Posts
2026
On computation, quantum mechanics, the structure of difficulty, and what a universe that allowed shortcuts would have to look like
I just have to wait out you
The cost of continual learning scales with generality
Claude's Constitution
Ithaka, California
It will drive you mad (and it's boring)
I'm going to be better this year now I'm not a founder
2025
From San Francisco
A letter to the C-suite
Scaling care beats scaling compute
For people who care about falsifiable stakes rather than vibes
The Case for Inheriting Nothing and Building Everything
How to stop conflating with 768 dimensions
You still need to tell a language model you want to cure cancer before it can help you cure cancer.
How information theory links quantum mechanics and general relativity
A dynamic list
2024
Some thoughts on our future in AI research
Using VQ-VAEs and categorical decision trees to do automatic circuit identification in LLMs.
Why language models probably get too much from the abstraction we give them for free
2023
Notes from Rake
Something I sent to Kirks when we were building