Why competing with AI on knowledge is a race humans will lose

Executive overview

AI models now know more facts than any human. Trying to stay ahead of AI on knowledge is the wrong goal — knowledge has become cheap, and verification and judgment have become more valuable.

True intelligence is generating new concepts, transferring patterns across disciplines, and acting on curiosity. That is what education and careers should cultivate.

The urgent question is not how to beat AI, but how to use the cheapness of knowledge to unlock human potential that broken education systems currently suppress.

What AI has changed

• LLMs function as the world's greatest librarian: if it's been written, they've likely seen it
• Competing on knowledge retrieval against LLMs is unwinnable — like racing Usain Bolt against a motorcycle
• Knowledge is now cheap; how you use and verify it is now expensive
• Human judgment remains irreplaceable: you wouldn't want a librarian as your neurosurgeon or air traffic controller
• AI democratises access to book-level learning at a fraction of university costs
• The gap AI cannot fill: how right questions are derived, what the next questions in a field might be

What intelligence actually means

• Ability to reason and make inferences — not regurgitation of facts
• Creating new concepts and stringing ideas together in deep ways
• Recognising transferable patterns across disciplines to advance another field
• Designing systems from scratch — setting the dials, not just operating them
• Deep expertise through relentless commitment to a niche is also a form of genius
• Schools at every level rarely recognise or reward these forms of intelligence

The Ramanujan lesson

• Ramanujan — a mystic autodidact, two-time college dropout — left behind three notebooks of formulas that shaped 20th-century mathematics
• His story was a turning point: the first time Ken Ono heard his father revere someone who wasn't a perfect student
• Ramanujan's work on modular forms underpinned the proof of Fermat's Last Theorem
• The key question: how many undiscovered Ramanujans exist today, and how do we find them?
• The Spirit of Ramanujan program surfaced hidden talent — including Ono's former student, now his boss

What's broken in education

• Students optimise for test scores and GPA as checkboxes, not for curiosity
• Debt traps force people into careers they hate — purgatory where identity is surrendered to loan repayment
• Children learn about gravity by stacking blocks and giggling — that natural wonder is what education should preserve
• Pressure to get into the right school starts in middle school; wonder is squeezed out early
• Over-valuing speed and perfection in test-taking trains the wrong skills for the next Einstein
• Universities still matter for the human element: mentorship, right questions, next questions in a field

What to cultivate instead

• Act on curiosity — bravery to follow it matters more than credentials
• Passion for the world generates the motivation to engage with hard problems: climate, conflict, culture
• The oddballs and the deeply committed niche experts are both forms of genius worth recognising
• AI as collaborator: it won't laugh at naive questions and will fill knowledge gaps without judgment
• Who owns your identity? You do — don't cede it to a system built around checkboxes