About this episode
Plus Why GPT-5 Flopped AI’s Next Hustle? Recruiting People to Train More AIForget robots hiring humans—AI’s hiring trainers. Mercor built an AI recruiter that interviews candidates, and guess what? The real cash cow wasn’t hiring—it was recruiting folks to teach more AI. That means your next job might be training our algorithm overlords, not working for them.This AI Startup Claims It Can “Eliminate Crime” But Has a Wild Surveillance EmpireFlock Safety runs an army of AI‑powered cameras—over 80,000 across the U.S.—and says it’s out here trying to “eliminate crime.” But the reality’s messier. It’s not just about footage—it’s a surveillance network blowing up with privacy alarms, legal headaches, and serious civil‑liberties chatter.Why GPT-5 Flopped—and What AI Needs to Do NextGPT-5 was hyped as almost AGI, but it stumbled on math, riddles, and reliability. Critics say the “bigger is better” scaling mantra has hit a wall. Instead of chasing size, experts argue AI should borrow from cognitive science—world models, core knowledge, neurosymbolic blends—to actually move past statistical mimicry.Vibe Coding Is Here—Engineers Aren’t Replaced, They’re Now Tech ConductorsAI isn’t killing coding—it’s evolving it. In "vibe coding," engineers describe what they want in plain English, and AI converts that into code. Suddenly, engineers are orchestras, not typists—spending more time designing systems, not debugging syntax. The grind’s automated; creativity’s upgraded.Why Teachers Want Student Effort, Not AI’s “Perfect” ShortcutA San Francisco teacher caught a student with completely polished writing but zero understanding. Rather than perfection, she values effort—the grit of digging into a text, wrestling with structure, and building academic resilience. Because shortcuts may polish papers, but they dull the skills that actually let students think for themselves.Are Zuckerberg’s AI Dreams Actually Risky? Experts Warn “End of Humanity” Mig
