AI can speed up a lot of the manual labour needed to bring even an almost dead, 10-year-old device back to life. My Xiaomi Mi Max 2 (oxygen, MSM8953) was dead. With AI help and my own knowledge, I ported/rewrote the missing drivers, and it now runs Nura (postmarketOS) with working camera, torch, SIM, phone calls and SMS. Only the loudspeaker is still broken.
Everything is up for anyone who wants to install it, or to understand the changes and manually write the patches for mainline: https://github.com/sv1sjp/mi-max2-oxygen-msm8953-nura-drivers Testing, feedback and PRs welcome.
No, they published them on GitHub. Meaning they aren't peer-reviewed. In the current scientific system, peer review is almost always necessary for your work to be taken seriously. If the original authors can get their work published first, they still get precedence. Plus, there have been a number of cases where X discovers and publishes something, Y rediscovers and publishes it a few years later, and Y gets the credit for whatever reason (usually racism or money).
The scientists were naive to hand over their ideas to a shady closed-source model. I don't think what OpenAI did is acceptable, but it is something that already happens without any LLM involvement, so I'm just not as shocked as you are.
Being published on Github makes it not as bad, though it seems the intent is for the peer review to happen there. I also don't think I need to point out the harm in forcing people to publish before they are ready by stealing their work.
In at least some cases, the mathematicians shared their work with OpenAI researchers asking to collaborate. I'll accept your acknowledgment that it's bad and ignore the irrelevant dig about being shocked.
That's not how this works! That's not how any of this works! You submit to some journal, who sends your work to 2-3 other scientists in the field (often anonymously, to reduce bias). There are cases of scientists ignoring journals and just posting their results to repositaries like arXiv, but that's the exception rather than the rule. Anyway, GitHub is not an acceptable place to publish your findings, since it allows you to make changes later on.
That, again, is an issue that existed long before LLMs.
I know that's not how it's supposed to happen. But the Github repository is getting attention and review, like this: https://kingy.ai/blog/openai-math-722-manuscripts-results-proofs-compute-costs/
I still don't get if you're still trying to excuse something bad and getting worse due to AI by saying it was always happening. I have acknowledged that it was happening.