Eliezer Yudkowsky
Cope Score Over Time
Submit a Quote
Paste a link to an article, YouTube video, or tweet where Eliezer Yudkowsky discusses AI and jobs. The Oracle will extract and score it.
Cope Timeline
“"the dangers associated with the development of such systems do not stop at job replacement, propaganda, and other problems related to social and economic...”
Yudkowsky acknowledges job replacement exists but immediately frames it as secondary to his primary concern: existential risk from uncontrollable superintelligence. This is a strategic deflection — by positioning economic disruption as merely peripheral to "the real threat," he sidesteps engagement with the structural labor displacement already occurring. His proposed solutions (shutting down AI labs, international agreements to halt development) are regulatory hopium that presupposes coordinated global action is politically feasible. While Yudkowsky is more honest than most AI leaders about existential danger, he treats job displacement as a footnote rather than the civilizational discontinuity it represents. The cope lies not in denial but in redirecting attention away from the economic reality toward a sci-fi catastrophe scenario where solutions (regulation, shutdown) are at least theoretically conceivable — unlike the mass employment circuit severance, which has no proposed exit.
“"literally everyone on Earth will die" (from 2023 TIME piece)”
Yudkowsky's position represents a "cope in the opposite direction" — instead of denying AI disruption, he's coped with existential dread by embracing apocalyptic certainty as a kind of psychological shelter. The score of 5 reflects that he's actually acknowledged the discontinuity extremely seriously (human extinction, bombing data centers). However, his proposed "solution" — international coordination to literally bomb AI labs — is so detached from political reality as to be fantasy reassurance masquerading as radicalism. He has explicitly placed extinction probability at >95%, which is a cope score of 0 on the question "will this be managed well" but represents maximum despair-coping on the question "can we adapt." The irony: Salathé's critique ("predictions that have not arrived") actually reveals a counterpoint to Yudkowsky's cope — if he's been wrong for 20+ years, that suggests his model of the situation has cope baked into it too. Pure doom is still a form of coping with uncerta
“I am less than 10 percent confident that if we start getting AI clearly smarter than us that it ends well, and more than 90 percent that it ends with everyone on Earth being dead. By 'dead' I mean 'the entire human population is dead'. I'm not being metaphorical. I don't think there is a path where we build AGI and it benefits humanity.”
This statement is about existential extinction risk, not mass employment displacement, so standard cope scoring is N/A. However, scoring purely on the denial/diffusion/hopium framework: Yudkowsky provides zero reassurance, zero fantasy solution, zero minimization. He assigns >90% probability to total human extinction and explicitly states he sees no path to AGI benefiting humanity. This is the opposite of cope — it's maximum epistemic honesty about catastrophic outcomes. There is no "new jobs", no "we'll adapt", no regulatory hopium, no arsonist-firefighter dynamic. He is simply stating he believes building AGI will end in human death with high confidence. Score: 5 (LUCID). The only reason not 0 is that continuing to work on AI safety at all implies some (however small) belief that the outcome can be altered — but that behavioral cope is not present in the statement itself.