>>5834361>>5834216>>5833695>>5833654>what do AI companies do with those prompts. You know the ones I mean... "those" special onesI am not too sure, but I reckon United Kingdom Prime Minister Rishi Sunak is very excited about it. This is his actual daily facial expression. Meanwhile Elon is pretty smug, he knows what is coming
I am not too familiar with this Ted Chiang (?? his name sounds a bit like a form of 4chan /pol endearment) this book author behind the Arrival film lol Denis Villeneuve lol but anyway, I read an article he wrote about how an interesting determinant of generative LLM credibility would be the extent to which the companies used the output of their own LLMs to train datasets for future iterations (I know some algorithms do this, eg alphago reinforcement learning plays against itself, but they are structured differently)
If the AI companies are unwilling to use their own generated LLM output to train their own data models and instead rely upon other external datasets (ie appropriated / stolen work from others hehe) to train their ChatGPT etc whilst foisting their hallucinated data output onto the wider user audience, that is telling you something about the confidence they have in the utility of ChatGPT
