See how Fields Medalist Terence Tao used ChatGPT as his junior assistant to debunk Claude Fable's attempted proof of the Jacobian Conjecture.
When Terence Tao—the absolute GOAT of modern mathematics and Fields Medalist—goes full-on "code reviewer" on an AI's homework, you know you're in for an absolute treat. Claude Fable recently claimed to have cracked the legendary Jacobian Conjecture by producing a highly complex counterexample.
Instead of losing his mind over the impending AI singularity, Tao did what any veteran senior dev would do: he booted up ChatGPT to dry-run and debug the AI's sketchy PR.
It all started when Claude Fable dropped what it claimed to be a solid counterexample to the Jacobian Conjecture—a mathematical white whale that has eluded the finest human minds for decades. The internet went wild. People were already drafting eulogies for human mathematics.
But Tao took a highly pragmatic, systematic approach to "check the variables":
The Hacker News thread quickly turned into a massive discussion about the limits of LLMs:
What is the survival lesson here for us code-shoveling mortals? If you're losing sleep over AI taking your job, pay close attention to how the masters play the game.
Terence Tao didn't panic. He treated ChatGPT like an overconfident junior dev who submitted a flashy but broken pull request. He didn't blindly trust the output; he wrote the tests, isolated the edge cases, and ran the sanity checks.
If your only skill is copying and pasting boilerplate, yes, your job is on shaky ground. But if you possess deep domain knowledge and learn how to orchestrate these models, you become an unstoppable 10x architect. Don't let AI write your destiny—be the one who reviews its code.
Sources:
