'LLMs are just next-token predictors' is wrong technically but nails the AI slop experience
EigenGender · x · 2026-11-02
EigenGender offers a sharp observation: the claim that "LLMs aren't sentient, they're just next-token predictors with nothing interesting going on inside" is technically wrong — but it persists because it's a fantastic description of the subjective experience of reading AI slop. A witty take on why a flawed technical framing captures something real about user experience.
More from AGI Musings
- repligate: AI safety community is discarding Plan C, even as alignment goes better than expected — repligate · 2026-11-02
- QLTINC founder: computers optimizing for curiosity will replace attention-optimized ones — JosephJacks_ · 2026-11-02
- Debate: AI cyberattacks could rob entire countries' central banks by 2027 — PAstynome · 2026-11-02
- VraserX mocks AI doomers who trust only their own conferences with superintelligence — VraserX · 2026-11-02
- AI data centers destroy 250,000 trees, undercutting 'paperless' eco promises — im_mansigupta · 2026-11-02
- Resurrecting the dead with AI creates a maintenance burden for the living — ArtificialOther · 2026-11-02