A.I. Safety Is So Back + Mythos Mayhem with Nikesh Arora + Hot Mess Express
Summary
Hosts Kevin Roose and Casey Newton discuss the Trump administration’s sudden pivot toward AI safety regulations, driven by concerns over Anthropic’s powerful Claude Mythos model. They analyze the political contradictions in US-China tech relations and the internal government turf wars over AI oversight. Later, Palo Alto Networks CEO Nikesh Arora joins to explain how Mythos is accelerating the discovery of critical cybersecurity vulnerabilities, forcing a rapid global response to patch digital infrastructure.
Topics discussed
Intro and IBM sponsorship
Introduction: AI safety and the Hot Mess Express
AI disclosures and the shift in US AI policy
Factions within the Trump administration on AI safety
Regulatory capture vs. genuine safety concerns
The rise of AI-driven cyber vulnerabilities
Trump's China trip and NVIDIA chip exports
Global AI race and the need for regulation
Interview intro: Nikesh Arora on AI cyber threats
Palo Alto Networks' findings on AI-discovered bugs
Using AI to find and fix code vulnerabilities
The shrinking window for patching zero-day exploits
AI in cybersecurity: Attackers vs. Defenders
Consumer security risks and phishing attacks
Impact of AI on cybersecurity hiring and jobs
Outro and sponsor messages
Venmo ending public transaction feeds
Amazon employees and AI token consumption
Students booing AI at commencement speeches
Samsung TV lawsuit over artist's image
eBay's failed $55B takeover bid and GameStop
Copyright dispute between AI companies
Madonna ad controversy on Grindr
Elon Musk's comments on OpenAI governance
Credits and Geico ad
Listen ad-free on Castria