The hosts analyze OpenAI's new releases, emphasizing the move to autonomous agents that can operate a user's computer, though they note these tools are still 'software-brained.' They discuss enterprise AI adoption, highlighting the 'reverse information paradox' where companies risk giving away proprietary knowledge when using AI models. The episode also covers the AI 2040 report's optimistic path to superintelligence, Apple's trade secret lawsuit against OpenAI, and Illinois's new AI safety law requiring transparency frameworks and third-party audits. The AI Safety Index grades leading companies poorly, with Anthropic earning a C+. A cheating scandal at Brown University reveals students using AI on take-home exams, with scores dropping dramatically when the final was in-person. The hosts also discuss foreign state actors spreading misinformation about AI data centers and share practical AI use cases for competitive analysis.