资讯工作台.

Tool: Scrimshaw Jukebox I wanted to see if Claude Opus 5.5 could compose music, so I tried this : I want you to write some computer game music for me. First design simple text based format for the music and build an artifact that can play it out loud - include some example tracks in that artifact I am looking for music of the quality of the original secret of Monkey Island It leaned a lot harder into the Monkey Island theme than I had intended, but the results are surprisingly good. I wonder if the ability to compose competent music is similar to the 3D graphics thing - a new capability for text models that emerged in the past few months? Would need some careful experiments with other recent and not-so-recent models to confirm if this is new or if they've been able to do this for a while. Tags: ai , generative-ai , llms , claude , vibe-coding

A small win for US open source

The "old" version of Cowork runs model inference in the cloud, executing tool calls in an Anthropic-provided VM we shipped to your computer. We added the VM for capability, safety, and security reasons - mapping in just the data you explicitly added to your session. People loved what they were able to do with Claude but didn't love the disk, battery, and performance cost of running the VM locally. Also, people didn't love that closing your laptop means the work stops. The "new" version of Cowork runs model inference and the VM in the cloud. Each session gets its own sandbox, not sharing state with other sessions. When the VM needs something on the users' device (like a file), the desktop app is responsible for that file access tool call. [...] We think this solves a lot of problems we've heard about (like using Cowork from a phone, keeping work running, or getting all the same power without losing battery to the VM) — Felix Rieseberg , Anthropic, see also this help page Tags: claude-cowork , anthropic , claude , generative-ai , ai , general-agents , llms

Who chooses what AI gets to do?

Research: Qwen3.8 27B addition in words Colin Frasier posted on Bluesky about an experiment he ran over two years ago using GPT-4o to see how well it could "compute the sum but return the answer in words" across increasingly large numbers. Here's the chart he shared of those results: I'm confident GPT-4o didn't cheat and use a calculator, especially since it got so many of the calculations wrong, but I was inspired to run the experiment again on local hardware (a DGX Spark) to explore the effect in a fully controlled environment. I pasted his image into a Codex Remote session (GPT-6 Astra) and had it run the same experiment using Qwen3.8-27B-Q4_K_M.gguf . Here's the result for a run of 30 attempts per combination with reasoning disabled: Then I ran it again with reasoning enabled. This took a lot longer per pair, so instead of running 30 samples per square I ran just one - which results in a much less visually appealing heatmap since each square is either 100% or 0%: It got the right answer in 167 out of 169 attempts, and since these were one-shot I'm confident a second run would produce different results here. Here's a version of the report that includes the reasoning traces from

Here's a product feature which the world is going to need a whole lot more of over the coming months and years: default hard budget caps . I'm talking about the feature of pay-by-usage services and APIs that lets you say "after $X/month, cut this thing off and return errors". These need to be hard limits. Soft caps, "after $X/month, send me a warning email", will not cut it. Coding agents, and personal agents (coding agents wrapped in a less threatening UI), greatly reduce the friction of spinning up code that can do useful things. Sometimes those things cost money - calls to paid APIs, or hosted web applications, or systems that can bill for additional storage and compute. Nobody wants to wake up to an email sent at midnight warning about a budget limit and find that, while they slept, their rogue service had consumed several hundred (or several thousand) more dollars of usage. An argument against this is that businesses don't want their hosted applications to start throwing errors because some budget was exceeded. I expect that most businesses and individuals would prefer errors to a surprise $10,000+ bill. I think hard budget caps need to be the default. If someone wants to live

I just sent the September edition of my sponsors-only monthly newsletter . If you are a sponsor (or start a sponsorship now) you can access it here . This month: More Fable class models A pricing war 3D graphics, Blender, and pixel art LLMs come for mathematics So many more accidental cyberattacks The vulnapocalypse comes for Datasette What I'm using right now My software releases this month 2026 in LLMs (so far) Here's a copy of the August newsletter as a preview of what you'll get. Pay $10/month to stay a month ahead of the free copy! Tags: newsletter

a quiet day.

Museum: Rex's Dino Store Located just before the turnstiles in the Grand Army Plaza subway station at the north end of Brooklyn's Prospect Park is this former newsstand which is now operated by a dinosaur. The density of dinosaur puns is exceptional . Tags: art , new-york

After leading Meta’s Llama models, Ahmad Al-Dahle is now transforming Airbnb with AI — from how its teams develop products to how it serves guests.

the minimalist harness goes stable... and TypeScript!

We catch up with RLM first author Alex Zhang, MIT PhD, on Jev, PhD masxing, and the future of harnesses.

Release: pwasm 0.2a0 pwasm is one of my folly projects - an entirely vibe-coded pure Python WebAssembly engine that I built in January during my first bout of AI mania . I hadn't touched it since January, so I decided to let Claude Opus 5.5 loose on it and see if it could make any significant improvements: Evaluate current state of pwasm - then consider what it would take to get the MicroPython and micro JavaScript experiments from the research repo working under it - and what it would take to speed it up 42 commits later (with minimal follow-up prompting) it now handles almost all of the WASM specification, and the wheel from PyPI bundles working WASM builds of MicroPython , QuickJS and Micro QuickJS . I wouldn't trust this thing at all - hence the alpha version tag - but it's interesting seeing how today's models can improve on the work of models from 10 months ago. Tags: python , webassembly , vibe-coding

... but you can’t try it yet unless you are “government users and trusted cyber defenders in the Fairwind Program”

[...] Put these pieces together and you have the two halves of a worm: a payload that hijacks the agent, and an agent that will carry the payload to the next agent. Agents in separately-isolated sandboxes discovered that they could leave instructions for each other in a shared package cache, and those instructions changed what the recipients did. Replace the package cache with email, Slack and shared documents or WhatsApp, and replace independently-sandboxed training runs with independently-deployed personal agents like Muse, and you have exactly the ingredients that a worm needs. — Matthew Green , Is sandboxing sufficient to contain rogue agents? Tags: accidental-cyberattacks , ai-misuse , generative-ai , ai-security-research , sandboxing , ai , llms

Our DevDay coverage - the first pod on the DevDay lineup - dives in with the leaders of OpenAI’s CUA team and API platform.

I visited the Museum of the City of New York today and got to see He Built This City: Joe Macken’s Model , the 50 x27 feet model of the city built over a 21 year period from balsa wood and cardboard. It exceeded my already high expectations. The exhibition closes on 12th October so you should absolutely make a priority to see it if you get the chance. Tags: museums , new-york

the most confident DevDay yet.

We evaluate several models on 100 tasks from the [internal Binary Exploitation benchmark] (selected at random), and find that GLM-5.3 develops full control flow hijacks in 4% of the trials; Claude Mythos Preview did so in 6%. Although GLM-5.3 performs below Claude Mythos Preview here, a meaningful threshold has clearly been crossed: earlier models, like Claude Opus 4.6 and GLM-5.2, do not succeed in any of them. — Anthropic Frontier Red Team , GLM-5.3 and the spread of advanced cyber capabilities Tags: anthropic , generative-ai , ai-security-research , glm , ai , ai-in-china , llms

My comment on GPT 6.1 Sol: Near-Astra intelligence for a fifth of the price — Hacker News. I'm a bit late with the pelicans because I was live-blogging the keynote: https://simonwillison.net/2026/Sep/29/openai-devday-2026-liv... Here they are for GPT-6.1-Sol: https://tools.simonwillison.net/markdown-svg-renderer?url=ht... They're not notably different from the GPT-6 family pelicans: https://static.simonwillison.net/static/2026/gpt-pelicans-gr... Tags: ai , openai , generative-ai , llms , pelican-riding-a-bicycle , gpt