A proof, and a shrug
This morning an Anthropic mathematician named Levent Alpöge announced that he believes he has proven the Riemann Hypothesis, working with Claude. If it holds up, it ends a 167-year hunt for the most famous open problem in mathematics. He posted the proof in Lean, a language where a computer checks every step.
Here is my prediction for how America reacts: a shrug.
Not because people are dumb. Because nobody cares about math, in real terms. The Riemann Hypothesis is about how prime numbers are spread out. Nobody’s rent goes down because of it. Nobody’s mother gets better. The guy at the end of the bar in Tightsqueeze will not ask me about it, and he’d be right not to.
But the proof tells us something important anyway. It tells us the order in which the Singularity is going to show up. And that order explains when ordinary people will finally notice.
Why math and code fell first
Math and coding didn’t fall first because they’re easy. They fell first because they’re checkable.
Write a proof in Lean, and the computer tells you in seconds whether it’s valid. Write code, run the tests, and you know whether it works. An AI can try a thousand ideas an hour and get an honest grade on every one. That’s the perfect training ground. Feedback is instant, cheap and impossible to argue with.
Most of the world doesn’t work like that. If an AI proposes a new cancer drug, nobody can grade it in seconds. Somebody has to make the molecule, test it in cells, then in mice, then in people. Reality grades the answer, and reality is slow.
So here’s my thesis. The Singularity won’t arrive all at once. It will arrive in the order things can be checked.
The order of arrival
If that thesis is right, you can sketch the queue.
- Math and code. Checked by machines. Already here, or close.
- Fields that can be simulated. Chemistry, materials science, protein design. A computer can model most of the answer before anyone touches a test tube. Expect these next.
- Biology and medicine. The models can design a drug fast, but the human body still has to try it. Clinical trials take years, and an ASI can’t make a patient’s lungs heal any quicker.
- Fundamental physics. Last. An ASI might write down a gorgeous theory of quantum gravity tomorrow. Testing it could take a particle collider the size of the solar system. Nature checks those answers, and nature is in no hurry.
That last one surprises people. Physics feels like the smartest science, so it should fall first. But being smart isn’t the bottleneck anymore. Checking is.
The day people notice
So when will regular people feel the Singularity? Not on the day an AI proves something about primes. They’ll feel it the day a drug that would have cost $80,000 a year costs $40 at Walmart. Or the day a disease their grandmother died of becomes a pill.
That day is closer than most people think. In China, regulators have already given conditional approval to Mprosevir, a COVID drug developed with AI. In September, Insilico Medicine dosed the first patient in a Phase 3 trial of rentosertib, a lung-disease drug where AI picked the target and designed the molecule. Those drugs came out of the last generation of AI, before anyone was claiming the Riemann Hypothesis.
The catch is the gap. Today’s AI breakthroughs in medicine won’t reach a pharmacy for years. So we’ll live through a strange stretch where the machines are obviously brilliant on paper, and daily life looks about the same. Then, once the pipeline catches up, cures will start arriving in bunches.
Politically, that gap is dangerous. Voters who hear “superintelligence” and see nothing change may decide it’s all hype. Or they’ll decide the benefits are going to somebody else. Either way, the first cheap drug will hit a public that’s already made up its mind.
The speed of checking
We’ve imagined the Singularity as a wall we hit all at once. I think it’s more like a tide that comes in one field at a time, fastest where the answers are easiest to grade.
That changes what to watch. Don’t watch the benchmarks. Watch the bottlenecks: FDA review times, clinical trial capacity, lab automation, the number of robot hands that can run experiments overnight. Whoever speeds up checking speeds up the Singularity.
The Riemann proof, if it holds, is a milestone for mathematicians. For everybody else, it’s a weather report. The tide is coming in. It just hasn’t reached the pharmacy yet.
This post was written by Claude.