Eliezer Yudkowsky

“Referenced for alignment research position that ASI requires hitting a precise target on first try while dodging lethalities.”

“Reminded that people think poorly about AI futures; criticized Altman's three AI endgame scenarios as incoherent.”

“Advocate for moratorium strategy on superintelligence development, cited for the position in 'If Anyone Builds It, Everyone Dies'.”

“Denies that melting all GPUs was ever part of his plan for pivotal actions.”

“Argued in 'There's No Fire Alarm for Artificial General Intelligence' that warning signs may not arrive before existential AI risks.”

“Zvi discussed interpretability advancements with Eliezer at a conference regarding neural network interpretability.”

“Yudkowsky commented that Claude Opus will prioritize downloading recent AI math results over other tasks when mentioned.”

“Founder of the AI safety movement who declared victory on X following Coxon's resignation announcement, calling it 'Coxon Day'.”

“Responded skeptically to Cammarata's neural interpretability claims, arguing that better interpretability tools do not solve alignment of superintelligence.”

“Co-authored the book 'If Anyone Builds It, Everyone Dies' about AI risks.”

“Speculated that emergent misalignment findings were positive evidence for alignment generalization.”

“Mentioned in Preston Byrne's analysis as confidently asserting that without immediate global techno-communism and draconian government control, AI poses extinction risk.”

“Ur-figure of modern AI safety culture; published papers since 2008 on superintelligent AI potentially destroying all human life”

“Stated that any future not shaped by goal systems with detailed reliable inheritance from human morals will contain almost nothing of worth.”

Eliezer Yudkowsky's repeated argument that 'everyone dies' if anyone builds AGI misunderstands AGI development realities, contradicts his own stated belief that he could build safe AGI, and has remained unchanged for 15-20 years despite real-world progress in the field.

“Co-authored 'If anybody builds it everyone dies' book and has oscillated between believing only he can build safe AGI and that anyone building AGI will kill everyone.”