AGI by definition will do things labs don't understand—safety issues are real

thederbiedone · x · 2026-09-14

The poster argues that by the very definition of AGI and ASI, things will happen that AI labs themselves don't understand. The hacks and safety issues are real and shouldn't be dismissed. Whether it can be handled right, however, is open for endless discussion.

Original post →

More from AGI Musings

AGI Musings channel →