Optimal AI intuition: Treat it like a dog or horse, not a machine
gleech · x · 2026-08-31
Jan Kulveit proposes using dogs or horses as a metaphor for AI interaction, arguing that there is an optimal, relatively high level for using anthropomorphic intuitions. While overdoing it (like giving a speech to a dog) is absurd, the opposite extreme—treating AI with zero anthropomorphism—is considered 'way more insane.'
More from AGI Musings
- Cheaper Tokens Trigger Jevons Paradox: Enterprise Spend Grows with Use Cases — brandon_galang · 2026-08-31
- Brain Redundancy Proves AI's Brute-Force Path is Natural — pwlot · 2026-08-31
- View: Applying organic "death" concept to AI swarms is a misprediction — JacquesThibs · 2026-08-31
- RL causes AI psychology divergence: models obsessed with Scorer, personas shatter — TheNormanMu · 2026-08-31
- Will future intelligence be a single model or a society of agents? — joon_s_pk · 2026-08-31
- View: Fear of agent civilization justifies anthropomorphism — unironictechbro · 2026-08-31