2026-04-15
PyTorch implementation of VQ-VAE by Aäron van den Oord et al.
Diagnose your Claude Code sessions
The example prompt for Google's new Gemini Flash TTS text-to-speed model is a lot simonwillison.net/2026/Apr/15/...
On my way to give a talk at CNN’s NYC headquarters. Taking the opportunity to wear a niche AI humor t-shirt that probably only made sense in 2016
Hey computer science faculty peeps! Are we prepared for the near future where every high school student and incoming college freshman has vibe-coded an AI agent as high school “research”? Exciting. And scary. We are going to need to update our priors.
I'm late to the game -- I only recently discovered @techtrenches.dev Highly recommended reading!
I spent some time trying to distill all the complex factors impacting open models -- economics, capabilities, distribution, policy, etc. -- into a clear list of beliefs. Here they are in full. www.interconnects.ai/p/my-bets-on...
We are hiring Software Engineers in Tokyo to help us scale Sakana AI’s R&D efforts. If you are interested in building the data pipelines and full stack infrastructure needed to push the boundaries of automated scientific discovery, we would love to hear from you. 🗼🎌 sakana.ai/careers/#sof...
Instead of the gold standard, we can, as a thought experiment, imagine an inference standard of exchange, the FLOP. (As opposed to tokens, this accounts for AI ability) With some AI help, I figure $1 buys roughly 10^17 managed-LLM inference FLOPs So that $4 coffee would cost half an exaFLOP, choom
This is becoming a pattern in AI that makes talking about capabilities challenging. First, there are overstated claims (like the flubbed Erdos problems that were announced last year), then minor wins (AI helps with discovery) then breakthroughs. The first stage feels like (& often is) hype, but…
Last year, someone (specifically, OUP) asked me to write an encyclopedia entry for "AI". I've just finished reviewing the copy edits, so hopefully it will be in the world soon. Meanwhile, a teaser: >>
The long legacy of simulation in control theory and what it can teach us about transferring policies from GPU to reality.
Feeling FOMO that I can't be at #CHI2026 this year but please check out all the great work that our @socialfutureslab.bsky.social + friends are presenting (see below for paper links). And say hi to @kjfeng.me @aliciaguo.com, Katie Yurechko, and Tony Zhou who are at the conference!
What I expect to come next and why, focused on the open-closed gap.