InfoHawk CEO Robert Leathern calls claims that AI could kill humans 'a bit hyperbolic'

Summary

InfoHawk co-founder and CEO Robert Leathern joins the show to discuss President Trump's post on AI slowdowns and the debate over AI safety. Leathern argues that while AI safety and security concerns are real, claims that AI could kill humans are somewhat hyperbolic. He suggests calls for an AI slowdown may be driven more by a desire to control AI spending, as hyperscalers have been in an escalating feature race and may have room to rein in capex given their lead over Chinese open-source models. He also comments on Anthropic's safety statement and notes uncertainty over Google's response.

• President Trump posted that calls for AI slowdowns are due to U.S. leadership and warned not to kill the golden goose.
• Robert Leathern says AI safety and security concerns are real but runaway AI is less likely than isolated incidents.
• Leathern believes some slowdown calls are political and may help companies control AI spending.
• Hyperscaler AI capex has soared, but Leathern says companies may have room to rein in spending.
• He notes U.S. companies still have a gap over Chinese open-source models.
• Leathern calls Anthropic's safety statement notable because unilateral disarmament is a real give.
• Google's potential response to the Trump post remains uncertain.