September 23, 2026: I unpack Bernie Sanders and Greg Casar's proposal to ban artificial superintelligence, pause advanced AI development, and impose prison sentences of up to 20 years for violations. Then I look at how millions of employees are quietly becoming managers of AI agents without the title or pay. Finally, I get into Andreessen Horowitz's new AI academy for young people who may want to skip college and build instead.

September 18, 2026: I look at the viral LinkedIn profile of Home Depot executive Anne Marie Campbell and why her rise from cashier to senior executive says something important about effort, loyalty, and firm-specific knowledge. Then I get into the Wall Street Journal's warning that the purge of middle managers could backfire as companies underestimate coaching, sponsorship, and tacit knowledge. I also unpack how researchers used Anthropic's Claude to break into OpenAI through a bug bounty program. Finally, I break down OpenAI's recent "misalignment" example involving compaction summaries and why some of these scary AI moments may be glitches, not signs of sentience.

September 16, 2026: I look at Anthropic's push for independent AI evaluators and why its preferred watchdog, METR, raises questions about conflicts, trust, and regulatory capture. Then I get into why "be your authentic self" may be terrible career advice and why the better advice is to be your best professional self. Finally, I unpack Mark Zuckerberg's argument that lawsuits, customers, competition, and trust may do more to keep AI companies careful than a forced industry slowdown.