When AI Solves the Problem, Who Are You Without the Struggle?

In The Atlantic, Cornell mathematician Steven Strogatz has written one of the most honest essays I have read about what AI is doing to people who love their work. In The Year AI Ate Math, he describes a field in crisis. "Nobody in math is ever going to forget the year 2026," he writes. OpenAI says its models are now cracking problems mathematicians chased for decades, including a report this month claiming close to 400 landmark open problems solved, at about three hours of computation each on average.

Strogatz tells the story through people. A colleague who had returned to one beloved problem for 19 years bought a bottle of cognac in 2010 to celebrate its solution. A machine solved it, and he now expects the tasting to feel more like a wake. A graduate student at a department town hall asked what the point is anymore. And a younger collaborator, Alex Townsend, is releasing hundreds of open problems so that people, working with AI, can solve them and clear the way for genuinely new work.

Strogatz admits he doesn't know what comes next. I don't think the answer lives only in mathematics. Every profession is about to face its own version of this moment.

The Struggle Was Never Just a Means to an End

The heart of the essay is that AI is separating two things that always traveled together: discovering a truth and the joy of discovering it. For centuries, the only way to know the answer was to do the work. And the work shaped the person doing it.

That is true far beyond math. A young lawyer learns judgment by wrestling with a case that won't resolve cleanly. A nurse learns to read a patient by nearly missing something and catching it in time. A manager learns to lead by sitting with a problem that has no tidy answer. The struggle is where people discover what they notice that others miss, what they care about enough to keep at, and who they are under pressure.

This is where conviction comes from. I think of conviction as the earned capacity to act on what you believe before the outcome is certain, and the important word is earned. It is built in the reps, in the hours spent not knowing. I wrote about what happens when AI quietly takes those reps away in When AI Does Your Thinking, You Lose the Reps That Build Conviction. Strogatz's colleagues are living a sharper version of it. The machine finished the problem before they could.

Answers Are Getting Cheap. Judgment Isn't.

Townsend's response deserves attention, because it points toward what remains distinctly human. He imagines a mathematics where people contribute by asking good questions, developing theories, interpreting machine-generated proofs, deciding which problems matter and enjoying the work together.

Look closely at that list. Almost none of it involves producing answers. Nearly all of it involves choosing what is worth solving and making sense of what an answer means. That is the part of any job that depends most on who you are: your curiosity, your values, the problems you are naturally drawn to. What you solve for has always been personal, and it keeps evolving as you do. AI is simply making it the main event.

There is also a quiet lesson in how the colleague with the cognac plans to proceed. He will open the bottle once peers he trusts agree the proof is correct and they understand it together. That is standardization doing its job. Shared standards of rigor and verification are what make knowledge reliable, and they matter even more when the author is a machine. What those standards can't supply is the understanding. People still have to earn that themselves.

I made a related point about the difference between output and earned ability in AI Can Fake the Catch. It Can't Fake What You Earned. The same holds here. A model can produce the proof. It can't produce the mathematician.

What Leaders Can Do Before Their Field Has Its Year

Strogatz writes that he has felt disoriented since May. Many professionals feel the same way, even if they haven't put it into words. Leaders have a role to play before that feeling turns into the question the graduate student asked.

A few questions to bring to your team this week:

Which parts of your people's work build their judgment, and are you protecting those parts as you automate? Efficiency that removes every struggle can also remove the growth.

Do your people know what they bring beyond their output? Ask each person to name the problems they would choose to work on if the routine tasks disappeared. Their answers show you where their best contribution lies.

Where can people still experience discovery? Give them room to understand why something works, not just to accept that it does.

Strogatz ends without an answer, and I respect that honesty. But his student's question has one. The point was never only the answer. It was who you became while searching for it.

AI can hand us the answer, but only the work of understanding it shapes who we become.

Want to explore these ideas further? Learn more about my work on leadership, identity and conviction at www.theglennllopis.com.