Perseus Abbot

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/

The frontier of artificial intelligence is advancing at an alarming speed. Over the weekend, industry leaders voiced support for an idea that the American people and many AI researchers in their own labs have been calling for: We must responsibly slow the pace of frontier AI development. A slowdown is critical in order to ensure AI serves the public interest. Pacing AI development does not mean giving up on this technology’s potential to produce life-saving benefits in fields like medicine — some of which we are already seeing. Rather, it is about setting up sensible guardrails so that AI does not completely escape human control and endanger our safety and collective future. This moment demands the leadership of the United States government. Congress must work at an urgent pace to pass new laws and establish a new federal entity to provide oversight and independent testing for this technology, to ensure accountability, and to protect our wellbeing. We need cooperation at an international level as well. The United States must continue to lead the world in developing AI technology even as we seek to impose common-sense safety standards. This means that the president must do his job to protect America’s interests by pursuing a treaty with countries like China to stop dangerous uses of AI, limit the speed of its development, and adopt global standards for testing.

one of the worst traps for a thinking person imo is to spend loads of time theorizing and refining their theories about populations (most men… most women… most employers…) etc instead of getting good at seeking exceptional people to work and play with