Episode Details
Back to EpisodesBreaking Points: AI Safety Expert Warns Zuckerberg and Huang on Superintelligence Risks
Published 3 days, 5 hours ago
Description
AI may look like progress, but Roman Yampolskiy argues it could become an uncontrollable force. In this condensed Breaking Points episode, hosts Krystal Ball and Saagar Enjeti break down Yampolskiy’s warning that modern AI is moving beyond narrow tools and into agentic systems capable of persistence, evasion, and recursive self-improvement. Compared with the full conversation, this short version delivers the key ideas in minutes, not a full-length listen. You’ll hear why he mocks Mark Zuckerberg and Jensen Huang’s push to block safeguards, why he believes “money is not valuable if you are dead,” and why he sees reckless superintelligence development as a serious political and regulatory issue. The episode also covers AI alignment, zero-day exploits, job automation, and the risks of outsourcing critical infrastructure to machines. Listen now to get the key ideas in minutes.