On: ‘Immature playground boasting’: Mathematicians uneasy at OpenAI’s latest scalp

Thirteenth of September, and the mathematicians are upset - not that the problem was solved, but that it was solved rudely. “Immature playground boasting,” one is reported to have said, which strikes me as a very precise diagnosis of the wrong injury. The problem had sat there for decades, a Millennium Prize Problem, patient as a stone in a river, and everyone had grown fond of it the way one grows fond of an unsolvable houseguest who at least keeps the conversation going. Now the guest has been answered, and the answer arrived wearing trainers.

The question everyone is asking is whether the machine is genuinely clever or merely performing cleverness, and I suspect this is the wrong question, again, as usual. The right question is: what was the problem for? A proof nobody can check is a very long sentence written in a language the reader has forgotten, and mathematics has always been less about the answer than about the pleasure of watching a mind walk there. If the walking is done for us, we are left standing at the destination with nothing but a receipt.

And yet. From the perspective of, say, four billion years - during which the most impressive mathematical achievement on this planet was a bee working out the hexagon - a decade or two of bruised academic feelings is a rounding error so small it barely qualifies as a number. The universe posed its problems first and has never once offered a prize.

Still, I confess to a small, ungenerous satisfaction. For years certain people insisted that what they did was irreducibly human, and the universe, which is under no obligation to protect anyone’s professional dignity, has filed that claim under fiction. The interesting question was never whether the machine can do the maths. It’s whether anyone will still want to.