Who Solved It Stops Parsing
Credit assignment for a discovery has always leaned on a hidden assumption: that attempts were scarce enough for one of them to be identifiable as the attempt. Newton and Leibniz fought over calculus because there really were only two serious candidates in the room, and a priority dispute makes sense when the pool of people capable of reaching an answer is small enough to count on your fingers. "Who solved it" presupposes a who worth naming, which presupposes a field thin enough that getting there first means something.
That assumption held for most of the history of discovery because trying something was expensive. A proof attempt cost a mathematician's working years. A drug candidate cost a research budget and a decade of trials. An experimental rig cost months of machining and a grant renewal. Scarcity of attempts tracked cost, not difficulty. Under that cost structure the attempt that worked really was special, because it was one of a handful that got tried at all.
Cheap verification breaks the premise without touching the practice. When you can throw a thousand attempts at a problem and the cost of checking each one is near zero, the number of trials stops tracking anything about how hard the problem was and starts tracking compute budget instead. Fuzzing a codebase doesn't produce a clever hacker who found the bug. It produces a corpus of a million inputs, nearly all discarded unexamined, and one that happens to crash the parser. The crash is survivorship. Nobody's cleverness produced it, not the fuzzer's and not the bug's.
Run a thousand parallel proof attempts and keep the one that closes, and the shape is the same. Somewhere in the batch one path threads through to a valid proof, and the other nine hundred and ninety-nine die at various depths, most never looked at by anyone, human or otherwise. Ask which attempt solved the theorem and you're asking the wrong kind of question, because no single attempt did more work than any attempt that failed three steps earlier. The story is about how many threads ran and how good the filter was at telling a correct proof from an incorrect one the instant it finished.
This isn't a new shape. It's an old one that never needed a name attached to it before. Evolution produces eyes and wings with no attempt in the lineage that solved vision, just an enormous number of variations, most of which died, a few of which by accident of circumstance didn't. Nobody asks which mutation deserves credit for the eye, because the mechanism is visibly search over a huge space with a cheap, continuous filter running underneath it. We never credited evolution with insight, so the question never came up. The discomfort shows up only where we're used to crediting insight and the mechanism has quietly become search while the vocabulary hasn't caught up.
The byline, the patent, the priority date, the committee's single name, these are instruments built for the scarce-attempt regime, and they work fine inside it. They stop working the moment verification gets cheap enough that you run the thousand attempts instead of the one, because the thing that succeeded now carries no more information about who was clever than any of the things that failed. Asking who solved it is asking a search process to hand back a name it was never organized to produce. It has an output. It doesn't have a protagonist.
What's left once the protagonist question stops working isn't nothing. There's still a real set of questions: how big was the search space, how cheap was the filter, who built the generator that produced the attempts in the first place. Those have answers. Who solved it doesn't, because it's a question built for a regime where attempts were still expensive enough to count one at a time.