The Future of Addictive Design + Going Deep at DeepMind + HatGPT

This episode of Hard Fork dives into two major themes: the legal reckoning facing social media platforms and the ethical frontiers of AI development. In the first segment, hosts Kevin Roos and Casey Noon discuss landmark jury verdicts in Los Angeles and New Mexico, where Meta and YouTube were found liable for designing addictive platform features that harmed users, particularly minors. These cases mark a potential turning point in internet law, challenging the long-standing protection of Section 230 by framing platform design as a form of product liability. The hosts debate whether features like infinite scroll and recommendation algorithms constitute defective products, with Casey arguing that the scale and behavioral engineering of platforms like Instagram make them uniquely harmful, while Kevin raises concerns about overreach and the blurring line between content and design. The conversation then shifts to AI, with a deep dive into Sebastian Malaby’s new book, *The Infinity Machine*, which explores Demis Hassabis’ vision for superintelligence at Google DeepMind. Malaby portrays Hassabis as a driven, spiritually motivated scientist obsessed with solving humanity’s greatest problems, yet also deeply aware of the existential risks of AI. The episode closes with a playful round of Hat GPT, where the hosts riff on recent tech news including AI-generated fruit reality shows, a banned AI agent writing angry blogs, a leaked Claude coding harness, and a controversial AI podcast startup that records Zoom meetings without consent. The tone balances serious concern with sharp humor, underscoring the growing tension between technological innovation and societal responsibility.