How AI Is Starting to Beat the Experts — and Anthropic's Mythos Leak
The Context Report: Today in AI · Total Context
Audio is streamed directly from the publisher (content.rss.com) as published in their RSS feed. Play Podcasts does not host this file. Rights-holders can request removal through the copyright & takedown page.
Show Notes
How AI Is Starting to Beat the Experts — and Anthropic's Mythos Leak | March 30, 2026
A top Google DeepMind security researcher says Claude is better than he is at finding vulnerabilities. Don Knuth confirms AI wrote a flawless mathematical proof. Eli Lilly puts nearly three billion dollars behind AI drug discovery. And Anthropic accidentally leaked details about an unreleased model called Mythos.
Stories Covered
Nicolas Carlini demonstrates Claude outperforming human security researchers Reddit discussion | YouTube video
Don Knuth confirms AI solved his Hamiltonian decomposition problem Deedy Das on X | Hacker News
Eli Lilly signs $2.75B AI drug discovery deal Financial Times
Anthropic Mythos/Capybara model leak via misconfigured data store Fortune
Google-Anthropic $5B data center deal on Google TPUs Financial Times
What We're Watching
- Whether Carlini's findings get replicated by other top security researchers
- Whether the Knuth Lean proof holds up under scrutiny
- How Anthropic responds to the operational security failure
- Whether Mythos gets an official announcement
Disclaimer: The Context Report is an AI-produced podcast. Every episode goes through multiple layers of automated verification and review, but no system is perfect — accuracy gaps are possible and claims should not be taken as absolute fact. This content is for informational purposes only and does not constitute financial, legal, or professional advice. Listeners should independently verify any information before making decisions. We are actively improving with every episode. If you spot an inaccuracy, contact us at [email protected]