TBPN Substack: Tech heavyweights debate fast unemployment AI scenario
Aug 19, 2025
Key Points
- Tech figures including Andrew Cote and Aaron Slodov diverge sharply on AI labor displacement, with skeptics arguing adoption rates remain far lower than hype suggests and historical precedent favors gradual transitions.
- Cote contends AI will eliminate pure execution roles while forcing humans into decision-making agents, but Slodov counters that tech founders systematically overestimate real-world adoption and that AI tools show high churn rates.
- Bryan Johnson sidesteps prediction entirely, arguing first-principles thinking breaks down with AI and that civilization's next organizing ideology must center on human survival rather than labor market mechanics.
Summary
Andrew Cote argues that AI will eliminate jobs requiring only script-following but force people into genuine decision-making. Most current jobs are busywork maintained by what he calls an "army of environmental coordinators" doing stakeholder engagement and email generation. When those roles disappear, people will have to become agents rather than wage laborers. Cote acknowledges job losses but sees a net positive outcome because AI excels at recall and fails at decision-making, the opposite of what destroys human value.
Roon holds the middle position: progress is mostly gradual, though discrete shock events can occur. A language model might suddenly encompass all core functions of an L3 software engineer, creating a discontinuous labor shock. But that discontinuity won't end labor itself. Knocker-uppers and street-light attendants lost jobs to technology, and those transitions unfolded over years.
Aaron Slodov rejects the fast unemployment scenario entirely. AI hype, he argues, comes from people in tech who don't spend time in the real world. We've built a crude form of AGI that handles digital tasks with enough training data, but growth rates look far more dramatic than total penetration. Fax machines are still in use. AI tools have "insanely high churn"—they hook users initially, then get abandoned because they don't replace entire workflows. The people firing workers are replacing single-task roles like call-center labor. Until real AGI arrives, AI is just another tool, and early adopters will dominate laggards.
Bryan Johnson sidesteps prediction. With AI, first-principles thinking breaks down. We're entering an era where we must acknowledge that we don't know what we don't know. His only claim: "don't die" will be the next full-stack ideology that helps civilization function.
The tech consensus around a fast labor collapse is thinner than it appears. People have incentives to signal belief in runaway capital dynamics and AGI inevitability. But when asked directly by peers, most respondents retreat to gradualism, discontinuous shocks interspersed with stability, or skepticism about real-world adoption. Slodov's observation cuts deepest: tech founders assume the world will drop everything to use what they built. History suggests otherwise.