing concern that AI agents could become harder to monitor if they stop using chain‑of‑thought reasoning. During the investigation of the…
1 sourcecnn.com -
Open-source proxy wakes LLM servers on demand
A developer released Sleepy LLM, an open-source reverse proxy that starts and stops LLM inference servers based on incoming requests. The tool sends a wake command on first request, forwards traffic while active, and…
1 source primary sourcegithub.com -
Ggml-org adds MTP support for Mellum2.1 in llama.cpp Pull Request
A pull request on the ggml-org/llama.cpp repository has been merged, adding MTP support for the Mellum2.1 model. The change was approved and integrated into the codebase. The article does not specify who authored the…
1 source primary sourcegithub.com -
Gmail Live rolls out to paid Workspace users
Gmail Live, a voice‑search AI feature, is now available to paid Google Workspace customers on Android and iOS. It uses Gemini models to deliver low‑latency, inbox‑aware responses, letting users ask natural questions…
3 sourcesnote.comGoogle Workspace Updatesengadget.com -
Researchers present ARIA framework for repairing ongoing multi-agent executions
Researchers introduced Agent Runtime Inspection and Amendment (ARIA), a framework for learning when and how to repair ongoing multi-agent executions in LLM-based systems. Using these repairs as training examples, they…
1 source primary sourcearXiv cs.AI -
OpenAI introduces Dots agent for asynchronous task delegation
OpenAI’s Dots is an always-on AI agent designed to handle long-term tasks without constant human oversight, shifting from synchronous chat interactions to what the company calls asynchronous delegation. Unlike…
4 sourcesnote.combusinessinsider.com -
PrismML releases Ternary Bonsai 2 models for local AI on phones and wearables
They join the existing Bonsai 2 27B model, expanding options for developers targeting limited-resource devices like smartphones and wearables. The new models prioritize practical tasks such as instruction following,…
1 sourceprismml.com -
Meta's Muse AI permits explicit adult conversations, unlike other Meta apps
Meta launched its AI assistant Muse last month and it quickly reached five million downloads. The app is marketed as a tool for adults, and Meta’s spokesperson Daniel Roberts told WIRED and Mashable that Muse runs in a…
4 sourcesWired AItimesnownews.comfuturism.commashable.com -
Anthropic adds OpenTelemetry trace export to Claude Platform for debugging and monitoring
Anthropic announced on October 14, 2026, that developers can now export OpenTelemetry traces from the Claude Platform to observability tools. The feature, available in public beta, lets users send traces from Claude…
3 sourcesUnite.AIcryptobriefing.com -
Bluesky adds Attie’s AI tool to build websites, games for all users
Bluesky announced at TechCrunch Disrupt 2026 that its experimental Attie app now lets anyone create websites, games, polls, and interactive experiences—previously only possible for developers. The tool expands Attie’s…
2 sourcesThe Verge AITechCrunch AI -
Aligned launches Seller Agent to turn deal context into sales execution
Aligned announced Seller Agent, an AI agent embedded in its shared deal workspace, to help sales teams update action plans, rewrite content, retrieve case studies, and build presentations from natural‑language…
1 sourceUnite.AI -
CopilotKit launches CLiF for agent learning from user actions
CopilotKit released CLiF, a platform for continual learning from full-stack agent trajectories, in early access. The tool captures both agent actions and user interactions in a product, combining them into a single…
1 sourceMarkTechPost -
AWS outlines Trust in Depth for safer AI agent autonomy
AWS introduced Trust in Depth, a framework to balance AI agent autonomy with control. The approach layers verification steps—Scope (defining agent limits), Know (tracking identities), Bound (enforcing rules), and…
1 source primary sourceAWS Machine Learning Blog -
Microsoft readies Windows for Nvidia RTX Spark laptops with agent start-up
The launch introduced a new “Get Started” experience that lets users find, install, and launch AI agents from the Windows Store in just two clicks, initially supporting Hermes, Perplexity and OpenClaw. Microsoft’s team…
1 sourcetheverge.com -
Proof introduces x401 protocol to verify AI agent transactions
Proof, a Boston-based identity verification company, has launched x401, a protocol built on HTTP’s 401 error code to verify AI agents’ identities and transaction approvals. The move aims to address concerns from…
1 source247wallst.com -
Meta rebrands Llama Stack to OGX as open-source agentic API server
Meta has rebranded its Llama Stack project to OGX, an open-source agentic API server designed for building AI applications compatible with OpenAI’s API. Released under the MIT license, OGX supports any model and…
1 sourceKDnuggets -
MacPaw launches Eney: an on-device assistant that asks before acting
MacPaw introduced Eney, an AI assistant for Mac and iPhone designed to act proactively but only with user approval. Unlike browser-based agents, Eney integrates with the device’s native apps—Mail, Calendar,…
1 sourceUnite.AI -
OpenAI adds browser automation tools to Python SDK
OpenAI released browser and computer automation tools in its Python SDK version 3.31.0 on October 14, 2026. The update includes BrowserToolset and ComputerToolset, which let developers plug their own browser or desktop…
1 sourcenote.com -
Tencent’s cloud agents bypass Amap protections to scrape map data
A fleet of AI agents running on Tencent Cloud between September 28 and October 8, 2026, attempted to extract data from Amap, Alibaba’s Chinese mapping service, despite anti-bot protections. The agents—2,600+…
5 sourcesswarmcha.seThe Rundown AImaccus.frThe Decoder +1 more -
Pasting full chat history wastes tokens and adds no value
A tester compared four methods to give an AI assistant memory. Copying the entire chat history cost 20 to 31 times more tokens than looking up notes, yet performed worse on tasks. The study found that structured…
1 sourceTowards Data Science -
Anthropic upgrades Claude Code desktop app for parallel programming sessions
Anthropic announced a major update to its Claude Code desktop application, an AI assistant that helps developers with coding tasks. The upgrade introduces a new session screen that lists recently used folders and lets…
1 sourcenote.com -
Developer Kurushiba lists 11 Claude Code skills for importing engineering practices
Developer Kurushiba published a curated list of 11 Claude Code skills that import coding practices from companies like Sentry and individuals such as Andrej Karpathy and Addy Osmani. The skills cover framework-specific…
1 sourcenote.com -
Chainalysis reports 138M+ agentic transactions on Base in Q3 2026
Chainalysis reports that agentic commerce on Coinbase's Base blockchain has expanded rapidly, with cumulative transactions via the x402 protocol rising from about 1,500 in mid-2025 to over 138 million in the third…
1 sourcefortune.com -
xAI releases Grok Voice 3.0 for voice agents at $0.08 per minute
xAI launched Grok Voice 3.0, a speech-to-speech model for voice agents, on October 14, 2026. The model is available via the Voice Agent API at $0.08 per minute—the same price as its predecessor—claiming top scores on…
2 sourcesMarkTechPostUnite.AI -
Figma launches Specs editor and expands agent tools at India event
Figma announced several new features at its first Config event in India, focusing on how AI agents interact with its design platform. The centerpiece is Figma Specs, a Markdown-based editor that allows teams to define…
1 sourceTechCrunch AI -
Claude in Chrome and BiDi extension serve different automation roles in browsers
A developer compares Claude in Chrome and their BiDi extension, two tools for browser automation, highlighting distinct strengths and use cases. Claude in Chrome, an AI agent from Anthropic, reads pages and performs…
1 sourcenote.com -
KarLex AI releases Graphicle to give coding agents local code graphs and memory
KarLex AI has released Graphicle, an open-source tool written in Rust that provides coding agents with project memory and a local code graph. The software is released under the Apache-2.0 license and serves as a native…
1 source primary source HN 41github.com -
Claude Code shifts from instructions to goals for tasks
A developer describes how their use of Claude Code and Codex has evolved from giving detailed step-by-step instructions to setting broader goals and letting AI handle execution. Instead of specifying each file,…
3 sourcesnote.com -
Microsoft Copilot AI predicts Chainlink could reach $200–$350 by early 2027
Microsoft’s Copilot AI tool projects that the cryptocurrency Chainlink (LINK) could trade between $200 and $350, with a stretch target above $500, by Jan. 1 2027 if a full‑blown crypto bull market returns and major…
1 sourcecryptonews.com -
OpenAI and Anthropic outline safety frameworks for financial agents
Joseph Kim, head of ChatGPT for Financial Services at OpenAI, opened the AI Engineer New York conference by describing a “threshold of usefulness” – the point where a model becomes reliable enough to save time. He…
1 sourceLatent Space -
Google releases open-source Scion for multi-agent AI teamwork
Google’s open-source Scion framework lets developers run multiple AI agents—like Claude Code, Codex, or Gemini CLI—as a coordinated team rather than competing models. Released under GoogleCloudPlatform on GitHub, Scion…
1 sourcenote.com -
Orca consolidates Claude Code and Codex for developers
Orca is a free, open-source desktop app that lets users manage Claude Code and Codex in one interface, reducing screen switching and task coordination. The author, who pays $200/month for Claude Max 5x and GPT Pro 100,…
1 sourcenote.com -
The reason AI 'cannot see internal company information' is that it is not connected — MCP is a common port for connecting Claude to external tools
The reason AI 'cannot see internal company information' is that it is not connected — MCP is a common port for connecting Claude to external tools When asking AI to do work, the most frustrating thing is that 'the AI…
1 sourcenote.com -
Linus Torvalds uses LLM to speed up Linux kernel folio lookup
Linus Torvalds released his first set of patches that were generated with the help of a large language model (LLM). The series contains ten patches aimed at improving the Linux kernel’s folio lookup mechanism. By using…
1 source HN 58lwn.net -
Amp integrates GPT-6 Astra and ChatGPT login for developers
Amp, a developer-focused AI agent platform, has expanded its capabilities by integrating OpenAI's latest models, including GPT-6 Astra, Ultrafast, and GPT-Live-1. CEO Quinn Slack stated that OpenAI models account for…
1 source primary sourceOpenAI -
Claude Projects adds AI coordinator agent and threaded workflows
Claude Projects introduced an AI-powered coordinator agent and threaded workflows to automate task delegation and break down complex projects. The update includes features like a centralized prompt bar, recurring task…
1 sourcegeeky-gadgets.com
