To EAs and AI safety people I would especially emphasize the point that we are underestimating the need for communicating with and involving the public. I think getting massive popular buy-in is critical for passing and enforcing the kind of robust AI governance legislation that would significantly buy time and reduce x-risk. The anger at data centers shows the potential energy is there, but whether and how that energy translates into legislative accomplishments probably depends heavily on whether and how AI safety people are in touch with the rest of America and the world.
This applies even/especially to race dynamics; e.g., achievements in nuclear deescalation were driven substantially by popular movements within the respective blocs.
Disclosure: I’m a member of PAISC, but my opinions are my own.
To EAs and AI safety people I would especially emphasize the point that we are underestimating the need for communicating with and involving the public. I think getting massive popular buy-in is critical for passing and enforcing the kind of robust AI governance legislation that would significantly buy time and reduce x-risk. The anger at data centers shows the potential energy is there, but whether and how that energy translates into legislative accomplishments probably depends heavily on whether and how AI safety people are in touch with the rest of America and the world.
This applies even/especially to race dynamics; e.g., achievements in nuclear deescalation were driven substantially by popular movements within the respective blocs.
Disclosure: I’m a member of PAISC, but my opinions are my own.