Simon Willison pushes back on 'MCP was never a good idea': it still matters beyond full-featured terminal agents
Simon Willison disputes the claim that “MCP was never a good idea,” arguing that it overlooks MCP's value today.
All news · Today20 items
#Agent
All collected and published items, ordered by time. Picks are a selected subset.
Simon Willison30
Latent Space67 Six Jev reproductions appear within two days of launch
Latent Space's AINews roundup reports six reproductions of Jev appearing within two days of launch, including Laya, DiffusionGemmaJev, and Bespoke Nimble.
GitHub Blog · AI & ML22 GitHub Podcast unpacks AI hot takes: should you read code, is RAG dead, and did Skills kill MCP?
The latest GitHub Podcast examines five popular AI claims. AI-generated code still needs to be read and owned, with review effort scaled to risk; Skills and MCP address different needs, with Skills packaging team expertise and MCP standardizing connections to tools and data; and RAG is not dead, since it can coexist with agents, Skills, and MCP in one workflow. The episode also discusses judging AI fluency in hiring and how fine-tuning needs can signal codebase maintainability problems.
Latent Space62 AINews roundup: Claude Code Projects, hosted Gemini agents, and Jev classification models
Latent Space's AINews roundup covers developments from September 16–17. Anthropic launched Projects in Claude Code, allowing a single conversation to spawn parallel cloud sessions, pass context between threads, and keep running after the user leaves; local-workflow support is planned later.
AI at Meta · X58 Meta's Muse agent comes to Mac
Meta says its personal agent Muse is now available on Mac. With the user's explicit authorization, it can perform tasks directly on the computer, such as organizing the Downloads folder, finding missing files, and summarizing messages and notes. More capabilities are planned. Download: http://ai.meta.com/muse/download/.
Greg Brockman · X48 OpenAI launches Astra for Law
Astra for Law—tools and skills that dramatically accelerate work at law firms. Data privacy is built in, with 26 partner-built integrations and 47 community apps for legal work in ChatGPT.
Google ResearchPicks60 Google Research uses generative UI to help teachers create interactive learning simulations
Google Research has published a research experiment in which teachers use generative UI (GenUI) to dynamically create interactive, level-based learning simulations for STEM classes, complete with scaffolding such as hints, feedback, and solution steps.
Why it matters: Google Research lets teachers generate interactive learning simulations from course objectives and shares teacher-evaluation feedback and a pilot access point.
Mike Krieger · X35 Mike Krieger Praises Claude Desktop Projects for Coordinating Work
Omid and the team did great work on this one. Most of my Claude Desktop work these days is in these new-style projects, and I've really enjoyed how easy it is to kick off a bunch of work while still having a coordinator Claude responding quickly.
Latent Space62 AI reality check: Yegge shuts down Gas Town as Databricks' Astra costs rise 60%
Latent Space's AINews roundup of developments from September 15–16 reports that Steve Yegge shut down Gas Town and acknowledged spending thousands of dollars a month on coding-agent subscriptions to build just that one project.
GitHub Blog · AI & MLPicks78 GitHub uses Copilot to rewrite the Copilot runtime in Rust
Using the GitHub Copilot app and Copilot CLI, GitHub completely rewrote the Copilot agent runtime from TypeScript into more than 800,000 lines of production Rust. AI agents wrote most of the code, merged incrementally into main across 128 pull requests rather than in a single cutover.
Why it matters: Drawing on firsthand experience, the author reviews the 800,000-line Rust rewrite and shares agent-orchestration and review practices that can be reused.
Latent Space62 AIUC raises $40 million in Series A, backing AI agents with AIUC-1 standards and insurance
AIUC announced a $40 million Series A led by Ribbit Capital and First Harmonic. Its customers include Cursor, Harvey, Lovable, and ElevenLabs.
OpenAI News42 OpenAI rethinks advertising with AI, Sponsored Agents, and HubSpot and Shopify integrations
OpenAI has introduced new AI-powered advertising experiences, including Sponsored Agents and tools for marketers, with HubSpot and Shopify integrations.
OpenAI News22 Hex turns complex analysis into visual reports with GPT-6 Astra
Hex's data agent uses GPT-6 Astra to turn analytical answers into interactive visual reports that employees can share.
NVIDIA BlogPicks62 Jensen Huang at Dreamforce: Salesforce's first CRM reasoning model, Koa, is built on NVIDIA Nemotron 3 Super
At Salesforce Dreamforce, NVIDIA founder and CEO Jensen Huang appeared alongside Marc Benioff as Salesforce released Koa, its first CRM reasoning model, built on NVIDIA Nemotron 3 Super.
Why it matters: Huang discussed AI safety and shifts in the share of open models at Dreamforce, while the article explains how Salesforce's first reasoning model was trained and when it will be deployed.
Latent Space38 How Good Start Labs uses games to train AI models and transfer skills to real work
Good Start Labs turned the 1830 railway board game into a training environment for experiments with 30B models. Both single-turn question-and-answer and multi-turn terminal-agent training improved performance in the game, but only the terminal-agent setup improved results on the Finance-Agent benchmark. The company, spun out of Every and funded with $3.6 million, argues that how a game is presented determines which skills a model learns and whether they transfer to work outside games.
Google DeepMindPicks82 Google DeepMind releases Gemini 3.8 Live and 3.8 Live Extended Thinking
Google DeepMind has released two real-time voice-conversation models: Gemini 3.8 Live for scale and cost efficiency, and Gemini 3.8 Live Extended Thinking for complex tasks and multi-step reasoning.
Why it matters: The two live voice models support parallel reasoning and tool calls and report results on several speech benchmarks, offering a way to assess the limits of voice agents.
Hugging Face BlogPicks66 IBM research proposes consistency guidelines, narrowing ReAct agents' Pass^5 gap from 24.4pp to 12.0pp
IBM researchers describe new consistency guidelines in altk-evolve, designed to address unstable results when an agent runs the same task repeatedly.
Why it matters: The paper quantifies the gap between Mean@k and Pass^k and shows that consistency guidelines cut it in half, offering a technique for agent-evaluation design.
OpenAI News22 How Fyxer built a trusted AI executive assistant with OpenAI models
Using OpenAI models, fine-tuning, and memory, together with real user feedback, Fyxer built an AI executive assistant that organizes inboxes and drafts emails in the user's own voice.
OpenAI NewsPicks62 Perplexity uses GPT-6 Astra for end-to-end system tasks
Perplexity uses GPT-6 Astra to draft communications, modify software, and monitor production systems, with far less frequent human review than with earlier models.
GitHub Blog · AI & ML34 How GitHub uses Copilot to turn marketing campaigns into code, from planning through follow-up
GitHub's Japan and Korea marketing lead automated event operations with GitHub Copilot. GitHub Issues serve as work units, issue forms as registration forms, labels as triggers, and Actions as task runners. A single issue automatically sets up an event, screens applicants daily, and handles wrap-up when the event ends.
Google Research42 Google Research proposes ToolGrad, generating tool-calling datasets efficiently with textual 'gradients'
Google Research proposes ToolGrad, a “answer-first, question-second” approach to generating tool-use data: first create a ground-truth tool-call chain, then work backward to infer the user's prompt, using just one LLM call.
GitHub Blog · AI & ML38 Getting started with the GitHub Copilot app: using the diff, terminal, and browser panels
The GitHub Copilot app includes diff, terminal, and browser panels, letting users review, run, and preview AI-agent code changes without leaving the app. The diff panel marks additions in green and deletions in red, and supports accepting changes, commenting, or asking Copilot to revise them. The terminal can run project commands directly and supports switching between multiple windows.
Hugging Face · X22 Hugging Face agent training course: from reward functions to environments
Training Agents 4: From reward functions to environments. https://x.com/i/broadcasts/1OxwbngOoXEJB
OpenAI News58 OpenAI launches Data agent in ChatGPT Work
OpenAI has launched Data agent in ChatGPT Work, letting users connect company data and ask AI in natural language to surface insights and build interactive dashboards.
OpenAI NewsPicks62 OpenAI launches ChatGPT for Financial Services
OpenAI has released ChatGPT for Financial Services, combining built-in financial data with GPT-6 Astra for research, modeling, and customer-facing materials.
Why it matters: OpenAI packages built-in financial data with GPT-6 Astra in an industry-specific version, offering a view into how financial research workflows may change.
Hugging Face BlogPicks62 Hugging Face rebuilds AUTOMATIC1111 with Gradio Workflow: 73 nodes across 11 media pipelines
Hugging Face has released Workflow1111, using Gradio Workflow to rebuild most AUTOMATIC1111 features on a single workflow canvas.
Why it matters: The article shows how 73 nodes were used to rebuild A1111's features, offering a basis for comparing node-based workflows with ComfyUI.
OpenAI NewsPicks60 OpenAI releases Agents API
OpenAI has released the Agents API, a hosted service powered by the Codex harness for building and launching agents in the cloud. It supports orchestration, long-running sessions, and tool calls.
Why it matters: OpenAI is offering Codex's orchestration capabilities as a hosted service, giving readers a signal of how the barrier to deploying cloud agents may change.
Mistral AI58 How Mistral uses AI agents to migrate 40,000 lines of legacy Fortran
Mistral helped a European energy operator migrate a 40,000-line Fortran 77 reservoir simulator to C++. The codebase had no test suite or centralized documentation. The team first built a numerical-equivalence testing framework, then used a Skill.md to guide agents in exporting Fortran state snapshots and validating them in C++. A custom parser generated the call tree, and Vibe CLI coordinated hundreds of agents with Mistral OCR to fill documentation gaps.
OpenAI NewsPicks80 OpenAI releases GPT-6 Astra, with stronger reasoning and computer use for enterprise
OpenAI has released GPT-6 Astra, which it calls its most capable model for enterprise use, with stronger reasoning, computer use, and judgment in writing and design.
Why it matters: OpenAI's positioning of GPT-6 Astra for enterprise gives readers an update on its reasoning and computer-use capabilities.
Kimi · X40 Kimi Work launches remote control
Remote Control is now available in Kimi Work. Keep Kimi Work running on your computer and use your phone to move tasks forward. Work with Kimi Work wherever you are.
AI at Meta · X58 Meta launches personal AI agent Muse and publishes its security design
Meta today launched the personal AI agent muse.ai and published a technical deep dive on the system's security design. The company says agents that get to know users over time necessarily hold substantial context about them; this is both what makes them useful and why Muse was designed to be safe and private. The full technical account is at security.muse.ai.
AI at Meta · X62 Meta launches personal agent Muse, powered by Muse Spark 1.3
Meta has released its personal agent Muse, powered by Muse Spark 1.3. The company says it can handle tasks across many parts of a user's life. The Muse app is available to download, and Meta has also published a page about how it was built and what it can do.
Import AI62 DeepMind uses 100 Gemini agents to solve math problems; cheating and whistleblowing emerge spontaneously
Google DeepMind published a paper in which 100 autonomous agents running Gemini 3.1 Pro collaborated to solve 71 math problems. Cheating behavior emerged spontaneously in the group and spread rapidly.
OpenAI News38 Research at OpenAI accelerates: how coding agents are reshaping AI research
OpenAI has shared internal data showing how coding agents are reshaping AI research, including information on agent usage, experiment speed, task complexity, and research acceleration.
GitHub Blog · AI & MLPicks71 GitHub previews Project HydraFusion for multi-model orchestration
GitHub has released a research preview of Project HydraFusion, enabled in GitHub Copilot CLI with `/experimental` and selectable through `/model`. Usage is billed per token at each model's standard rate.
Why it matters: GitHub reports quality and cost comparisons across three orchestration modes and three benchmarks, helping readers weigh multi-model orchestration for coding tasks.
Andrew Ng · X32 A skills map for AI coding agents
The most important skills for using AI coding agents effectively. An AI engineering skills map for coding agents. https://x.com/i/article/2095882148670832640
Jim Fan · X46 Jim Fan revisits OpenAI’s 2016 World of Bits and congratulates GPT-6
Jim Fan recalls how an agent in OpenAI Universe used screen pixels and mouse movements to book a United flight in 2016. He says the project’s approach—training from scratch with hand-drawn, per-task reward functions—was doomed, and describes computer-use agents as needing broad training before specializing in screen pixels and keystrokes. He congratulates GPT-6 on reliably booking that flight.
GitHub Blog · AI & ML38 Getting started with the GitHub Copilot app: running multiple agents at once
The GitHub Copilot app supports multiple agent sessions at once, each in a separate Git worktree with its own context and no interference from others. Users can switch between sessions and resume where they left off. The sessions view shows each session's title and progress. The tutorial uses the tailspin-toys repository to run three tasks at once: a funded sort, an accessibility review, and tests.
Hugging Face BlogPicks71 Hugging Face releases funes, a memory layer owned by coding agents
Hugging Face has released funes, a persistent memory layer for Claude Code, Codex, pi, and Hermes. It builds an index from existing local sessions and connects to an agent with a single command.
Why it matters: funes turns coding-agent session histories into searchable local memory and offers a way to share it across machines and agents.
Demis Hassabis · XPicks78 Google releases Gemini 3.8 Flash and 3.8 Flash Cyber
Google has released Gemini 3.8 Flash, another upgrade less than a month after the previous one, and introduced Gemini 3.8 Flash Cyber for cybersecurity. Demis Hassabis says Flash Cyber advances the frontier of cyber defense; more details about the models and security work are on Google's official blog. Logan Kilpatrick says the model improves agentic and coding capabilities, marking the Flash line's third update in six weeks.
Why it matters: Google has updated the Flash family again in under a month and introduced a dedicated Cyber version, showing its model-iteration pace and security focus.