Domingos: no one can post-train models without regressions, so AGI isn't near
pmddomingos · x · 2026-11-14
Computer scientist Pedro Domingos argues that no one knows how to post-train a model to improve at one thing without degrading it at another — and that this unsolved tradeoff implies AGI is not at hand.
More from AGI Musings
- AGI endgame: either a balanced multipolar world or a centralized panopticon — CedricMakes · 2026-11-14
- "60% of people around you have some form of AI psychosis," says Alsop III — StewartalsopIII · 2026-11-14
- AI won't make everyone equal—it will split the smart from the rest worse than money — prasenx · 2026-11-14
- Toddlers learn new verbs only from responsive partners; a call to rerun the test with AI — AnnaCiaunica · 2026-11-14
- Developer muses: in Murphy's world, billions of interdependent events make rare failures daily — bingxu_ · 2026-11-14
- Solo dev: agents made me a hundred-decisions-a-day generalist — and unhireable — rudrank · 2026-11-14