Eliezer Yudkowsky
"I THINK WE ARE ALL GOING TO DIE"
THE ARCHIVE
History · cartoons · headlines
THE TIMELINE6 sourced events
HE TRIED TO ACCELERATE AI. NOW HE SAYS IT'LL KILL US
Yudkowsky started an institute to accelerate superintelligence, then decided the real problem was surviving it. He coined 'Friendly AI' in 2001 — superintelligence built to share human values, the idea that seeded the AI alignment field, where he's still the most uncompromising voice. His famous fight was with Robin Hanson over FOOM, staged on Overcoming Bias itself: Yudkowsky said superintelligence detonates overnight, Hanson said it climbs gradually. The industry has raced like it's gradual ever since — which, per 'If Anyone Builds It, Everyone Dies', is exactly what gets everyone killed.
2000
HE STARTS AN INSTITUTE TO GET TO SUPERINTELLIGENCE FASTER
Yudkowsky co-founded the organization that became MIRI in 2000. MIRI’s own history says its early goal was to help accelerate the arrival of smarter-than-human AI.
SOURCES: MIRIfirst-party
2003
THEN HE DECIDES THE REAL PROBLEM IS NOT DYING
MIRI’s history says Yudkowsky shifted toward the problem of ensuring advanced AI would pursue desirable goals rather than merely accelerating its arrival.
SOURCES: MIRIfirst-party
2006
THE ARGUMENT MOVES ONTO THE INTERNET
Yudkowsky became a principal contributor to Overcoming Bias, writing extensively about rationality, catastrophic risk and artificial intelligence.
SOURCES: LessWrongreported
Feb 2009
LESSWRONG BECOMES ITS OWN COMMUNITY
LessWrong was created in 2009 using Yudkowsky’s writing as seed material, becoming a major online home for rationalist and AI-risk discussion.
SOURCES: LessWrongreported
Jan 30, 2013
THE SINGULARITY INSTITUTE BECOMES MIRI
The organization renamed itself the Machine Intelligence Research Institute, emphasizing technical research on advanced AI and alignment.
SOURCES: MIRIfirst-party
Sep 16, 2025
THE DOOM ARGUMENT BECOMES A BOOK
Yudkowsky and Nate Soares published If Anyone Builds It, Everyone Dies, presenting their case that sufficiently powerful unaligned AI could cause human extinction.
SOURCES: MIRIfirst-party
