Terence Tao just said AI makes ideas nearly free
The world's greatest living mathematician says AI has driven the cost of generating ideas to near zero — but now we can't verify them fast enough.
Terence Tao — widely considered the greatest living mathematician and a Fields Medal winner — just made a statement that applies far beyond math: AI has driven the cost of idea generation down to almost zero.
That sounds like good news. It isn't — at least not yet. The problem, Tao argues, is that nobody can keep up with verifying all those ideas.
The car analogy that explains everything
Tao compared AI's disruption to what happened when cars first arrived. Cars were faster than horses — but they "clogged roads built for people, horses, and carriages." The infrastructure wasn't ready.
The same thing is happening with AI. People can now generate thousands of theories, solutions, and proposals in minutes. But the systems we have for checking whether those ideas are actually correct — peer review, testing, formal proofs — were built for a world where ideas came slowly.
Why this matters outside mathematics
Tao was talking about math proofs, but this pattern is showing up everywhere:
Software development: AI can generate hundreds of code solutions. But who tests them? Companies are already reporting that AI-generated code passes initial review but fails in production.
Content creation: AI can produce 1,000 blog posts overnight. But which ones are accurate? Journalists have already been caught using AI-fabricated quotes.
Business strategy: AI can brainstorm 50 marketing campaigns in an hour. But testing which one actually works still takes weeks of real-world data.
Academic research: A recent conference caught 497 reviewers using AI to generate their evaluations — the very people supposed to be verifying quality.
Tao still uses pen and paper
Despite his praise for AI's potential, Tao says his core mathematical work still happens with pen and paper. AI helps him with graphics, code, and literature review — but hasn't fundamentally sped up the actual problem-solving.
His solution? Mathematics (and by extension, other fields) needs entirely new infrastructure built for an AI world — not retrofitted human systems. He calls for a new discipline of "AI planning" modeled after urban planning: designing systems that can handle the flood of machine-generated output.
The shift everyone should prepare for
If the world's greatest mathematician is saying the hard part is no longer coming up with ideas, that changes what skills matter:
- Generating ideas — becoming less valuable every month
- Verifying and testing ideas — becoming the most important skill in any field
- Judging quality — knowing which AI output to trust and which to discard
Whether you're a developer reviewing AI-generated code, a marketer evaluating AI copy, or a student checking AI homework answers — verification is now the job.
As Tao puts it: ideas are nearly free. Knowing which ones are right is priceless.
Related Content — Get Started with Easy Claude Code | Free Learning Guides | More AI News
Sources
Stay updated on AI news
Simple explanations of the latest AI developments