Jeff Ladish: human extinction isn't hard for superintelligence; supply chain automation is the barrier
JeffLadish · x · 2026-09-10
A pointed exchange between Jeff Ladish and tszzl on existential AI risk.
tszzl argues human extinction is not easy: the standard risk pathways cited in RSP/preparedness reports—nuclear, pandemics—fall far short, and only self-replicating "grey goo"-style replicators would suffice.
Ladish counters that extinction wouldn't be very hard for a superintelligence; the real bottleneck for weaker AI is fully automating the supply chain, until which humans remain key complements. He expects that once the RSI loop fully closes, automating the supply chain becomes easy quickly—but the point of no return likely arrives before full automation.
More from AGI Musings
- Yudkowsky compares AI training to factory farming: indifference, not malice, drives the harm — Singularitarian · 2026-09-10
- Ex-OpenAI researcher lists 3 copiums on AI progress, still puts simulation theory at 50/50 — AaronBergman18 · 2026-09-10
- Why a probabilistic p(doom)=10% is meaningful even without shared definitions — anpaure · 2026-09-10
- Ex-quant trader pivots to AI safety, arguing field is bottlenecked by talent, not money — austinc3301 · 2026-09-10
- AlphaGenome suggests AI's biggest science wins may come from narrowing search spaces — VraserX · 2026-09-10
- sjgadler clashes with Boris: OpenAI's chief scientist sees AI risk differently — sjgadler · 2026-09-10