3 ms·
I've seen this up close during my time in ML research. I'd often jokingly mention an idea, only for someone to get a state-of-the-art paper out of the exact sa
by screye 4y ago
I've seen this up close during my time in ML research.
I'd often jokingly mention an idea, only for someone to get a state-of-the-art paper out of the exact same idea within a year of that. It annoyed me at first, but with time, I realized that this was a rather familiar occurrence.
Many top labs are tight-lipped even when talking about peripheral topics, because so many of their peers are within months of reaching the same answer, and even the smallest hint might lead to a peer getting there first. My conversations with top researchers indicate that all research in ML is about 'getting there first' rather than producing something that requires a mythical epiphany.
It does take away from the glamor of it a little bit. If Salieri had been months away from composing Mozart's best work himself, that would have taken away from the timelessness of his work.
- DoreenMichele 4y agoMy recollection is that at least two other people came up with the idea of evolution around the same time. Charles Darwin got credit for it for whatever reason, but it seems likely the world was at a point where the pieces were there and just needed to be put together and several people did put it together but only one got credit for it.
- screye 4y agoThe one that blow my mind the most was the simultaneous invention of calculus. Modern Calculus can be reasonably thought of as the most important mathematical leap forward of the last 1000 years. If there was ever an invention that can be considered a piece of unquestionable genius, then that's calculus. And somehow, it was simultaneously invented by 2 different people within a few years of each other. https://en.wikipedia.org/wiki/Leibniz%E2%80%93Newton_calculus_controversy https://en.wikipedia.org/wiki/Leibniz%E2%80%93Newton_calculu...