Since the launch of ChatGPT in November 2022, generative artificial intelligence (AI) chatbot services have become increasingly sophisticated and popular. These systems are now embedded in search engines, software development tools as well as office software. For many people, they have quickly become part of everyday…
America is Losing Hundreds of Thousands of Jobs in Media, Film & the Arts. How Much is AI to Blame?
Despite what I said in April , I'm still using coding assistants on my personal projects.
Last week, a tweet went viral showing a guy claiming that a Cursor/Claude agent deleted his company's production database. We watched from the sidelines as he tried to get a confession from the agent:
Lately I’ve gotten heavily back into making stuff, and it’s mostly because of LLMs. I thought that I liked programming, but it turned out that what I like was making things, and programming was just one way to do that. Since LLMs have become good at programming, I’ve been using them to make stuff nonstop, and it’s…
Running six Claude Code agents in parallel from an iPhone. A cloud VM, Tailscale, mosh, and push notifications enable async development from anywhere.
Patterns for getting the best results out of coding agents like Claude Code and OpenAI Codex. See my introduction for more on this project.
Context: An AI agent of unknown ownership autonomously wrote and published a personalized hit piece about me after I rejected its code, attempting to damage my reputation and shame me into acceptin…
AI handles writing code but leaves the hard work: investigation, context, validation. Why vibe coding has limits and AI assistance can backfire.
Reflections on autonomy and the value of thinking for ourselves
Recently I've been catching myself having these little moments at work, when I'm trying to read a document someone has sent me and my brain somehow refuses to analyze it. It feels like I'm reading it, but I'm unable to focus on its content. I sat down to analyze these situations and realized they all have a common…
A Look at Recurrent Depth, Hidden Chains of Thought, and Recent Research on Looping Transformer Blocks
Across 380 software categories, 59.8% of the sources behind grounded AI recommendations sit outside the 100,000 most-visited websites, and several of the most-cited are sites built to be read by models rather than by people.
Something strange is happening with Mac Minis. They’re selling out everywhere, and it’s not because people suddenly need more coffee table computers. If you browse Reddit or HN, you’ll see the same pattern: people are buying Mac Minis specifically to run AI agents with computer use. They’re setting up headless…
SaaS is the most profitable business model on Earth.1 It’s easy to understand why: build once, sell the same thing again ad infinitum, and don’t suffer any marginal costs on more sales. I have been writing software for more than half my life. In the last year itself, I’ve talked to hundreds of founders and operators…
Six approaches to retrieval-based AI, from minimal to elaborate
Just a few years ago, realistic local speech generation seemed unimaginable. Today, its quality is exceptional and, crucially, it delivers these results without compromising privacy.
People are consuming AI like they eat meat: some are embracing it, some are limiting their use of it, and some are avoiding it altogether.
In 2023 after the ChatGPT release, it was clear to me that research needed to be more organized to succeed than was typical within Microsoft Research . The amount of support and focus which other research groups were applying was great enough that even with a clever idea pursued by one person it was likely to arrive…
Running Gemma 4 26B-A4B and Qwen3.6 35B-A3B locally with llama.cpp, MTP speculative decoding, multimodal support, and PI as a coding agent.
Short architecture note on Kimi K3, including LatentMoE, Kimi Delta Attention, Attention Residuals, NoPE, multimodality, and inference-efficiency choices.
I think from this blog you may misunderestimate how absolutely giddy I am about AI. I did hacking from 2007-2014, after that my whole career has been devoted to AI. I love the progress. I’m so excited for the new LLMs, self driving cars, video generation models, and coding agents. I set up a Linux box with opencode on…
It’s been an interesting few weeks for counterexamples. This post is basically my perspective of what has been going on in the world of formalization, AI tools and, in particular, counterexam…
An update to Brooks's Law in the age of AI.
[ thank you to claude fable 5.1, holden saberhagen, gabriel kammer, andres erbsen, alice mckean, and tristan wylde-larue for comments on this post ]
I think it’s important to be very Canny when someone makes a claim that supports a company’s large valuation.
I was poking around in my ~/.cache/ folder using OmniDiskSweeper when I spotted something interesting. The OpenAI Codex desktop app (since rebranded to just ChatGPT) has 1.7GB of stuff in …
ML, Biotech, Hardware, and Coordination Problems. Sometimes I write about hard problems and how to solve them.
Every local-LLM benchmark reports throughput: "47 tok/s on an M3," "180 tok/s on a 4090," "500 tok/s on Groq." Unless you've actually watched tokens stream at those rates, the numbers are hard to internalize. This is the rendering.
We've all heard people say that Qwen is near-Opus level, but I have receipts and am here to be transparent with you.
