Claude Fable 5 reportedly disproved the 87-year-old Jacobian conjecture

An AI model called Claude Fable 5 has apparently cracked an 87-year-old problem in pure mathematics called the Jacobian conjecture. The claim is trending hard on Hacker News and spreading across social media, but mathematicians have not verified it yet. As of this morning, it’s still in the “interesting if true” category.

The Jacobian conjecture is a problem about polynomial functions that’s been open since 1939. If the proof holds, it’s a legitimate breakthrough. But pure math proofs take time to check, especially when they come from unexpected sources. The math community is cautious, and for good reason. Automated theorem provers and AI models have made claims before that didn’t survive scrutiny.

What matters for you is not whether this particular proof is real. It’s that we’re now in a world where an AI can generate something plausible enough to get 629 upvotes and 387 comments on Hacker News before anyone knows if it’s correct. That’s a new kind of noise. If your team is using AI to generate analysis, code, or strategy documents, you need a way to separate “sounds right” from “is right.” You need humans in the loop who can verify, not just accept. And you need systems that flag when an AI output needs expert review before it gets treated as fact.

This is exactly the kind of thing we build into an AI command centre, where outputs get routed to the right people for verification before they become decisions. The Jacobian conjecture might be solved, or it might be a very convincing dead end. Either way, the lesson is the same: AI can now produce work that looks authoritative enough to fool a crowd. Your job is to make sure it doesn’t fool your business.