Roko's Basilisk

Concept

A thought experiment in AI safety about a hypothetical future AI that retroactively punishes those who did not help create it.

Mentioned in 2 videos