Why We Study Fundamental Math Question in AI Era

Every Friday I have a Math Quiz. I usually finish early and end up with a gap before my next Stat class — that sliver of time has become when Adam and I talk.

I found him sitting on the steps, lost in thought. I asked him what was on his mind. He said he was trying to rest, but there are so many things to fix. I asked if he’d slept well. He hadn’t. I told him I hadn’t either — I’d been running a fever this week, felt pretty awful, and had a lot weighing on me.

Adam is always patient with me. Even when my English slips and he can’t quite catch what I’m saying, he never makes me feel bad about it. He just asks me to repeat myself, and somehow makes me feel encouraged instead of embarrassed.

We talked about personal things for a while. Then I asked him something more general — something I’ve been sitting with for a long time:

I study CS. AI is developing so fast. Why should I still be solving these foundational math and stat problems by myself?

He didn’t catch it the first time. I rephrased. He paused, said “that’s a good question,” and told me this isn’t a question that only applies to math or stats — it applies to everything. You can let machines do anything for you. Then you can go on vacation.

I told him that wasn’t really what I was asking.

He thought for a moment, then said something that stuck with me. Maybe he’s right that AI has moved fast enough that we genuinely don’t need to do basic math by hand, or write basic code from scratch anymore. But the point of doing it yourself was never really the answer. The point is the process — the way working through something slowly builds your capacity to think, sharpens how you reason, grows the part of you that notices things.

Then he asked me to sit with a question of my own:

If you were the employer, why would you hire someone? What would make that person special?

I’ve asked this same question before — to my Math instructor, who did her undergrad at Peking University. Her answer was different. She told me that the reason we get to use such intelligent AI today is because the previous generation sat down and learned all of these foundational subjects. If none of us learn them now, eventually no one will know them — and no one will be able to develop the next generation of AI.

I didn’t really agree with her. AI’s trajectory isn’t going to reverse itself. The knowledge she’s describing isn’t going to vanish just because fewer undergraduates grind through problem sets by hand. Her answer felt more like an appeal to duty than a real reason — like she was defending the shape of her own education rather than telling me what the work is actually for.

Which is maybe why Adam’s answer stayed with me and hers didn’t. He didn’t tell me to do the work for the sake of some future generation. He pointed me back at myself.