interpretability might be the most important scientific problem in the world right now, as I don't think we can solve alignment if we don't know what we're building. I think I have good news about it. I co-authored early work in mechanistic interpretability at OpenAI and came away pessimistic, because neural networks appeared to fold into arbitrary tangled knots. years later, I started openai's first large-scale automated interpretability project, finding about 2% of neurons legible with the primitive tools we had with modern agentic tools and a much closer look, the picture looks totally different. the vast majority in models up to trillions of parameters are legible, and their circuits beautiful. many famous examples of tangled neurons turned out to be misunderstood. The roadblocks I expected mostly aren't there. I now think ambitious interpretability will be solved by the time we need it most, and possibly quite soon, as automated scientists keep improving. it may be a bitter lesson for interpretability: the simple thing, at absurd scale, works already-clean.goodfire.com/

thanks for telling us. 3 fucking weeks late

I think there's maybe a 20% chance within the next year that the entire internet will simply no longer be available for a week or longer Think hard about what would happen in that scenario and make sure you can cope, whatever that entails

Three way UK team up announced tonight.

nathan fielder has an almost incomprehensible film making talent and he’s wasting it on some quite boring subject matter. there is nothing very interesting about elizabeth holmes or billy evans

who is working on a LLM with original sin ?