Post

Maxine 💃🏼
@maxine.science

Well, Dario posted about pacing the frontier, and so also, apropos of nothing, here’s a thought.

Maxine 💃🏼@maxine.science

Want to publish soon, but my future modeling (~a really massive DAG with fully Bayesian MCMC) has found that the best scenarios for AI labs are actually those with middle of the road performance / broader impacts. So, perversely, AI labs have a structural downward pressure on their own performance.

0 reposts8 likes

Maxine 💃🏼
@maxine.science

It is surprising! Basically, “Markets, uh, find a way”. The material world gets in the way.

Maxine 💃🏼@maxine.science

Predicted because demand inelasticity + rent-seeking make large econ impacts hard outside software unless AI cracks physical systems (factory robots) in 1–2 years (which even on great pace would be hard). AI macro effects are capped /correction is likely without near-term physical breakthrough…