Shawn Powers

Taken together, these findings point to two distinct barriers to AI governance. First, a material barrier—driven by limited fiscal and institutional capacity—prevents some states from acting even when risks are recognized. Second, an ideological barrier, rooted in regulatory skepticism and market-oriented political preferences, constrains action... See more

Taken together, these patterns suggest that Republican governors and partisan alignment contribute to ideological and institutional inertia, while older population bases further reinforce legislative inaction. In these contexts, states tend to prioritize conservative fiscal management and established policy areas over emerging forms of... See more

There is also a genuine risk that overly prescriptive legislation ends up imposing tests or rules that don’t actually improve safety but that waste a lot of time (essentially amounting to “safety theater”)—this too would cause backlash and make safety legislation look silly.

I believe the only solution is legislation—laws that directly affect the behavior of AI companies, or otherwise incentivize R&D to solve these issues.

The unique value of interpretability is that by looking inside the model and seeing how it works, you in principle have the ability to deduce what a model might do in a hypothetical situation you can’t directly test

ecall that these AI models are grown rather than built,

But there’s no way to know that for sure, and when we’re talking about risks to humanity, it’s important to be paranoid

Millions of people talk to Claude about an astonishingly diverse range of topics, which makes it impossible to write out a completely comprehensive list of safeguards ahead of time.

Chatbots Can’t Be Trusted to Spot AI-Generated Videos