His name is Terence Tao. And what he said about cognitive friction and AI last March has not left me since.
He won the Fields Medal, the highest honor in mathematics, at 31. Most mathematicians consider him the most gifted pure mathematician alive. He holds a professorship at UCLA and directs special projects at IPAM, the Institute for Pure and Applied Mathematics.
He is not a tech optimist. He is not an AI evangelist. He is the person most qualified on earth to tell you whether AI is actually useful for serious intellectual work.
Last March, he stood up at an OpenAI Forum and said something I have not been able to stop thinking about.
What Terence Tao Said
“We lived in a world of cognitive friction until very recently. Every task required us to use our brain. We just thought this was the cost of doing something intellectual. But now we have AI, and these frictions can come down to zero.”
Then he declared AI ready for primetime.
Not eventually. Now.
What Cognitive Friction Actually Means
Most people assume the hard part of intellectual work is the insight. The breakthrough. The idea that changes everything.
It is not.
Most of a researcher’s time, most of a knowledge worker’s time, is not spent discovering. It is spent checking. Verifying. Computing. Searching literature. Testing a path, finding it false, deciding whether the failure taught you anything useful. Reformatting. Re-running. Starting over.
That is cognitive friction. The tax on thinking. The cost of doing something intellectual that produces no insight of its own.
AI does not do the discovering.
It does the checking.
And when the cost of checking drops, you can try more things. That is the relationship between cognitive friction and AI.
Tao said it plainly: “I will try crazier things. You can vibe on the blackboard, and then if there’s a computation that neither of us wants to do, we can just get our AI tool to finish that.”
He also noted that literature searches, finding what has already been done, what has already been tried, what has already failed, are now dramatically faster and more accurate than before.
The friction is coming down. Across every domain of knowledge work.
The 23-Year-Old and the 60-Year Problem
Here is the story about cognitive friction and AI that stops every room I speak in.
A 23-year-old used ChatGPT to solve a mathematics problem that had been unsolved for 60 years.
In 80 minutes.
That is not AI replacing mathematicians. That is AI lowering the cost of exploration enough that a 23-year-old could finally afford to check a hunch that would previously have been too expensive, too time-consuming, too friction-laden, to pursue.
The idea was always there. The capacity to test it cheaply was not.
OpenAI’s Chief Research Officer Mark Chen framed the institutional ambition this way: the goal is not to win a Fields Medal. It is to enable 100 mathematicians to do that for themselves.
That is the shift. Not replacement. Acceleration.
This Is Not Just a Math Story
Tao made a specific prediction in 2003. He said that 2026-level AI, when used properly, would be a trustworthy co-author in mathematical research.
He made that prediction 23 years ago. It came in on schedule.
The implications extend well beyond mathematics. The same friction collapse Tao is describing in pure math is happening in drug discovery, materials science, climate modeling, theoretical physics, and, critically, in every organization trying to outthink its competition.
When cognitive friction drops, the advantage shifts to the people who have the most ideas worth testing. That is when cognitive friction and AI benefit the user.
What This Means for Your Organization
The IDEAS ingredient, one of the five capabilities in the Kryptonite Defense that no AI can replicate, automate, or replace, is not about having ideas.
It is about having enough ideas worth testing, and the capacity to test them faster than your competition.
Cognitive friction was the tax on that capacity. Every hour your team spent checking, verifying, reformatting, and re-running was an hour not spent generating and testing the ideas that actually move your organization forward.
AI is reducing that tax.
But here is the part that most AI conversations miss entirely.
The reduction in cognitive friction benefits the people who have built the ideas worth testing. The researcher with ten strong hypotheses can now test all ten instead of two. The professional who spent years building judgment, expertise, and original thinking gets dramatically more leverage from the friction reduction.
The professional who outsourced their thinking to AI has nothing left to accelerate.
Tao is not saying AI makes everyone smarter. He is saying AI makes the people who already built something worth accelerating much faster.
That is the IDEAS ingredient question your organization needs to answer right now: Do your people have enough ideas worth testing that the cost reduction changes your output?
If the answer is yes, you are being accelerated.
If the answer is no, you are being exposed.
The Tax Is Coming Down. Are You Ready?
Terence Tao declared AI ready for primetime in March 2026.
Not eventually. Now.
The question is not whether cognitive friction is collapsing. It is. The question is whether your organization has built the foundation to take advantage of it, or whether the reduction in friction simply reveals how little original thinking was happening in the first place.
Those prepared need not fear the forces at work.
Find out where your organization stands: realmikeevans.com/scorecard
Read more about the five capabilities that determine who thrives in the age of AI: Knowledge Is Now Worth Zero, But These Five Things Are Not