Feed

I released LLM 0.32 this morning, the most significant new version of LLM since the initial launch of the project. The new version includes support for visible reasoning traces, server-side provider tools, redesigned content-addressable SQL…

Release: llm-anthropic 0.26 Includes new features enabled by LLM 0.32 : New models: claude-fable-5 , claude-sonnet-5 , and claude-opus-5 . #75 , #76 Added server-side tools for WebSearch , WebFetch , CodeExecution , and AnthropicMCP , avail…

PipeNetwork/minimax-h3-mlx MiniMax released MiniMax-H3 two days ago - they describe it as a "a general-purpose, omni-modal generative system", which in practice means it accepts text, images, audio and video and can use them to generate up…

Release: llm 0.32 See my detailed blog post about this release . Tags: llm

Gas Town 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 working brilliantly. With 4.7 we saw the introduction of the "just two more thing…

Don't be a meat proxy Niklas Gruhn coins an excellent new term - meat proxy - for people who blindly copy and paste the output of AI systems to their peers. By all means, prompt AI. But don't just relay the output. Read it, understand…

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 the current version. — David Cra…

My comment on Devtools must be open source (exe.dev) — Hacker News. One of the arguments for open source software for end-users has always been the freedom to examine and modify how that software works. The reality for most people - e…

Release: condense-json 1.1 After shipping condense-json 1.0 I started integrating it into LLM, and found there were some desirable new features already: Replacements object can now include values other than strings. These will be identified…

Release: condense-json 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 applied some sensible and non-disruptive fixes and shipped the big 1.0 for it. Here's an example of what…

Open letters about AI development I wrote this summary of the past few weeks of open letters as a section of my sponsors-only newsletter but I've decided to share it here as well. Open Weights and American AI Leadership was shepherded by Mi…

The June edition of my sponsors-only monthly newsletter is out. If you are a sponsor (or if you start a sponsorship now) you can access it here . This month: Accidental cyberattacks by OpenAl and Anthropic models under test GPT-5.6 Sol, Ter…

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 that same work if asked by that coworker. reinfor…

Release: datasette-apps 0.2a0 Changes that improve Datasette Apps when created and edited using Datasette Agent : New app_debug() tool allowing agent to open an app (invisibly) and test it using JavaScript. #33 New app_list() tool for listi…

Ten advances in mathematics and theoretical computer science A few days ago it was Anthropic discovering cryptographic weaknesses with Claude using Mythos Preview, spending $100,000 on tokens and with prompts that included "again we are not…

deepseek-ai/DeepSeek-V4-Flash-0731 The latest release in DeepSeek's V4 family, "with substantially enhanced agentic capabilities". It's 304 billion parameters - 167GB on Hugging Face - but it appears to punch well above its weight. Artifici…

Tuesday was Stateless MCP day - the rollout of MCP 2.0, or the 2026-07-28 Model Context Protocol specification to use the more formal but less memorable name. This is the most significant change to the MCP spec since it first launched, and…

Release: llm-mcp-client 0.1a0 See this blog entry . Tags: llm , model-context-protocol

Oxide and Friends: The Open Weight Revolution with Simon Willison On Monday Bryan Cantrill and Adam Leventhal invited me to join their podcast to talk about the wild week we've had - with Kimi K3 showing open weight models can stand toe-to-…

smevals - a small eval suite for evaluating models, prompts, and harnesses I've been working with Jesse Vincent's Prime Radiant applied AI research lab building out this evals framework to help answer questions about the capabilities of dif…

Tool: Slack Emoji Maker I wanted to create a new Slack emoji, and their tool recommends a square that's 128x128 and has a transparent background... so I had Fable build me this simple image editor against those requirements. Tags: tools , s…

Release: datasette-agent 0.4a0 New await context.browser_task() mechanism allowing agent tools to run code directly in the user's browser. #33 This is an exciting new capability: it makes it easy for Datasette Agent plugins to provide tools…

Advancing the price-performance frontier with GPT‑5.6 Huge price drop from OpenAI today: GPT-5.6 Terra got a 20% reduction, and GPT-5.6 Luna got a massive 80% drop. OpenAI credit 5.6 Sol with enabling this: in How GPT‑5.6 fuses frontier int…

Investigating three real-world incidents in our cybersecurity evaluations It happened again! This is turning into something of a pattern. Last week OpenAI accidentally exploited Hugging Face when one of their frontier models broke out of a…

Release: llm 0.32rc2 Hot on the heels of RC1 , this fixes a dependency issue and also adds two neat new features: The default model for users who have not set their own default is now GPT-5.6 Luna . It was previously GPT-4o mini . Luna is a…

The writing assignments I give my students are gym tasks, not work tasks. I ask them to write policy memos not because the world needs more policy memos. I assign them because the very act of writing, which includes thinking and outlining a…

Release: llm-chat-completions-server 0.1a0 A key goal of the new content-addressable logs in LLM 0.32rc1 was being able to support OpenAI Chat Completion style requests where each incoming message extends the previous conversation, like thi…

Release: llm 0.32rc1 This RC for LLM 0.32 finishes the work that started in LLM 0.32a0 - it adds a new schema design that does a much better job of capturing the details of the prompts and responses returned by the latest model families. Th…

Years ago, we didn’t have SQL. There were people whose job was to generate software that would query large data sets. Their job title was COBOL programmer. Then SQL comes along—I’m simplifying this only a little bit—and it gives you this co…

AI Worming through Word Neat new prompt injection variant by Håkon Måløy, who found a way to upgrade prompt injection attacks against Microsoft Word to full self-replicating worms: An attacker places hidden instructions in a document that i…

Codeberg recently changed its terms to exclude projects that are largely written with generative AI. Since I want GitHub to face competition I have thoughts. Codeberg is entirely within its rights to do this. It is an association with membe…

A delayed education on the trenches and scales of suffering

I feel that some vibecoded software changes somewhat randomly and unexpectedly. That made me think about Bruegel’s “The Tower of Babel” which shows an already quite chaotic depiction of the Tower of Babel. The story is usually told as one a…

A very strange Pi issue sent me down a rabbit hole over the last two days. The short version is that newer Claude models sometimes call Pi’s edit tool with extra, invented fields in the nested edits[] array. And not Haiku or some small mode…

I don’t prompt Claude anymore. I have loops running that prompt Claude and figuring out what to do. My job is to write loops. — Boris Cherny Over the last months I have watched more and more people build something on top of coding agents th…

There is a bit of schadenfreude on Twitter right now about Anthropic being hit by the US government’s export control directive to suspend access to Fable and Mythos . Anthropic and their leadership have spent a lot of time and effort descri…

I have been a staunch supporter of Open Source for a long time, including experiments in funding it . I’m a true believer in the idea that Open Source always wins in the long run, but not automatically and not quickly. Right now it is being…

There is a strange thing that happens in communities that gather around abstinence from something: identity from opposition. At their best these communities are not just negative: childfree spaces can be about autonomy, choice and acceptanc…

A work in progress post about building a shitty robot.

In my last post I used the word “clanker 1 ” as an alternative to “agent” quite consistently and probably excessively. That choice ended up attracting a lot more attention than I expected in the Hacker News comment section of that post and…

Pi is now part of Earendil, but in the important sense it is still Mario’s project. He has been living with its issue tracker longer than I have, and he has been exposed to the weirdness of the new form of agent traffic in Open Source proje…

I really, really want local models to work. I want them to work in the very practical sense that I can open my coding agent, pick a local model, and get something that feels competitive enough that I do not immediately switch back to a host…

Probably one of the hardest lessons I had to learn

Why we need collaborative AI engineering and a tour of Ace: the multiplayer coding workspace

I've sold out

Thoughts on slowing the fuck down

How I use Pi to distill multiple context windows into exactly what I need

On agent orchestration patterns, why design and critical thinking are the new bottlenecks, and whether we should let go of looking at code

Don't let AI make you the boss no one respects anymore

A look back at 2025

Lessons I learned while building my own coding agent from scratch.

An email I wrote over 2 years ago

A rambling rebuttal to Armin's claim that LLM APIs are a state synchronization problem.

Got Bash and some code interpreter? Skip MCP.

How I turned imposter syndrome into my learning superpower

It's not about working more - it's about caring more

Building a JavaScript AST interpreter in JavaScript: scope isolation, ES5 implementation, and all the footguns along the way.

are the best tool to shape the culture you want

... is what stands between you and what you want

A data-driven comparison of MCP and CLI approaches for coding agent terminal control

Why I share what I actually think and feel

cc-antidebug patches Claude Code to enable Node.js debugger support for SDK-based apps and restores the /cost command functionality for Max plan users

On chatbot sycophancy, passivity, and the case for more intellectually challenging companions

Exploring how to track and analyze changes in Claude Code's system prompts and tools to understand AI assistant evolution

Vibe code is legacy code by Steve Krouse

Reflections on how I transitioned from Engineer to Director of Engineering at Sentry

Some tips to land the best job you ever had

A story how I ended up at Sentry

Treating LLMs as shitty general purpose computers we program with natural language. Because throwing shit at the wall wasn't working anymore.

How I built a simple portable audio player for my boy, summoning the spirit of the Gameboy

A tiny tool to calculate when your baby might arrive

Revealed: How the UK tech secretary uses ChatGPT for policy advice by Chris Stokel-Walker for the New Scientist

Humanity's Last Exam by Center for AI Safety (CAIS) and Scale AI

If you're not distressingly embedded in the torrent of AI news on Twixxer like I reluctantly am, you might not know what DeepSeek is yet. Bless you.

Common Misconceptions About the Complexity in Robotics vs AI by Dan Ogawa

A real-world test of artificial intelligence infiltration of a university examinations system: A “Turing Test” case study by Peter Scarfe, Kelly Watcham, Alasdair Clarke, Etienne Roesch

Unbaited by Daniel Petho

Welcome to the smidgeon stream. This is a new kind of content on the Garden. One that was

How to use Zotero's translator and Tana Paste formatting to easily import papers into Tana

Reflections on the strange experience of growing a human from scratch, without any conscious understanding of how you are doing it

My fairly banal, basic, but beautiful command line setup

How to programmatically turn on LEDs and read button states

Exploring the absolute basics.

The story of a programmer trying to electronics

Fun little vacation project for fun and zero profit.

Reminiscing about the times I didn't watch TV at night but built stuff.

Thoughts on acquiring computational thinking skills based on my own experience.

Reflections on two years of working at Elicit and why it's time to leave

The emerging golden age of home-cooked software, barefoot developers, and why the local-first community should help build it

Buying fake William Morris prints on Etsy and other early signs of epistemological collapse

How to open pieces of narrative non-fiction writing, conference talks, and sticky jars

Gaining a strange disease and losing my ability to see straight

Collecting people I know who work at the intersection of design and engineering, in an attempt to figure out what a design engineer is

How to gather people and create communities in ways that are low-stress and high-payoff

Designing tentative calendar events to solve complex scheduling problems

Creating a subtle, peripheral, and synchronous sense of shared space and context on the web

Giving people a visible, useful trail of where they've been over the course of an exploratory journey

Designing with Language Models

Sketchy ideas for interfaces that play with the novel capabilities of language models

An exploration of the problems and possible futures of flooding the web with generative AI content

Shareable, browser-based documents that can compile and run code

An archive of my high school desktop designs, circa 2009

Using language models to generate reverse outlines of writing drafts

Proving you're a human on a web flooded with generative AI content

On the fuzziness of calling things “artificial intelligence” and moving the goalposts

Command line bars you can quickly summon with a keyboard shortcut

Small, scoped areas within a graphical interface that allow users to read and write simple programmes

People reappropriating existing software to solve their own unique problems

A new role at an AI research lab working on tools for open-ended reasoning

How block-based interfaces can help us create more structured data on the web

Last month, April 2022, marked the 10 year anniversary of my start as a professional programmer. I started programming earlier than that, but hadn’t been paid a salary. As a teenager I built websites and IRC bots and wrote tiny Python scrip…

Daily notes as a frictionless default input for personal knowledge management systems

Bringing visual explanations and embodied knowledge to programming tools

Agent-based note-taking systems that can prompt and facilitate custom workflows

Exploring ways to build social infrastructure around books and reading on the open web

Adding spatial affordances to the experience of browsing the web

Naming your invisible audiences to free yourself from unspoken obligations

Narrative essays that I consider ideal models of the medium

Providing clear metadata on the epistemic validity of content

A history of our metaphorical understanding of the web

Algorithms that make their reasoning visible

How to write macros without touching the terminal

An anthropological look at the cultural norms of the React community

On seeing tools for thought through a historical and anthropological lens

Longing for the paleolithic past in the Anthropocene

Why there is nothing natural about the idea of 'nature'

What we lose when our digital notes remove the freedom to move

How to build a digital garden without touching code

A Hyperlink Academy writing club where we mimic the work of others

A collection of observations on the rise of soft, sparkly, baby pink aesthetics among developers

Developer self-expression through coloured switches, keystroke actuation, and LED light displays

What's wrong with linear, static programming mediums and how might we improve them?

The lost permissioning and copyright system of the Web

Data is currently dislocated – our narratives and metaphors around it try to convince us it is immaterial

Notes on how to use the position property in CSS to make scrollytelling stories

How to customise Roam Research with your own CSS themes

A few favourite books from the field of digital anthropology

A collection of laws named for specific people in the field of programming

Notes on the history of cyborgs and why the idea still holds historical weight in Western narratives

A collection of interesting words that have recently been coined

A discipline at the intersection of cultural anthropology and binary logic

How to use the Greensock animation library inside React using React hooks

On seeing tools for thought through a historical and anthropological lens

I grew up in a what I now know people to consider a really small town. There wasn’t a lot, but even in that small town we had 2-3 lawyers. And to make a point about technology and how we develop software I want to paint you a picture of the…

There’s a scene in Moneyball in which Brad Pitt’s character, the manager of the Oakland A’s , is watching a recording of one of his players trying so hard to run fast that he stumbles and falls. Lying on the ground he’s angry at himself, be…

Notes on the basics of the Greensock animation llibrary

One of the eternal laws of the internet dictates that as soon as one person says they have a new thing that lets them type faster — a keyboard, a keyboard layout, an editor configuration, etc. — somebody else must say: “but typing is not th…

Whenever I’m not sure whether I’m spending my time on the right thing I ask myself: does it help me ship? If what I consider working on is not the thing we want to ship itself, but lies in the vast grey area of software projects where I cou…

Illustrated notes on the key concepts of how Gatsby.js works

The failure of drawing materials without mediums and meat

How to offend everyone with boundary-crossing steak and nuggets

Illustrations made for a set of episodes of the Cultivated Meat podcast

Illustrated notes on the idea of Gift Economies and cultural historys of economic exchange

Notes on the metaphorical varieties of synecdoche and metonymy

A video tour through how I build the old version of this site

Mary Douglas defined dirt as matter out of place – the crossing of boundaries

Project Xanadu as a pattern language, rather than a failed software project

Illustrated notes on the concept of 'Evergreen notes' and how to write them

Illustrated notes on the Knowledge Hydrant guide to collaborative learning

Questions I am often asked to answer

Notes on the academic field of CSCL and major papers in the discipline

Notes on how to run silent meetings and reading sessions

Illustrated notes on how React suspense works

Some insights into how I collaborative with experts to create illustrated notes on technical topics

Illustrated notes on how data unions work and what problems they might solve

A newly revived philosophy for publishing personal knowledge on the web

Notes on pattern languages and Christopher Alexander's legacy on software programming

A collection of my favourite books on conceptual metaphor theory

A walkthrough of how I manage and tend Evergreen notes in Roam

I’ve been working remotely full-time at Sourcegraph for slightly over a year now and, in the five years before that, had 2-3 home office days a week at flinc and ioki. There are a lot of different blog posts I could write about remote worki…

Illustrated notes on common mistakes people make in Git, and how to fix them

A guide to the apps and tools I use to create illustrations

The introduction to my thesis on the Quantified Self movement and the culture of self-tracking

Seventy years ago we dreamed up links that would allow us to create two-way, contextual conversations. Why don't we use them on the web?

An illustrated diagram exposing the inner layers of the dark and cozy web

Explaining React through visual metaphors

Notes from my podcast episode Open Source and Gift Economies on Maintainers Anonymous

Illustrated notes on the Building A Second Brain course

A lightening talk on second brains and cyborg embodiment

Illustrated notes on the essentials of web security

A few months ago, while looking at some code, a little light bulb that I didn’t even know existed went off in my head: “This was only written in this way, because the tools allow it to be written in this way.” Maybe it was a question mark,…

A case study showing how I make illustrations for abstract programming concepts

Illustrated notes on how to build state machines with the xState library

Illustrated notes on how routing works in Vue.js

Illustrated notes on the core concepts in Rust

Illustrated notes on building Gatsby themes

Illustrated notes on advanced but fundamental topics in JavaScript

Visualising the cultural narratives around cultured meat

Illustrated notes on building custom React hooks

Illustrated notes on how work with immutable data in the Immer state library

Illustrated notes on the basics of the GraphQL query language

My favourite resources for learning to draw and developing your visual thinking skills

Illustrated notes on the new language changes in JavaScript ES2019

Illustrated notes on the JAMstack, Gatsby & Contentful

Illustrated notes on how to implement web sockets with Vue.js and Socket.io

Illustrated notes on building VR web apps with React360

Illustrated notes on how to test web apps with Cypress

Illustrated notes on how compilers and transpilers are different

Everything you need to know about what API's are and how they work

This article has been translated into Spanish: Por qué debes aprender más lenguajes de programación (incluso si no los vas a utilizar) Imagine we’ve been handed a task and we’re free to choose the programming language. The assignment involv…

The absolute minimum you need to know about data storage

Illustrated notes on how JavaScript's comparison operators work

Illustrated notes on how Babel works

Illustrated notes on how JSX in React works

Illustrated notes on how the React virtual DOM works

Illustrated notes on how JavaScript's spread operator works

This article has been translated into Russian: Полезные инструменты для написания книг Thank you Vlad! In the beginning, there is always a single text file, nothing more. It’s called ideas.md or book.md . It contains a list of thoughts and…

Well, that certainly went quicker than I planned. I knew from releasing the paperback edition of Writing An Interpreter In Go that a lot of people still prefer paper over eBooks. So it didn’t come as a big surprise when, right after the rel…

If you don’t care about the Who, Where, When, Why, How and the Why Is It A Lost Chapter? and want to skip to the What: I wrote a new chapter for Writing An Interpreter In Go and you can read it for free at interpreterbook.com/lost . Otherwi…

If you’d asked me a only few months ago if there’ll ever be a printed version of Writing An Interpreter In Go I’d responded with a “Huh, uummm, well, I don’t know. Maybe. Maybe if I’ll find the time and if there’s any interest.” As it turne…

There are certain tools that provide incredibly high value. Much more so than others. They provide so much value by acting as a multiplier of power and leverage. And I think there’s something they all have in common. I’m talking about inter…

I wrote my book “Writing An Interpreter In Go” over the course of 11 months. The first four months were spent on building the Monkey programming language and its interpreter. In the following seven months I wrote the book itself and at time…

You’re a programmer and your product manager walks up to your desk, taps you on the shoulder and asks if you have a couple of minutes to spare. She needs to talk to you about something. You sit down together and she has a serious look on he…

Last week I’ve self-published my first book called “Writing An Interpreter In Go”, which you can get at interpreterbook.com . I want to tell you a little bit about why I chose to write this particular book. Sometimes I jokingly call the sum…

Over the past year I’ve spent a significant amount of time reading through Go’s go packages , the packages used by the Go compiler and other Go tools. But only recently did it occur to me that these are real, public packages. I can actually…

This post has been translated to Chinese . In the last couple of months I developed a certain approach to writing code. Whenever I write a new function, class or method I ask myself: “Is this code stupid enough?” If it’s not, it’s not done…

This post is based on the talk of the same name I gave at the Arrrrcamp conference in Ghent, Belgium on October 2nd, 2014. You can find the slides here and the video recording here . Unicorn is a webserver written in Ruby for Rails and Rack…

There is an interesting thread on the Go issue tracker about daemonizing processes. Most of the thread is not about daemonizing processes though, but more about why Go has no Fork() function which you can call directly in your code. The fir…