~/talk

Upload or die: apocalyptic visions in AI

Talkat Collision 2016 · 2016-04-26..04-28· [large-language-models, tech-ethics]

speaker

name:Jerry Kaplan

links:none on file

resources

video:https://youtu.be/cDboMh3oKG4

slides:none on file

summary

Argues that AI poses no imminent existential threat and that AlphaGo used deep learning on massive data while chess engines used heuristic search, so these are different engineering tools not steps toward a single superintelligence. Says automation replaces tasks but machines lack independent goals, IQ scores for machines are meaningless, and uploading minds is modern religious wishful thinking. Points out practical risks like misused systems, inadequate testing, and contrasts those with larger threats such as genetic engineering, near Earth asteroids, and nuclear winters.

AI-written summary · about

chapters (27)

related talks