Eric Schmidt argues against a ‘Manhattan Project for AGI’
TechCrunchMAR 7
Tech Leaders Warn Against US 'Manhattan Project' for AI
Above: Eric Schmidt during the Future Investment Initiative Institute Priority conference in Miami on Feb. 20, 2025. Image copyright: Zak Bennett/Contributor/Bloomberg via Getty Images
The Spin
Narrative A
MAIM is the best strategy yet. It wields deterrence to temper the dangerous AI race and ensure that no single power ascends unchecked. It crafts a win-win situation, safeguarding humanity from AI threats without hindering all AI development. As was the case with nuclear weapons, this universal program would block bad actors from pursuing reckless agendas no matter how technologically advanced they become.
Narrative B
One only needs to look at ChatGPT's rapid two-year rise to know that the next AI event — superintelligence — is inevitable. Bad actors have already used basic AI to spread misinformation and impact important institutions like elections. If they get superintelligence first, democracy crumbles, which means MAIM's sabotage idea will slow the US down when speed is critical. Democracies must lead the AI race to defeat enemies who won't play fair.