My blog uses aggressive caching: it sits behind Cloudflare with a 15 minute cache header, which guarantees it can survive even the largest traffic spike to any given page. I've recently added a couple of dynamic features that work in spite of that full-page caching. Here's h...
The Five Levels: from Spicy Autocomplete to the Dark Factory
Dan Shapiro proposes a five level model of AI-assisted programming, inspired by the five (or rather six, it's zero-indexed) levels of driving automation.
Spicy autocomplete, aka original GitHub Copilot or copying...
One Human + One Agent = One Browser From Scratch
embedding-shapes was so infuriated by the hype around Cursor's FastRender browser project - thousands of parallel agents producing ~1.6 million lines of Rust - that they were inspired to take a go at building a web browser usi...
Kimi K2.5: Visual Agentic Intelligence
Kimi K2 landed in July as a 1 trillion parameter open weight LLM. It was joined by Kimi K2 Thinking in November which added reasoning capabilities. Now they've made it multi-modal: the K2 models were text-only, but the new 2.5 can handl...
Someone asked on Hacker News if I had any tips for getting coding agents to write decent quality tests. Here's what I said:
I work in Python which helps a lot because there are a TON of good examples of pytest tests floating around in the training data, including things lik...
One of my favourite features of ChatGPT is its ability to write and execute code in a container. This feature launched as ChatGPT Code Interpreter nearly three years ago, was half-heartedly rebranded to "Advanced Data Analysis" at some point and is generally really difficult...
the browser is the sandbox
Paul Kinlan is a web platform developer advocate at Google and recently turned his attention to coding agents. He quickly identified the importance of a robust sandbox for agents to operate in and put together these detailed notes on how the web br...
Kākāpō Cam: Rakiura live stream
Critical update for this year's Kākāpō breeding season: the New Zealand Department of Conservation have a livestream running of Rakiura's nest!
You’re looking at the underground nest of 23-year-old Rakiura. She has chosen this same site to ne...
Don't "Trust the Process"
Jenny Wen, Design Lead at Anthropic (and previously Director of Design at Figma) gave a provocative keynote at Hatch Conference in Berlin last September.
Jenny argues that the Design Process - user research leading to personas leading to user journ...
If you tell a friend they can now instantly create any app, they’ll probably say “Cool! Now I need to think of an idea.” Then they will forget about it, and never build a thing. The problem is not that your friend is horribly uncreative. It’s that most people’s problems are not software-shaped, and most won’t notice even when they are. [...]
Programmers are trained to see everything as a software-shaped problem: if you do a task three times, you should probably automate it with a script. Rename every IMG_*.jpg file from the last week to hawaii2025_*.jpg, they tell their terminal, while the rest of us painfully click and copy-paste. We are blind to the solutions we were never taught to see, asking for faster horses and never dreaming of cars.
Last week Cursor published Scaling long-running autonomous coding, an article describing their research efforts into coordinating large numbers of autonomous coding agents. One of the projects mentioned in the article was FastRender, a web browser they built from scratch usi...
[...] i was too busy with work to read anything, so i asked chatgpt to summarize some books on state formation, and it suggested circumscription theory. there was already the natural boundary of my computer hemming the towns in, and town mayors played the role of big men to drive conflict. so i just needed a way for them to fight. i slightly tweaked the allocation of claude max accounts to the towns from a demand-based to a fixed allocation system. towns would each get a fixed amount of tokens to start, but i added a soldier role that could attack and defend in raids to steal tokens from other towns. [...]
SSH has no Host header
exe.dev is a new hosting service that, for $20/month, gives you up to 25 VMs "that share 2 CPUs and 8GB RAM". Everything happens over SSH, including creating new VMs. Once configured you can sign into your exe.dev VMs like this:
ssh simon.exe.dev
Here...
Most people's mental model of Claude Code is that "it's just a TUI" but it should really be closer to "a small game engine".
For each frame our pipeline constructs a scene graph with React then:
-> layout elements
-> rasterize them to a 2d screen
-> diff that against the previous screen
-> finally use the diff to generate ANSI sequences to draw
We have a ~16ms frame budget so we have roughly ~5ms to go from the React scene graph to ANSI written.
Qwen3-TTS Family is Now Open Sourced: Voice Design, Clone, and Generation
I haven't been paying much attention to the state-of-the-art in speech generation models other than noting that they've got really good, so I can't speak for how notable this new release from Qwen is.
...
Most people's mental model of Claude Code is that "it's just a TUI" but it should really be closer to "a small game engine".
For each frame our pipeline constructs a scene graph with React then
-> layouts elements
-> rasterizes them to a 2d screen
-> diffs that against the previous screen
-> finally uses the diff to generate ANSI sequences to draw
We have a ~16ms frame budget so we have roughly ~5ms to go from the React scene graph to ANSI written.
Claude's new constitution
Late last year Richard Weiss found something interesting while poking around with the just-released Claude Opus 4.5: he was able to talk the model into regurgitating a document which was not part of the system prompt but appeared instead to be baked...
Previous work estimating the energy and water cost of LLMs has generally focused on the cost per prompt using a consumer-level system such as ChatGPT.
Simon P. Couch notes that coding agents such as Claude Code use way more tokens in response to tasks, often burning through many thousands of tokens of many tool calls.
As a heavy Claude Code user, Simon estimates his own usage at the equivalent of 4,400 "typical queries" to an LLM, for an equivalent of around $15-$20 in daily API token spend. He figures that to be about the same as running a dishwasher once or the daily energy used by a domestic refrigerator.
Detailed and thoughtful description of an open-book and open-chatbot exam run by Ploum at École Polytechnique de Louvain for an "Open Source Strategies" class.
Students were told they could use chatbots during the exam but they had to announce their intention to do so in advance, share their prompts and take full accountability for any mistakes they made.
Only 3 out of 60 students chose to use chatbots. Ploum surveyed half of the class to help understand their motivations.
jordanhubbard/nanolang
Plenty of people have mused about what a new programming language specifically designed to be used by LLMs might look like. Jordan Hubbard (co-founder of FreeBSD, with serious stints at Apple and NVIDIA) just released exactly that.
A minimal, LLM-frie...
Scaling long-running autonomous coding
Wilson Lin at Cursor has been doing some experiments to see how far you can push a large fleet of "autonomous" coding agents:
This post describes what we've learned from running hundreds of concurrent agents on a single project, coordi...
FLUX.2-klein-4B Pure C Implementation
On 15th January Black Forest Labs, a lab formed by the creators of the original Stable Diffusion, released black-forest-labs/FLUX.2-klein-4B - an Apache 2.0 licensed 4 billion parameter version of their FLUX.2 family.
Salvatore Sanfilipp...
[On agents using CLI tools in place of REST APIs] To save on context window, yes, but moreso to improve accuracy and success rate when multiple tool calls are involved, particularly when calls must be correctly chained e.g. for pagination, rate-limit backoff, and recognizin...
Our approach to advertising and expanding access to ChatGPT
The long-rumored introduction of ads to ChatGPT just became a whole lot more concrete:
In the coming weeks, we’re also planning to start testing ads in the U.S. for the free and Go tiers, so more people can benefit...
Open Responses
This is the standardization effort I've most wanted in the world of LLMs: a vendor-neutral specification for the JSON API that clients can use to talk to hosted LLMs.
Open Responses aims to provide exactly that as a documented standard, derived from OpenAI's R...
The Design & Implementation of Sprites
I wrote about Sprites last week Here's Thomas Ptacek from Fly with the insider details on how they work under the hood.
I like this framing of them as "disposable computers":
Sprites are ball-point disposable computers. Whatever ma...
When we optimize responses using a reward model as a proxy for “goodness” in reinforcement learning, models sometimes learn to “hack” this proxy and output an answer that only “looks good” to it (because coming up with an answer that is actually good can be hard). The philos...
Claude Cowork defaults to allowing outbound HTTP traffic to only a specific list of domains, to help protect the user against prompt injection attacks that exfiltrate their data.
Prompt Armor found a creative workaround: Anthropic's API domain is on that list, so they constructed an attack that includes an attacker's own Anthropic API key and has the agent upload any files it can see to the https://api.anthropic.com/v1/files endpoint, allowing the attacker to retrieve their content later.
Anthropic invests $1.5 million in the Python Software Foundation and open source security
This is outstanding news, especially given our decision to withdraw from that NSF grant application back in October.
We are thrilled to announce that Anthropic has entered into a two-y...
Featured
