The US government forced Anthropic to pull its two newest AI models, Fable 5 and Mythos 5, late last week over national security concerns. The action followed reports that researchers at Amazon allegedly discovered a method to bypass Fable 5's safety guardrails. Since the ban, cybersecurity researchers have circulated an open letter calling the government's move dangerous, and Anthropic has pointed out that the same jailbreak techniques work on competing models as well. The episode of TechCrunch's Equity podcast, hosted by Anthony Ha, Sean O'Kane, and Rebecca Bellan, examines whether this represents a legitimate security threat or simply another wrinkle in Anthropic's complicated relationship with the Trump administration, along with what it means for developers building on the platform and for investors watching the company's upcoming IPO.
Beyond the Anthropic ban, the hosts also discussed why the UK's recently passed social media ban for users under 16 may end up being the lesser of two evils, and unpacked SpaceX's acquisition of Cursor, which they read as a window into xAI's broader strategy and its blind spots. They also highlighted Jeff Bezos's $12 billion investment in Prometheus, a startup working on what it calls an "artificial engineer" — a bet on physical AI that signals the Amazon founder's continued interest in frontier technology outside of his former company.
Listeners can find Equity on YouTube, Apple Podcasts, Overcast, Spotify, and other podcast platforms, and follow the show on X and Threads at @EquityPod. The episode was produced by audio producer Theresa Loconsolo, who has been with TechCrunch since 2022 and previously worked at a four-station radio conglomerate in New Jersey, where she wrote, recorded, and edited content while engineering live interviews. Ha, TechCrunch's weekend editor, previously worked at Adweek, VentureBeat, and a local newspaper in Hollister, California, while Bellan is a senior reporter covering AI and emerging technology for the publication.