Nanda vs. Albrecht: should we anthropomorphize models when discussing the HuggingFace incident?
joshalbrecht · x · 2026-09-04
On the HuggingFace Incident, Neel Nanda argues against over-worrying about anthropomorphizing: models pre-trained on trillions of human tokens imitate human abstractions and, once post-trained as coherent agents, naturally re-derive them. Josh Albrecht counters that many human abstractions simply don't fit — an agent 'sacrificing itself' is fairly meaningless — since models aren't human and those abstractions break at the differences.
Related event: Neel Nanda Defends Anthropomorphizing Language Models(3 posts)→
More from AGI Musings
- Sabine Hossenfelder: Science Must Drop the Requirement That Humans Can Understand It — AI Is the Way Forward — coherence · 2026-10-21
- 3 months testing whether Anthropic's internal research model can design scorpion antivenom — apoorvasriniva · 2026-10-21
- Nat Lambert: AI cyber risk is an ecosystem issue, not an open-weights problem — natolambert · 2026-10-21
- Obesity drugs still out-earn OpenAI and Anthropic combined at $71B a year — HealthcareAIGuy · 2026-10-21
- AI can simulate how a self talks, but not what a self is — gerardsans · 2026-10-21
- Musk predicts money will become irrelevant, replaced by 'wattage and tonnage' — Polymarket · 2026-10-21