Researcher pushes back on Anthropic's model personhood narrative: 'never was science'
gerardsans · x · 2026-09-07
Researcher Gerard Sans argues that anthropomorphic narratives of AI personhood don't survive hard facts: inference is local, stateless, and atemporal. Two independent queries sharing no priors break the narrative entirely.
Key points:
- A 'local bias or compression fossil' is not proof of personhood or a self — it's 'single forward pass archaeology'
- He credits Anthropic's safety and alignment work as technically decent, but says projecting personhood onto vectors doesn't match the underlying architecture and creates unnecessary noise for researchers
Related event: Developer Pushes Back on AI Personhood Narrative: Models Are Just Software(2 posts)→
More from AGI Musings
- Senior researcher rebuts 'one-person pretraining' fear: your fundamentals matter more, not less — RexDouglass · 2026-10-23
- PHP, certificates, railways: a 48-hour wave of hacking incidents unpacked — David Ondrej · 2026-10-23
- NBER releases 'The Economics of Transformative AI' edited by Agrawal and Brynjolfsson — JMateosGarcia · 2026-10-23
- Pedro Domingos: AI Is an Intelligence Lever — the Smarter You Are, the Smarter It Makes You — pmddomingos · 2026-10-23
- Anil Seth: Consciousness Is a Property of Life, Not Computation — Building Conscious AI Is a Bad Idea — anilkseth · 2026-10-23
- Pedro Domingos predicts 2027 will see the most security vulnerabilities fixed in history — pmddomingos · 2026-10-23