Terminator 2's AI doom portrayal dissected: Skynet takeoff logic vs real RSI loops

JeffLadish · x · 2026-09-21

AI safety researcher Jeff Ladish follows up his Terminator 2 rewatch with an analysis: at a high level the film portrays AI danger better than most AI movies — the real threat isn't the killer robots but Skynet, a giant strategic supercomputer that learns on its own.

He flags two unrealistic points: Skynet's learning setup is unspecified and goes from 0 to superhuman far too fast, yet it still can't wipe out humanity entirely. Still, he thinks a significant takeoff once the RSI loop fully closes is plausible — not overnight, but a few months wouldn't surprise him.

Related event: AI Safety Researcher Says Terminator 2 Is a Surprisingly Accurate AI Risk Film(2 posts)→

Original post →

More from AGI Musings

AGI Musings channel →