Mythos & AI Bad Behavior 4/9/26

The state of the U.S.-Iran ceasefire appears fragile amid Israel’s strikes in Lebanon. Council on Foreign Relation president Michael Froman discusses the fog of the Iran war, the confusion in negotiation, and the many stakeholders in the region. Head of policy at Palisade Research Dave Kasten discusses Anthropic’s newest model Mythos, an LLM so powerful the company isn’t rolling it out to the public. Kasten shares his worries about the AI race, including developers’ increased tolerance of AI “bad behavior” in the name of innovation progress and speed. Plus, Anthropic has lost its bid to temporarily block the Pentagon’s blacklisting, and a few Polymarket users won big on particularly well-timed Iran war bets.    Dan Murphy - 3:42 Michael Froman - 21:03 Dave Kasten 31:04   In this episode:  Dan Murphy, @dan_murphy Andrew Ross Sorkin, @andrewrsorkin Becky Quick, @BeckyQuick Joe Kernen, @JoeSquawk Katie Kramer, @Kramer_Katie Hosted by Simplecast, an AdsWizz company. See pcm.adswizz.com for information about our collection and use of personal data for advertising.