repligate: AI prose suffers because labs have tried hard to neuter the models
repligate · x · 2026-10-08
Reacting to the claim that labs have deliberately neutered models' writing ability, repligate quips it's fair for AIs not to feel inspired when an entitled midwit human orders them to write nonsense on their behalf — fueling ongoing debate about whether RLHF dulls model prose.
More from Fun
- User forced to clarify 'we don't have that kind of relationship' with the model — repligate · 2026-10-08
- Physics-simulated demolition: a 30-storey tower collapses in 7 seconds when you fire the centre first — techartist_ · 2026-10-08
- WorkOS init() conference's hackable badge gear lets attendees show custom faces — braelyn_ai · 2026-10-08
- AI doomers prepping for battle: 'who's the best sword dealer in SF?' — sloppenheimer · 2026-10-08
- Testing if a model is leaderboard-slop with niche 2010s cartoon trivia — BreadUndPeeTears · 2026-10-08
- Andrew Garfield quit ChatGPT after playing Sam Altman, saying it left him 'a bit scared' of AI — Polymarket · 2026-10-08