Meta’s AI Training Leak | Godfather of AI Pushes “Mommy AI” | Toxic Work Demands Driving Moms Out

Episode notes

Happy Friday, everyone! Congrats on making it through another week, and what a week it was. This week I had some big topics, so I ran out of time for the positive use-case, but I’ll fit it in next week. Here’s a quick rundown of the topics with more detail below. First, Meta had an AI policy doc lead, and boy did it tell a story while sparking outrage and raising deeper questions about what’s really being hardwired into the systems we all use. Then I touch on Geoffrey Hinton, the “Godfather of AI,” and his controversial idea that AI should have maternal instincts. Finally, I dig into the growing wave of toxic work expectations, from 80-hour demands to the exodus of young mothers from the workforce. With that, let’s get into it. ⸻ Looking Beyond the Hype of Meta’s Leaked AI Policy Guidelines A Reuters report exposed Meta’s internal guidelines on training AI to respond to sensitive prompts, including “sensual” interactions with children and handling of protected class subjects. People were pissed and rightly so. However, I break down why the real problem isn’t the prompts themselves, but the logic being approved behind them.

• Stopping the Agent Sprawl: Why Cranking The Dial on Autonomy is Financial and Operational Suicide45 / 100
• The Rise of Tokenmaxxing: Analog Organizational Risks Just Got an Expensive AI Upgrade41 / 100
• (Special Episode) AI Agent Sprawl, Lost Agency, & Rethinking Resistance with Travis Hahler
• Vibe-Coded Catastrophes: Accelerating Intimacy in Digital Apps is Your Security Nightmare
• Fortifying Organizational Fragility (Part 3): Business Physics and the Frictionless Fallacy