ponder

ponder.ooo

17 October 2026 19:56

re: ai safety bill: ppl already do projects w/ serious safety considerations, & a good framework looks less like "ppl at metr w/ close personal ties to industry leaders decide which models civilians can use" & more "a licensed professional engineer has to sign off on every step of your operations"

ponder

ponder.ooo

17 October 2026 14:55

the ai companies are happy to collaborate with the regime's insane wars of aggression and the domestic operations of ICE. it is insane to joke abt "first they came for the artists, then the mathematicians" re: ppl currently participating in a real fascist project targeting fascists' usual targets

ponder

ponder.ooo

17 October 2026 12:17

the biggest open problem in ai safety is how do you get ai researchers and ai safety researchers to stop collaborating with fascists

Sir Reginald, Duke of Chutney

hyperplanes.bsky.social

17 October 2026 12:11

I think it's obviously that a statement in mathematics evaluates to true or false, whereas art is purely subjective

David

yonderdavid.bsky.social

17 October 2026 12:10

Mathematicians disagree about this, too. Just as they disagree about whether they belong first to the humanities or the sciences. But I think a big part of why team discovery feels as they do is that they are navigating a system of locally inviolable rules.