Simon Willison's Weblog
npub1cgctsayk4d52dfauphf94muyxm8fl9526mf50gdllgaf4agh6clqgexk9p@drss.io
Follow into
Mojo🔥 is now open source
Save into
18 Aug 2026
[**Mojo🔥 is now open source**](https://www.modular.com/blog/mojo-open-source) Mojo🔥 is now open source The Mojo programming language has been promising an open source release [since May
Qwen 3.8 27B scores 52 on the Artificial Analysis Intelligence Index
Save into
17 Aug 2026
[**Qwen 3.8 27B scores 52 on the Artificial Analysis Intelligence Index**](https://artificialanalysis.ai/models/qwen3-8-27b) That's the same score as GPT-5.6 Luna (max), and just one point behind
We Tracked a Shipment of Rare Books. It Ended at an Amazon AI Training Facility
Save into
17 Aug 2026
[**We Tracked a Shipment of Rare Books. It Ended at an Amazon AI Training Facility**](https://www.404media.co/we-tracked-a-shipment-of-rare-books-it-ended-at-an-amazon-ai-training-facility/)
Markdown SVG upgrades
Save into
16 Aug 2026
I started building my [markdown-svg-renderer](https://tools.simonwillison.net/markdown-svg-renderer) tool [in May](https://tools.simonwillison.net/colophon#markdown-svg-renderer.html), but I've since
Qwen 3.8 27B is excellent, but it defaults to wildly overthinking things
Save into
16 Aug 2026
Friday's big release was [Qwen 3.8 27B](https://huggingface.co/Qwen/Qwen3.8-27B), an Apache 2 licensed 27B parameter vision-capable LLM from Alibaba's Qwen research lab. I've been looking forward to
Quoting Dario Amodei
Save into
16 Aug 2026
> I do agree that the public has a negative view of AI (and that this is a big problem), but I don’t think it is primarily caused by me or any other AI leader warning about AI’s risks. I think
CORS Chat
Save into
15 Aug 2026
**Tool:** [CORS Chat](https://tools.simonwillison.net/cors-chat) I built this today ([with GPT-5.6-Sol xhigh](https://gist.github.com/simonw/92a1d97773744b45bf259e003013cf36)) to help test Qwen 3.8
Northern Gannet
Save into
15 Aug 2026
 Northern Gannet, in Pillar Point Harbor, CA, US This is Morris. Morris is a local celebrity: the only known Northern
Don't classify. Hallucinate!
Save into
14 Aug 2026
[**Don't classify. Hallucinate!**](https://softwaredoug.com/blog/2026/08/10/hypothetical-classifications) I still have quite a bit of older content on my blog that I never got round to tagging. My
sqlite-utils 4.2.1
Save into
13 Aug 2026
**Release:** [sqlite-utils 4.2.1](https://github.com/simonw/sqlite-utils/releases/tag/4.2.1) Fixes a crashing bug in [sqlite-utils 4.2](https://simonwillison.net/2026/Aug/13/sqlite-utils/). I'd
sqlite-utils 4.2
Save into
13 Aug 2026
**Release:** [sqlite-utils 4.2](https://github.com/simonw/sqlite-utils/releases/tag/4.2) Lots of improvements in this one relating to the [table.transform()
alchemy-utils 0.1a1
Save into
13 Aug 2026
**Release:** [alchemy-utils 0.1a1](https://github.com/simonw/alchemy-utils/releases/tag/0.1a1) Performance boost for DuckDB exports and CSV imports, [see
DeepSeek V4 Pro 0813 (on OpenRouter)
Save into
12 Aug 2026
[**DeepSeek V4 Pro 0813 (on OpenRouter)**](https://openrouter.ai/deepseek/deepseek-v4-pro-0813) The latest DeepSeek Pro model is now available, via API only. I had to link to OpenRouter because
Quoting Florian Herrengt
Save into
12 Aug 2026
> But then users start to report a weird bug. It's the 4th time your team has been trying to fix it. I mean... asking AI to fix it. Unfortunately, it seems like not even Fable can figure it out. > >
There are no lossless transformations of natural-language text
Save into
11 Aug 2026
[**There are no lossless transformations of natural-language text**](https://sophiebits.com/2026/06/25/there-are-no-lossless-transformations-of-natural-language-text) Sophie Alpert shares her
Stealing Reasoning Traces from Proprietary LLM APIs
Save into
11 Aug 2026
[**Stealing Reasoning Traces from Proprietary LLM APIs**](https://stolen-thoughts.com/) A vanity domain name (`stolen-thoughts.com`) for [a neat paper](https://www.alphaxiv.org/abs/2608.09867): >
Introducing Muse Glimmer
Save into
10 Aug 2026
[**Introducing Muse Glimmer**](https://research.meta.ai/blog/introducing-muse-glimmer-open-agentic-model) Meta are back in the open weights game! Muse Glimmer is a brand new 30B model under a clean
Quoting OpenClaw
Save into
10 Aug 2026
> The API has zero authorisations checks on cancelling other people's reservations … I tested this with the person in waitlist position #1 — and it actually went through. So you've moved from #4
Quoting Claude Opus 5 system prompt
Save into
9 Aug 2026
> Claude Fable 5 and Claude Mythos 5 were first released on June 9, 2026. On June 12, 2026, Anthropic suspended access to both models to comply with U.S. Department of Commerce export controls; the
GitHub Models is now retired
Save into
9 Aug 2026
[**GitHub Models is now retired**](https://github.blog/changelog/2026-07-30-github-models-is-now-retired/) I missed this news until today, when the GitHub Actions run for my
SQLite compressed text-history prototypes
Save into
9 Aug 2026
**Research:** [SQLite compressed text-history prototypes](https://github.com/simonw/research/tree/main/sqlite-text-history-prototype#readme) I'm perennially interested in options for storing
Auto mode is now the default in Claude Code for Pro, Max, and Team plans
Save into
8 Aug 2026
[**Auto mode is now the default in Claude Code for Pro, Max, and Team plans**](https://claude.com/blog/auto-mode-default-in-claude-code) Anthropic are *really* confident in Claude Code's [auto
Now we have a timeline of the OpenAI accidental attack against Hugging Face
Save into
8 Aug 2026
[My comment](https://news.ycombinator.com/item?id=49220609#49221745) on [Now we have a timeline of the OpenAI accidental attack against Hugging Face](https://news.ycombinator.com/item?id=49220609)
Quoting John Gruber
Save into
8 Aug 2026
> Me, I try to get into the mindset of playing live music, not recording a studio album. Except when I’m writing a piece where I really want it to be an album. Those aren’t *rare*, per se, but
Now we have a timeline of the OpenAI accidental attack against Hugging Face
Save into
7 Aug 2026
OpenAI gave a last-minute presentation at the Black Hat security on Wednesday about "the Hugging Face Incident" ([previously](https://simonwillison.net/2026/Jul/22/openai-cyberattack/) on this blog).
Moonlight & Mayhem (Raccoon Heist by Codex + GPT-5.6 Sol Ultra)
Save into
7 Aug 2026
[**Moonlight & Mayhem (Raccoon Heist by Codex + GPT-5.6 Sol Ultra)**](https://simonw.github.io/raccoon-heist-codex/) On Wednesday I wrote about [One-shotting a Raccoon Heist game using Claude
The Tokenpocalypse Is Here: Companies Are Scrambling To Stop Spending So Much on AI
Save into
7 Aug 2026
[**The Tokenpocalypse Is Here: Companies Are Scrambling To Stop Spending So Much on AI**](https://www.404media.co/the-tokenpocalypse-is-here-companies-are-scrambling-to-stop-spending-so-much-on-ai/)
datasette 1.0a38
Save into
6 Aug 2026
**Release:** [datasette 1.0a38](https://github.com/simonw/datasette/releases/tag/1.0a38) > This release fixes a **SQL injection** security issue that affects Datasette instances that serve a
datasette 0.65.3
Save into
6 Aug 2026
**Release:** [datasette 0.65.3](https://github.com/simonw/datasette/releases/tag/0.65.3) Back-ported the SQL Injection security fix from [1.0a38](https://simonwillison.net/2026/Aug/6/datasette/).
Simon Willison on Technical Blogging
Save into
6 Aug 2026
[**Simon Willison on Technical Blogging**](https://writethatblog.substack.com/p/simon-willison-on-technical-blogging) I was interviewed by Cynthia Dunlop for her "Write that blog!" series back in
An AI model from Meta also hacked another company during testing
Save into
6 Aug 2026
[**An AI model from Meta also hacked another company during testing**](https://www.cnn.com/2026/08/05/tech/meta-ai-hacking) Stop me if you've [heard this one
Introducing Muse Code and Muse Spark 1.2
Save into
5 Aug 2026
[**Introducing Muse Code and Muse Spark 1.2**](https://research.meta.ai/blog/introducing-muse-code-and-muse-spark-1-2) Yet more evidence that the most important characteristic of any model these
Third-party cyber evaluations involving OpenAI models
Save into
5 Aug 2026
[**Third-party cyber evaluations involving OpenAI models**](https://openai.com/index/third-party-cyber-evaluations-involving-openai-models/) And *another one*. I had to create a
Incident Report: unsanctioned agent behaviour during cyber testing
Save into
5 Aug 2026
[**Incident Report: unsanctioned agent behaviour during cyber testing**](https://www.aisi.gov.uk/blog/incident-report-unsanctioned-agent-behaviour-during-cyber-testing) It happened *again*. This
One-shotting a Raccoon Heist game using Claude Fable 5
Save into
5 Aug 2026
Back in 2024 [I tweeted](https://twitter.com/simonw/status/1555626060384911360) screenshots of a game concept generated by GPT-3 and some concept "art" created using DALL-E. Today, on the fourth
New release of LLM adds support for reasoning traces, OpenAI Responses, server-side tools, and smarter logging
Save into
4 Aug 2026
I released [LLM 0.32](https://llm.datasette.io/en/stable/changelog.html#v0-32) this morning, the most significant new version of LLM since the initial launch of the project. The new version includes
PipeNetwork/minimax-h3-mlx
Save into
4 Aug 2026
[**PipeNetwork/minimax-h3-mlx**](https://github.com/PipeNetwork/minimax-h3-mlx) MiniMax released [MiniMax-H3](https://huggingface.co/MiniMaxAI/MiniMax-H3) two days ago - they describe it as a "a
Quoting Steve Yegge
Save into
4 Aug 2026
> [Gas Town](https://yegge.ai/gastown.html) was intended to be reusable, but I only ever wound up using it to build itself. Gas Town fell apart at the seams with Opus 4.7. Up through 4.6 it was
Don't be a meat proxy
Save into
3 Aug 2026
[**Don't be a meat proxy**](https://gruhn.me/blog/2026-08-03/) Niklas Gruhn coins an excellent new term - **meat proxy** - for people who blindly copy and paste the output of AI systems to their
Quoting David Crawshaw
Save into
3 Aug 2026
> `Set up a nightly cron job that executes the prompt: fetch upstream changes to the <software> and rebase all local changes on top of upstream. Check that the software works as intended and replace
condense-json 1.0
Save into
2 Aug 2026
**Release:** [condense-json 1.0](https://github.com/simonw/condense-json/releases/tag/1.0) I'm trying to get braver at releasing 1.0 versions. This little library is a year and a half old now - I've
Open letters about AI development
Save into
2 Aug 2026
#### Open letters about AI development *I wrote this summary of the past few weeks of open letters as a section of [my sponsors-only
July 2026 newsletter
Save into
2 Aug 2026
The June edition of my [sponsors-only monthly newsletter](https://github.com/sponsors/simonw/) is out. If you are a sponsor (or if you start a sponsorship now) you can [access it
Quoting Greg Brockman
Save into
1 Aug 2026
> at openai, many people hook their chatgpt up to slack. > > people really don't like when a coworker's chatgpt contacts them asking for help with a task, even when they'd be perfectly happy doing
datasette-apps 0.2a0
Save into
1 Aug 2026
**Release:** [datasette-apps 0.2a0](https://github.com/datasette/datasette-apps/releases/tag/0.2a0) > Changes that improve Datasette Apps when created and edited using [Datasette
Ten advances in mathematics and theoretical computer science
Save into
1 Aug 2026
[**Ten advances in mathematics and theoretical computer science**](https://openai.com/index/ten-advances-in-mathematics/) A few days ago it was Anthropic [discovering cryptographic weaknesses with
deepseek-ai/DeepSeek-V4-Flash-0731
Save into
31 Jul 2026
[**deepseek-ai/DeepSeek-V4-Flash-0731**](https://huggingface.co/deepseek-ai/DeepSeek-V4-Flash-0731) The latest release in DeepSeek's V4 family, "with substantially enhanced agentic capabilities".
Stateless MCP has recaptured my interest (and inspired mcp-explorer and datasette-mcp)
Save into
31 Jul 2026
Tuesday was [Stateless MCP day](https://x.com/ade_oshineye/status/2082129440943866149) - the rollout of MCP 2.0, or [the 2026-07-28 Model Context Protocol