cd /news/artificial-intelligence/the-jacobian-conjecture-falls-to-ai · home topics artificial-intelligence article
[ARTICLE · art-65898] src=greyenlightenment.com ↗ pub= topic=artificial-intelligence verified=true sentiment=· neutral

The Jacobian conjecture falls to AI

Levent Alpöge at Harvard University used an AI system called Fable to disprove the 87-year-old Jacobian conjecture, providing a 216-character counterexample. The discovery went viral and highlights AI's growing role in mathematical research, though experts note that human expertise remains essential for interpreting results.

read2 min views1 publishedJul 20, 2026

This went viral today, “AI’s solution to 87-year-old riddle takes mathematicians by surprise.”

Levent Alpöge at Harvard University wrote on X on 19 July that the Jacobian conjecture – which academics have spent decades trying to prove was true – is actually false, giving a tiny, 216-character counterexample as proof.

I will admit I wish I had thought of this (not the conjecture), but just running it into a prompt. This is the math-equivalent of winning the lottery, where the price is worldwide fame and publication in a top journal, which, TBH, is better than money away. We’re going to see even more of this, if it hasn’t happened already. Everyone is going to start full-sending open conjectures into all the models at massive scale, in the hope of hitting math paydirt.

It makes me ask, “What is the point of studying math?” when AI and luck matter more, either in terms of choosing problems that are tractable with AI or crafting the perfect prompt. Or, “Why even bother going to university for 4-10 years when you can just run prompts for open conjectures?” Imagine how pissed of you’d feel to have studied this problem for years, only to be beaten by AI. You get no credit for failed attempts. This guy gets all the fame even though it was technically Fable that found the solution.

As someone notes below:

Just to take a step back here, Levent had the highest GPA at Harvard, won the top undergrad math research prize and most frequently collaborates with a Fields Medalist.

We don’t have enough human mathematicians to ask questions to AI and spend time understanding the results

[https://t.co/GhPQzofdCT]— Prakash (@8teAPi)

[July 20, 2026]

This agrees with my post, “AI is not going to make mathematicians obsolete.” Levent Alpöge, who found the counterexample, is from Harvard, which is as prestigious as it gets as far as credentials go. Rather than AI making mathematicians obsolete, it’s a tool used by mathematicians, much like Mathematica. Other mathematicians were recruited to check the proof and to create the LEAN proof. As far as I know, there has been no example of a layperson making a breakthrough with math using AI.

There is always some low hanging fruit that in hindsight seems obvious. You have to keep your eyes peeled. AI is a backdoor to math when combined with attentiveness of whoever uses it to find these low hanging problems.

── more in #artificial-intelligence 4 stories · sorted by recency
── more on @levent alpöge 3 stories trending now
sponsored brought to you by zahid.host 4,200+ EU-deployed projects
reading about agents? ship yours in a single git push.

Run your AI side-project on zahid.host

EU-based hosting, git-push deploys, automatic HTTPS, no cold starts. Free tier with a custom domain — perfect for shipping the agent you just read about.

$git push zahid main
Live at https://your-agent.zahid.host
Get free account → Pricing
from €0/mo · no card required
LIVE [news/the-jacobian-conject…] indexed:0 read:2min 2026-07-20 ·