AIWhy accelerate AI research if you see a risk of catastrophe?
A user assumes most people accelerating work at AI labs put the risk of catastrophe close to zero, and asks those who believe current models pose zero risk to say so publicly.
TLDR
A user questions why anyone with a nonzero “p(doom)”—an estimated probability of AI catastrophe—would help accelerate research, training data work, or training at AI labs. They assume most people doing that work put the risk close to zero. Separately, they ask those who believe the risk from current models is zero to make that position explicit to the public.
Combined views
20.6K
3 Sources, first seen 6h ago
Why accelerate AI research if you see a risk of catastrophe?
A user assumes most people accelerating work at AI labs put the risk of catastrophe close to zero, and asks those who believe current models pose zero risk to say so publicly.