causal
Speed determines whether trial-and-error AI safety works
Whether AI safety can be achieved by gradual trial-and-error depends on speed: if AGI is a century or two away, safety will likely develop incrementally as with aviation (early crashes taught us to make flight safe without destroying the world), but if full substitutability between human and AI labor produces very fast growth or an intelligence explosion, the slow trial-and-error process societies use to mitigate new technologies is substantially disempowered, greatly increasing risk.
causalpending
Speaker
Will MacAskillEvidence Quote
“if things are moving a lot faster than we're used to i think that increases the risks greatly”
Created: 6/14/2026, 1:56:48 AM
Your notes
Loading notes...