OpenAI declares AI race “over” if training on copyrighted works isn’t fair use - Ars Technica
OpenAI declares AI race “over” if training on copyrighted works isn’t fair use - Ars Technica
National security hinges on unfettered access to AI training data, OpenAI says.
You're viewing a single thread.
So the software cannot improve making connections between information, only just be fed more data.. I'm ignorant about how this works but that somewhat sounds like giving up.
That's all LLMs could ever do. This was never the path to generalized AI in the colloquial sense.
I don't think it is fair to say that. LLMs in the current form, or by themselves cannot do better. Does not mean, that this basis can be further evolved to do better. But that requires good research, and a lot of open source to cross check and evolve. There is no "business case" in that. Hoarding resources and knowledge and keeping other people from it on the other hand, is something the MBAs understand.
The software isn't doing nothing. What it does do, I would assume, could be better...