EPISODE · Apr 30, 2026 · 30 MIN
Zuckerberg's AI Agents Are Proving The Dead Workforce Theory
from They Might Be Self-Aware · host Hunter Powers, Daniel Bishop
Zuckerberg's AI agents are training on you. Meta is recording every employee's screen (mouse, keystrokes, screenshots) to teach the model how to do your job. Hunter and Daniel break down Meta's "Model Capability Initiative," the workplace surveillance pipeline disguised as productivity tooling, and the new cottage industry of AI agents companies buying defunct startups' Slack archives and Jira tickets because the public internet's training data well ran dry. They coin "dead workforce theory," roast the AI agent whose marquee feature is keeping two coworkers from accidentally throwing two birthday parties for Bob, and walk through the Bank of America playbook: January 2026, CEO Brian Moynihan says AI "isn't a threat to jobs." April 2026, 1,000 jobs cut, AI mentioned six times in the press release. Meanwhile at Anthropic, the Mythos model (the one they hid because it's "too dangerous") is loose. Someone guessed the URL by incrementing a "2" to a "3" and now there's a Discord channel handing out access. Sam Altman called it "fear-based marketing." (Pot, meet kettle.) The NSA is reportedly using it anyway, the same week Anthropic's CEO has been spotted at the White House. Earlier this year the same lab leaked the entire Claude Code source via a supply chain hack. Containment? Apparently optional. Plus: digital likeness rights, "just a little DNA sample," and Daniel's vision of six guys with bitcoin hoses going straight into their mouths while the rest of us audition for UBI. ⏱️ CHAPTERS 00:00 Gary on the dumpster payphone 02:06 Mythos breach: the URL guess 04:55 Altman calls it fear-based, NSA uses it 09:22 Meta's spyware trains your replacement 13:29 BofA's "not a threat" 1,000 layoffs 16:25 Selling dead startups' Slack archives 20:08 Dead Workforce Theory + Bob's birthday 23:35 Digital likeness, DNA, the bitcoin hose 26:42 Wrap-up: What About Bob, Gary takeover ⚡ Listen now & get self-aware before your tools do. 🎧 Listen on Spotify: https://open.spotify.com/show/3EcvzkWDRFwnmIXoh7S4Mb?si=3d0f8920382649cc 🍎 Subscribe on Apple Podcasts: https://podcasts.apple.com/us/podcast/they-might-be-self-aware/id1730993297 ▶️ Subscribe on YouTube: https://www.youtube.com/channel/UCy9DopLlG7IbOqV-WD25jcw?sub_confirmation=1 📢 Engage What's your number? Tell us in the comments: how much would your employer have to pay you to install Meta's screen-recording "capability" software, or is there no number that makes that okay? And while you're here: cupcakes or cake for Bob? He needs to know. New here? Subscribe for twice-weekly AI chaos. 🧠 They Might Be Self-Aware — but are we? #AIAgents #DeadWorkforceTheory #Anthropic
What this episode covers
Meta is installing screen-recording software on every employee to train the AI agents that will replace them, and Anthropic's "too dangerous" Mythos model is loose in the wild because someone guessed the URL. Hunter and Daniel cover the workforce surveillance arms race, coin "dead workforce theory," and dissect the AI agent whose flagship feature is preventing two coworkers from accidentally throwing two birthday parties for Bob.
NOW PLAYING
Zuckerberg's AI Agents Are Proving The Dead Workforce Theory
No transcript for this episode yet
Similar Episodes
No similar episodes found.
Similar Podcasts
No similar podcasts found.