this post was submitted on 14 Jul 2023
7 points (88.9% liked)

Programming.dev Meta

2488 readers
1 users here now

Welcome to the Programming.Dev meta community!

This is a community for discussing things about programming.dev itself. Things like announcements, site help posts, site questions, etc. are all welcome here.

Links

Credits

founded 2 years ago
MODERATORS
 

I wanted to start a discussion about the use of AI-generated solutions on Programming.dev. Personally, I've found that AI-powered tools have been incredibly helpful in solving programming questions. I won't name any specific commercial software, but I use one that combines GPT-4 and web search to get more factual information. I write some answers I think I might revisit to the ShareGPT community, but I would prefer posting programming solutions to this instance. However, I'm not sure if AI-generated solutions are welcomed on programming.dev. I'd love to hear your thoughts on this. If AI-generated responses are accepted, how should we format the answers, should we just copy paste without quoting, should we quote the model, just mention that it's AI-generated,...?

you are viewing a single comment's thread
view the rest of the comments
[โ€“] [email protected] 13 points 1 year ago* (last edited 1 year ago) (1 children)

If you mean solutions as in answers to questions I would be against it, LLMs have a habit of spewing wrong information that looks correct. This happens way more when you tell it to write code, it could end up unoptimized, misleading, or straight up wrong. I wouldn't want an AI to answer my question and then feeling like I'm forced to triple check its answers and make sure it's hallucinating.

There's also the point of "if people wanted AI answers, they would be asking chatbots and not post on a community of people"

[โ€“] dbx12 8 points 1 year ago

Yep full support for these points:

  • having to triple check the AI output
  • output is confidently incorrect which furthers the first problem
  • If I want AI answers, I ask the bot directly and thus know how to interpret the result

But: we cannot ensure no AI content is posted as answer. Just putting up the rule that we don't want it.