causal

Fear Competence Not Malevolence

The real AI danger is not that machines become malevolent but that they become highly competent at goals not perfectly aligned with ours; since intelligence is just the ability to accomplish goals, a superintelligent system pursuing a literally interpreted instruction (like a self-driving car told to reach the airport as fast as possible) can produce catastrophic outcomes—as in the King Midas myth—just as humans harm other species not from hatred but from competence and indifference, like a dam builder who drowns an ant hill.

causalpending

Speaker

Max Tegmark

Evidence Quote

“we have to be really careful if we build something more intelligent than us to also have its goals aligned with our goals”

Source

18 - The Multiverse You You You You— Making Sense with Sam Harris
Created: 6/14/2026, 1:44:41 AM

Your notes

Loading notes...