Eric Schmidt: Why AI needs divas, fast fails, and a plug

Schmidt argues that the core skill for an 18‑year‑old today is analytical, critical thinking—being able to distinguish between being marketed to ("also known as being lied to") and being given evidence and arguments. Practically, this means habitually checking claims (“Is that true?”), understanding scientific falsifiability, and refusing to pass on unverified assertions. In a world where AI enables "perfect misinformation" and hyper‑targeted feeds, this discipline becomes essential for functioning democracies and sane personal decision‑making.

Despite AI’s growing ability to write code, Schmidt still urges young people to learn Python, the lingua franca of AI systems and APIs. The strategic shift is from hand‑coding everything to orchestrating systems—calling AI APIs, gluing services together, and building products AI can’t specify on its own. Treat coding as a way to prototype games, apps, or tools that call AI, not as a dying artisan skill; you’ll understand what’s possible, know when AI is wrong, and retain leverage in an AI-dominated stack.

Schmidt outlines Google’s 70‑20‑10 rule: 70% of effort on the core business, 20% on adjacencies, 10% on wild bets. A tiny 10–15 person team (Google Brain) from that ‘10%’ generated tens of billions in profit. The key is to structure portfolios so many experiments can fail quickly without killing the company, while a few breakthroughs more than pay for the losses. Entrepreneurs should be explicit about which initiatives are core vs. adjacent vs. experimental, and relentlessly kill what doesn’t work rather than protect zombie projects.

Schmidt differentiates between ‘divas’ (brilliant, demanding builders like Steve Jobs or Elon Musk) and ‘knaves’ (self-serving actors). World‑changing products almost always originate from a small number of true divas, not committees. His recommendation: ally yourself with people markedly smarter and more driven than you, bias hiring toward technical problem‑solvers, and make the CEO effectively the chief product officer. Let product quality and user delight, not sales muscle or politics, dictate strategy; if the product is right, customers line up.

Schmidt believes AI will displace tasks, not eliminate work. Aging populations and labor shortages mean we need massive productivity gains: robots in manufacturing, AI assistants for knowledge workers, AI‑enhanced doctors and teachers. Dangerous, repetitive, and low‑judgment roles (e.g., security guards, some production tasks) will increasingly be automated, pushing people into more interpersonal, creative, or craft roles. Leaders should plan for reskilling and job redesign, not a jobless future—assuming a permanent UBI world is, in his view, naïve about human nature and ambition.