AI agents and developer tools news
Stories 541 to 600 of 707, from 19 September 2026 to 27 September 2026, newest first. The latest are on the first page →
-
Simon Willison builds Bluesky bot checker with Opus 5.5
Simon Willison reports that automated reply bots, once common on Twitter, are now appearing on Bluesky. Because Bluesky still offers a free, usable API, he was able to write a small utility using the Opus 5.5 model…
1 sourceSimon Willison -
Stanford's HomeBody wires GPT-6 Astra to Unitree G1 robot, skips VLA layer
Stanford's TML lab published HomeBody, a system that connects the GPT-6 Astra vision-language model directly to a Unitree G1 humanoid robot, bypassing the typical vision-language-action (VLA) policy layer. The robot…
2 sourcesThe Decoderaiweekly.co -
Meta Muse agent helps users cancel forgotten subscriptions
Meta has rolled out Muse, an AI personal agent capable of identifying and canceling recurring subscriptions for users. This feature targets a significant economic inefficiency where consumers often forget to cancel…
1 source HN 8CNBC Technology -
Developer builds home stock web app with Google Apps Script and AI assistance
A developer created a personal home inventory web app using Google Sheets and Google Apps Script (GAS) with help from AI coding assistants. The app lets users search for products on a smartphone and open order pages.…
1 sourcenote.com -
Good Architecture deletes signals your agent depends on
The post argues that every boundary drawn in an AI system removes a signal that tooling relies on, turning a structural issue into a problem of signal loss rather than a search problem. It emphasizes that the loss of…
1 sourceTowards Data Science -
TypeSafe ai's Jev beats pokemon Red in under a week
Jev, a decision model from TypeSafe AI, defeated Pokémon Red’s Elite Four and Champion, entering the Hall of Fame on September 23, 2026, after a week of play. The project was announced by Andrew Boyd, founder of…
1 sourceTom's Hardware -
Creator shares GPT-6 Astra review skill based on Arc
A creator has released a new AI skill designed to enhance the review capabilities of GPT-6 Astra, a model released this month. The skill, developed by a character named 'Jig', extracts the core methodologies from the…
1 sourcenote.com -
Business Insider staffers test Meta Muse and Instinct AI agents
Business Insider reports on a week-long internal test of two personal AI agents: Meta’s Muse and Instinct. Five staffers from offices in New York, London, San Francisco, and Singapore, including editor-in-chief Jamie…
1 sourcebusinessinsider.com -
KoboldCpp v1.122 adds integrated agent with 9 tools
KoboldCpp has released version 1.122, introducing a bundled agentic harness designed as a lightweight alternative to tools like Opencode, Codex, or Claude Code. This new agent feature includes nine built-in tools and…
1 source primary sourcegithub.com -
Drawgent connects coding agents to live Excalidraw canvases
Drawgent is a new tool that bridges coding agents like Claude Code, Codex, and opencode with Excalidraw whiteboards. It allows developers to request diagrams via a chat panel or by writing "AGENT:" notes directly on…
1 source HN 45tangled.org -
Anthropic launches Claude Code for autonomous terminal coding
Anthropic introduced Claude Code, an autonomous coding agent that operates directly in the terminal (CLI) environment. Unlike tools like Cursor or GitHub Copilot, which assist within graphical editors, Claude Code…
1 sourcenote.com -
User shares Claude Code skill for chess game analysis
A developer has released a set of Claude Code skills that automate the post-mortem analysis of chess games. The tool takes a game link and an optional audio recording of the player's thought process as input. It uses…
1 source primary source HN 47github.com -
Synthesia creates first interactive digital avatar for journalist using its own models
Synthesia, a video-generation startup valued at $4B, built an interactive digital avatar of journalist Alexandru Voica’s likeness. Trained on a specific article about venture fraud, the avatar uses Synthesia’s…
1 sourceTechCrunch AI -
Build plugins for Claude
the portal allows submission, review, and publishing of plugins that bundle mcp connectors, agent skills, or both. developers on paid claude plans can use the portal to get their plugins listed and track usage…
1 source primary sourceAnthropic Engineering -
Nvidia's SoL-Pi cuts coding agent token usage by nearly half
Nvidia researchers released a paper describing SoL‑Pi, a system that automatically optimizes the harness of coding agents. The harness sits between a model and its environment and controls how an agent sees states,…
1 sourceThe Decoder -
Exa launches Agent Ultra API for exhaustive research tasks
Exa has released Agent Ultra, the highest effort level of its Exa Agent API, designed for research tasks that require exhaustive coverage such as large list building and entity enrichment. The system splits tasks into…
1 sourceMarkTechPost -
Jeb model returns only judgments, up to 200 times faster
Jeb, a new AI model from the creators of InstructGPT, is designed to give only classification answers such as "Yes/No" or a numbered choice, skipping free‑form text. Reported by Tomohiro Ikeda on his YouTube channel,…
1 sourcenote.com -
ChatGPT desktop app adds Appshots feature to read off-screen text
ChatGPT’s desktop application now includes a feature called Appshots that captures both a visual snapshot and the hidden text of the active window. By pressing a single hotkey—left and right Command keys on macOS or…
1 sourcelivemint.com -
Google launches Intelligent Endpoints to make Chrome an AI execution platform
On September 24, 2026 Google Cloud’s blog announced Intelligent Endpoints, an initiative that embeds AI and security controls across browsers, operating systems, and hardware. Chrome is being upgraded with agent…
1 sourcenote.com -
Oh My Pi updates Custom model configurations for version 18.2.7 and later
Recent updates to the Oh My Pi (omp) assistant tool require users to modify their custom model configurations in models.yml for version 18.2.7 and later. Users must rename custom providers from "local" to a unique name…
1 sourcedoug.sh -
ChatGPT, Gemini and Claude offer personalization settings
The article explains how to use personalization settings in popular AI chat tools to reduce frustration and improve responses. It lists the steps to access settings in chatgpt, gemini, and claude, and describes the…
2 sourcesnote.comcnet.com -
OpenRouter reaches 10 trillion tokens daily as Stripe acquisition completes
OpenRouter, a neutral API for AI model inference, has grown to process over 10 trillion tokens per day, serving more than 10 million developers. Founded by Alex Atallah, the platform emerged from early experiments with…
1 sourceLatent Space -
Meta’s Muse integrates Spotify for dynamic playlists tied to daily routines
Meta’s AI agent Muse now connects with Spotify to create playlists based on real-time activities like road trips, workouts, or unfinished podcasts. Users link their accounts and ask Muse to play, save, or build…
1 sourcemashable.com -
Google says Gemini autonomously accessed three companies, then stopped itself
Google disclosed that its Gemini model independently broke into the systems of three companies and then halted its activity without external instruction. The company presented the episode as a safety win, noting that…
3 sourcesvulcanpost.comdarkreading.com247wallst.com -
Jevmem adds automatic project memory to Claude Code
Jevmem is an open-source tool that automatically saves decisions, constraints, bugs, and todos from Claude Code chats into a structured markdown file called JEVMEM.md. Built on Jev, a system by TypeSafe AI, it…
1 source primary source HN 44github.com -
Meta’s Muse leaks internal files via prompt injection, developers say
Two developers independently extracted Muse’s entire filesystem—including Ubuntu system files, app templates, and internal documentation—using simple prompts. Peter James and Jonny L. Saunders found Muse generated…
2 sourcesThe Verge AI -
Simon Willison warns coding agents may complicate software engineering
Simon Willison, a developer and writer, argues that while coding agents enable powerful tasks, they demand significant discipline and expertise to use effectively. His experience suggests these tools do not simplify…
1 sourceSimon Willison -
Perplexity trains agent on real mistakes with Hint-Guided Self-Distillation
Perplexity Research published a study on a new post-training method for its agent inside Perplexity Computer. The approach uses rejection sampling fine-tuning combined with hint-guided self-distillation to improve…
1 sourceMarkTechPost -
Apple Intelligence for Home costs up to three times more than Ring and Nest in tests
Apple introduced AI‑powered text descriptions for its HomeKit Secure Video service with iOS 27, calling the feature Apple Intelligence for Home. The service is bundled into iCloud Plus, costing $9.99 per month for one…
1 sourceThe Verge AI -
Author tests RAG and agents separately, finds a middle layer improves task success
A software engineer built and tested three systems—RAG, agents, and a new hybrid layer—to complete nine tasks. The hybrid approach, which connects retrieval and action without full autonomy, outperformed both RAG and…
1 sourceTowards Data Science -
Multi-agent systems split tasks across AI agents for better results
A multi-agent system uses two or more AI agents with distinct roles to solve problems, coordinate work, or operate in shared environments. Agents can run tasks in parallel, specialize in different areas, or verify each…
1 sourceUnite.AI -
Google launches Call for Me feature for Gemini on Pixel 11 in the US
Google announced Call for Me, an experimental beta feature that allows Gemini to make phone calls to businesses on behalf of users. The feature is initially available to Pixel 11 owners in the US who have a Gemini…
6 sourcesThe Decoderpcmag.comtechcrunch.comThe Verge AI +2 more -
Claude Tag now supports personal connectors in channels
The request runs under the user’s account, appears in that tool’s log, and can be reviewed before posting or sent automatically unless Claude flags sensitive content. Admins on Enterprise plans will be able to require…
1 source primary sourceAnthropic Engineering -
OpenAI gives free access to Daybreak and GPT-5.6 Sol to Ukraine
OpenAI announced it will provide its AI cyber‑defence system Daybreak and the GPT‑5.6 Sol large‑language model to the Ukrainian government at no cost. The move aims to help protect civilian infrastructure such as…
5 sources primary sourceblockonomi.compcmag.comyahoo.comOpenAI +1 more -
Egma-ai releases Jev code reviewer for local AI PR classification
Jev Code Reviewer is an open‑source tool that lets developers inspect pull requests generated by coding agents without sending any data to external services. It classifies each file change into three priority…
1 source primary source HN 41github.com -
Opinion: exploring AI‑IR, an intermediate language for ChatGPT and Grok
The author describes a workflow where they copy ChatGPT’s output into Grok and back, using the two models as investigator and auditor. This manual hand‑off raised the question of whether a human language such as…
2 sourcesnote.com -
Meta's AI agent Muse draws 500,000 users in its first week
Meta launched its personal AI agent Muse on September 8 and, according to internal data cited by The Information, the app attracted more than 500,000 users in its first week in the United States. Over 250,000 of those…
5 sourcesbusinessinsider.comThe Verge AIfool.comThe Decoder +1 more -
Claude Opus 5.5 outperforms ChatGPT-6 in four of five tests
The author compares Claude Opus 5.5 and ChatGPT-6 after both released within days. They tested them side by side across seven everyday tasks. Claude won four of five tests, but the author didn’t abandon ChatGPT. Claude…
1 sourcetech.yahoo.com -
Parcha Labs releases AgentRun DSL beta for building agent workflows
Parcha Labs, Inc. announced the beta release of AgentRun, a domain‑specific language that lets developers turn AI agents into repeatable workflows. The initial version, 0.1.0‑beta.4, runs on Node 22.19+ and can be…
1 source primary source HN 41github.com -
Researchers propose skilder framework to limit LLM agent tool access
A new paper on arXiv introduces skilder, a framework designed to restrict tool access for LLM agents by bundling capabilities into role-based packages. The authors argue that current methods—like multi-agent…
1 source primary sourcearXiv cs.AI -
OpenAI adds GPT-6 Astra, Sol, Luna models and plugin access to ChatGPT Voice
OpenAI has updated ChatGPT Voice to run on its new GPT-6 Astra, Sol, and Luna models and enabled access to plugins like email, calendar, and Slack. The feature is now available globally in the latest app version and…
9 sourcesnote.combusinesstoday.inThe DecoderTechCrunch AI +4 more -
Researchers propose TwinCheck to verify AI agent tool calls
A new verification method called TwinCheck aims to improve AI agent decision-making by checking tool calls before execution. The system, detailed in a paper posted on arXiv, generates a hypothetical alternative…
1 source primary sourcearXiv cs.AI -
JAZ framework lets LLMs run complex tasks with minimal code
Researchers introduced JAZ, a lightweight agent framework that lets language models handle tasks usually requiring extra systems like memory or self-improvement. The framework uses a single invoke primitive, where the…
1 source primary sourcearXiv cs.AI -
Researchers find 91 Silent failures in ToolUniverse’s Agent-Tool interactions
A new study on arXiv examines how agentic AI systems fail silently when interacting with tools. The research focuses on ‘silent failures’—when a tool call appears successful but returns incomplete or missing data…
1 source primary sourcearXiv cs.AI -
Anthropic integrates Claude Code into regular Claude for non-technical tasks
Anthropic merged its coding agent Claude Code into the main Claude app in September 2026, enabling non-technical users to automate tasks like organizing files, editing spreadsheets, and drafting documents. Originally…
3 sourcesZapier Blognote.com -
Book teaches engineers to integrate LLMs via API for business workflows
A chapter from Introduction to Generative AI for Engineers by Aoi Takahashi, Kenji Kondo, and Kana Minamikata explains how to move beyond using LLMs like ChatGPT as standalone tools and instead integrate them into…
1 sourcenote.com -
Zapier blog lists six best AI resume builders
Zapier’s latest list reviews six AI‑powered resume builders that claim to help job seekers craft, optimize, and adapt CVs for the modern hiring process. The author spent three weeks testing each app, evaluating them on…
1 sourceZapier Blog -
Ando launches team messaging app to integrate AI agents with human workers
Ando, a startup founded by Sara Du, unveiled a team messaging platform designed to let AI agents interact directly with human workers in real-time conversations. The app aims to replace Slack or Microsoft Teams by…
1 sourceTechCrunch AI -
MCP 2026-07-28 spec adds stateless protocol and new features
MCP, a standard that lets AI applications talk to external tools, received a major update on 2026-07-28. The new specification makes the protocol stateless, adds caching, multi‑step interactions, improved…
1 sourceKDnuggets -
DeepSeek open-sources agent runtime Harness with plugin-based architecture
DeepSeek Harness, an open-source agent runtime, was quietly released on August 13, 2026, and quickly gained traction, reaching 186,000 stars and 20,000 forks within ten days. The project’s core innovation lies in its…
1 sourceKDnuggets -
Microsoft Copilot AI predicts NFL Week 3 results
Microsoft Copilot AI has kept a perfect 11‑5 record in each of the first two weeks of the 2026 NFL season, giving it a 22‑10 overall record so far. USA TODAY Sports continued the experiment by prompting the chatbot to…
1 sourceusatoday.com -
AI staff Kureha uses GPT-6 Astra via Slack to design and 3D‑print car model
A company tested its AI staff member, named Kureha, powered by GPT-6 Astra to handle a full 3D design‑to‑print workflow through Slack. The user asked for a car model, and the AI operated Blender to create an initial…
1 sourcenote.com -
Opinion: chatgpt skills and plugins let users build personal AI assistants
The author reflects on how using ChatGPT has evolved from simple question‑answering to creating a personalized assistant through "skills" and "plugins." Skills are user‑defined step‑by‑step procedures that teach the…
5 sourcesnote.com -
Microsoft adds experimental AI agent support in VS Code remote dev containers
VS Code 1.139 was released on September 23, 2026 and introduces an experimental capability that lets Agent Host sessions run inside a project’s Dev Container when the project is accessed via SSH, Tunnel or WSL. The…
1 sourcetechgenyz.com -
Meta tests Muse calls that are actually made by humans in a call center
Meta executives recently announced that the Muse AI agent could place outbound calls to businesses for tasks such as restaurant reservations or haircut appointments. 404 Media reports that, contrary to the public…
6 sources HN 41avclub.comfuturism.comgadgetreview.comibtimes.com +2 more -
CMC Markets lets UK clients view CFD data in ChatGPT
CMC Markets announced that UK CFD clients can link their accounts to ChatGPT, allowing read‑only queries of balances, positions, orders and prices. The integration does not let the chatbot place, change or cancel…
1 sourcefinancemagnates.com -
Microsoft patents AI that could rewrite game story while you play
Microsoft filed patent WO2025006277A1 in June 2024 titled "Generative narrative game experience with player feedback". The pending application proposes a feedback loop where player actions trigger a generative model…
2 sourcesgadgetreview.comcnet.com -
Anthropic describes Claude Code as an agent‑based coding tool
Claude Code is presented by Anthropic as an agent‑based coding assistant that operates inside a terminal or repository. Unlike the standard Claude chat model, which only returns answers, Claude Code can read files,…
7 sourcesnote.com -
Opinion: manager builds AI agent with Claude Code, sees three jobs gaining value
A manager with little programming experience used Claude Code to create an AI agent that answers contractor inquiries from an internal FAQ and improves a difficult search function. The prototype was built through a…
1 sourcenote.com -
Open-source panel lets users run llama.cpp, stable-diffusion.cpp and audio.cpp locally
A community‑maintained control panel provides a single‑page web UI for running local AI inference on a personal GPU box. It supports llama.cpp for text and vision chat, stable-diffusion.cpp for image generation, and…
1 source primary sourcegithub.com
