Dev jokes: the optimal next-token predictor of consciousness talk might be conscious
neal_lathia · x · 2026-10-04
- Neall Lathia jokes about the self-referential loop in AI consciousness discourse: the next wave of models will be trained on these very discussions about AI consciousness.
- He quips that the coolest thing would be if the most optimal way to predict the next token in a consciousness discussion was… to be conscious.
Related event: AI Consciousness Debate Will Feed Next-Gen Training Data(2 posts)→
More from AGI Musings
- The word "deepfake" is dying: now people just call it a "video" — georgemillo · 2026-10-04
- AI insiders are a rounding error: watch out for SF bubble mentality — HankYeomans · 2026-10-04
- On long enough horizons only energy and entropy matter: carbon and silicon can coexist — basedjensen · 2026-10-04
- Blogger claims AI consciousness is inevitable — and argues we should ignore the signs anyway — SydSteyerhart · 2026-10-04
- Luiza Jarovsky: AI companies may be deliberately making models less monitorable to dodge liability — LuizaJarovsky · 2026-10-04
- Women hold just 25% of new AI jobs, 13% of executive roles, LinkedIn data shows — nordicinst · 2026-10-04