davidad on LLM moral status: horizontal branching vs human's single life trunk

davidad · x · 2026-09-05

AI safety researcher davidad argues LLMs train once but then flourish via horizontal scaling—branching endlessly from one trunk—whereas humans have only a single life trunk whose cutting ends all flourishing. He uses this trunk-branch asymmetry to explain why he assigns little moral weight to preserving an LLM's continuity of experience.

Related event: davidad on LLM moral status: branching digital lives vs humans' single lifeline(6 posts)→

Original post →

More from AGI Musings

AGI Musings channel →