News for hackers. Stuff that matters.

A comment on OpenAI’s policies hinder reproducible research on language models. Parent comment

Re:OpenAI’s policies hinder reproducible research on languag... by lionkor on Thursday March 23, 2023 @06:49PM (#35282547)
Cool, but sadly, as I said, it did not give a very useful answer. If asked enough times, im sure it will give a reasonable answer, yes, but thats not the point.

GPT4s answer is interesting, though

[ Parent | on HN ]
Re:OpenAI’s policies hinder reproducible research on languag... by Mike_12345 on Thursday March 23, 2023 @07:01PM (#35282693)
But all of the answers were correct and useful, and GPT-4 was perfect. Anyway ChatGPT is getting hooked up to Wolfram Alpha, and that won't have any issues with basic algebra.
[ Parent | on HN ]
Re:OpenAI’s policies hinder reproducible research on languag... by ugh123 on Sunday March 26, 2023 @12:21AM (#35310902)
If you already know the answer, but need an interesting and detailed way to explain/write it, then GPT4 can be an immense value-add. Even if you have to do some fact checking afterwards.
[ Parent | on HN ]