News for hackers. Stuff that matters.
Re:After ChatGPT disruption, Stack Overflow lays off 28 perc... by aurareturn on Wednesday October 18, 2023 @02:58AM (#37925591)
So how will the LLM answer your questions in the future if SO is greatly diminished?

I guess LLMs can learn directly from documentation websites. But there is no human feedback on correct solutions or better ways to do things that SO provides.

Re:After ChatGPT disruption, Stack Overflow lays off 28 perc... by Moldoteck on Wednesday October 18, 2023 @03:10AM (#37925662)
It'll get better and will be able to answer more questions based on sourcecode and documentation?
Re:After ChatGPT disruption, Stack Overflow lays off 28 perc... by zaphirplane on Wednesday October 18, 2023 @03:35AM (#37925846)
LLM do no answer, but kind of autocomplete
Re:After ChatGPT disruption, Stack Overflow lays off 28 perc... by Moldoteck on Wednesday October 18, 2023 @04:53AM (#37926295)
Debatable, but true, still, my point holds, it'll be able to autocomplete bettereven without direct hints / dataset from SO, esp taking into account GitHub & github issues
Re:After ChatGPT disruption, Stack Overflow lays off 28 perc... by zaphirplane on Thursday October 19, 2023 @01:56AM (#37939025)
Without the question text and the words explaining the answer ?

I think those words provide a lot of context to get to that specific response

Re:After ChatGPT disruption, Stack Overflow lays off 28 perc... by isoprophlex on Wednesday October 18, 2023 @04:56AM (#37926314)
Well, yes, and somewhat cynically also through eg. OpenAIs effort to pay programmers in low wage locations to create high quality train data related to programming in general (thus paying them to automate themselves out of a job)
Re:After ChatGPT disruption, Stack Overflow lays off 28 perc... by ofrzeta on Wednesday October 18, 2023 @10:56AM (#37929608)
I've seen support forum sites that have integrated LLMs to automatically generate answers. I doubt that feeding that back into LLMs will improve quality in the long run.