4 ms·
This is why I left math even after solving a 20 year old conjecture in grad school. Literally who cares who solved the problem just publish the results. Acade
by vatsachak 1mo ago
This is why I left math even after solving a 20 year old conjecture in grad school.
Literally who cares who solved the problem just publish the results.
Academia was always politics first results second and I AM GLAD that LLMs are becoming superhuman at math. I like better theorems, not better politics.
- tstactplsignore 1mo agoUh but here we have non-academics at for-profit companies playing politics, and the academic they're threatening being kind and overly generous?
- vatsachak 1mo agoIn math we have a thing called a "scoop"; another mathematician publishing a result that beats yours before you published it. The scooper hardly acknowledges the scoopee unless the methods used were orthogonal. The scooper gets the good journal and the scoopee's paper is usually one tier below. It seems like OpenAI heard of the rumor and then scooped them because their internal model is better/they have more compute. OpenAI has NO obligation to mention Tristan nor Levent, because they DID NOT steal their data.
- tzone 1mo agoYou are glossing over the fact that there is reasonable suspicion that OpenAI used privileged information to do “the scoop”. I.e. they used the fact that researcher used OpenAI tools to get advantage . Imagine if OpenAI opened up a high frequency trading arm and suddenly stole all the prompts and research that other HfT firms are doing through OpenAI tools and start making bank based on that . Wouldn’t that be straight up insane?
- somebodyyoumayn 1mo agohighly agree