kalomaze: labs mine user data for correlated novel failure classes, not one-off edge cases
kalomaze · x · 2026-09-09
- kalomaze clarifies his earlier point in a debate with radienren about how frontier labs benefit from user chat data.
- His claim: what labs realistically pin down are systematic/correlated classes of novel failure, not esoteric hyperfrontier problems experienced by a single human with zero targeted verifier constructions available.
- He adds that process improvements from user data at the frontier overwhelmingly don't look like gains monocausally attributable to individual power users — rather, transcripts showing widespread power-user frustration in a domain divert effort into that domain.
More from AGI Musings
- OpenAI reportedly spent $15M in tokens on Navier-Stokes proof, and it was worth it — bindureddy · 2026-09-09
- On why evolution produced consciousness: intelligence as the plausible intermediary — jessi_cata · 2026-09-09
- Neal Stephenson has written by fountain pen for 25 years — Baroque Cycle manuscript stood 42 inches tall — ashishkr9311 · 2026-09-09
- Jeff Ladish clarifies: he means intellectual siloing, not just time on Twitter — JeffLadish · 2026-09-09
- Jeff Ladish: researchers building superintelligence being intellectually siloed is extremely risky — JeffLadish · 2026-09-09
- Why aren't you part of a system optimizing for subjective experience? An AI-consciousness argument thread — jessi_cata · 2026-09-09