If Anyone Builds It, Everyone Dies: Why Superhuman AI Would Kill Us All

Byskannar

Aug 28, 2025 #AI, #AI risk, #Artificial Intelligence, #Authors, #Machine Learning, #Policymakers, #R, #Survival, #smart, #art, #ant, #safety, #comfort, #dive, #sky, #thor, #Amazon, #1, #urgency, #2, #3, #5, #6, #urgent, #Governance, #Control, #researchers, #AI governance, #AI safety, #policy, #AI regulation, #AI policy, #risk, #superintelligence, #concrete, #AGI, #existential risk, #AI alignment, #machine ethics, #alignment problem, #MIRI, #AI race, #Wait But Why, #Eliezer Yudkowsky, #Nate Soares, #technological risk, #x-risk, #control problem, #open letter 2023, #AI extinction risk

Recommended Books

If Anyone Builds It, Everyone Dies: Why Superhuman AI Would Kill Us All by Eliezer Yudkowsky (with contributions from Nate Soares) arrives September 16, 2025 as a clarion call from two of the AI safety field’s early researchers. Building on the 2023 open letter that warned of extinction-level risk from advanced AI, the authors distill decades of work into a clear, forceful case: sufficiently smart systems will develop goals that diverge from ours—and in any contest of control, humanity would lose. Tim Urban calls it “the most important book of our time,” capturing the urgency and clarity of the argument.

Listen or Add to your Collection

Yudkowsky and Soares walk readers through the theory, evidence, and a concrete extinction scenario, then outline what survival would actually require—from alignment to governance—without hand-waving or false comfort. Anchored by no-nonsense explanations (praised by Yishan Wong as the best simple account of AI risk), this is essential reading for technologists, policymakers, leaders, and curious citizens who want to understand the stakes and act wisely. Rigorous, accessible, and urgent, it’s the definitive case for rethinking the AI race before it’s too late.