this post was submitted on 03 Sep 2024
44 points (95.8% liked)

Programming

17349 readers
232 users here now

Welcome to the main community in programming.dev! Feel free to post anything relating to programming here!

Cross posting is strongly encouraged in the instance. If you feel your post or another person's post makes sense in another community cross post into it.

Hope you enjoy the instance!

Rules

Rules

  • Follow the programming.dev instance rules
  • Keep content related to programming in some way
  • If you're posting long videos try to add in some form of tldr for those who don't want to watch videos

Wormhole

Follow the wormhole through a path of communities [email protected]

founded 1 year ago
MODERATORS
[email protected]
[email protected]
[email protected]
44
Policy: Generative AI (e.g., ChatGPT) is banned (on Stack Overflow) (meta.stackoverflow.com)
submitted 2 months ago* (last edited 2 months ago) by [email protected] to c/[email protected]
22 comments fedilink hide all child comments
 

Obviously, given the subject matter, I had to let ChatGPT generate a summary for this:

The Meta Stack Overflow post discusses a policy decision regarding the use of generative AI tools, such as ChatGPT, on the platform. The key points include:

  1. Ban on Generative AI: The community has decided to prohibit the use of generative AI for answering questions on Stack Overflow. This is due to concerns about the quality and reliability of AI-generated content.
  1. Quality Control: The decision aims to maintain high standards for answers, as AI-generated responses may lack accuracy and context, potentially leading to misinformation.
  1. Community Feedback: The policy was influenced by feedback from the community, emphasizing the importance of human expertise in providing reliable answers.
  1. Future Considerations: The post suggests that while the current stance is a ban, the situation may be revisited in the future as the technology evolves.

Overall, the policy reflects a commitment to ensuring that the content on Stack Overflow remains trustworthy and valuable to its users.

top 22 comments
sorted by: hot top controversial new old
[–] [email protected] 25 points 2 months ago (1 children)

It's sad but i stopped writing answers or comments on SO years ago. I used to have all these optimistic ideas about people working together to collectively grow our shared knowledge. I guess Wikipedia and the Internet Archive keep barely hanging in there, but if anything those cases prove my point: without one extremely strong personality to hold the corruption in check, all these collaborative "digital commons" projects are a leadership change away from completely selling out all the work put into them. That can be feeding everything into AI but it's also monetization schemes and EULA changes to claim ownership of user submitted content and locking the public out of your site without accounts and subscriptions.

And usually the public's only recourse is to tear it all down and start again, waiting for the next con artist to come along and steal the village's prosperity.

[–] [email protected] 3 points 2 months ago

At least stackoverflow content is publically licensed

[–] [email protected] 18 points 2 months ago (1 children)

Is this post just highlighting the policy thats existed for 2 years, or has the policy changed?

[–] [email protected] 19 points 2 months ago* (last edited 2 months ago) (1 children)

It has now been finalized as of yesterday, from what I understand. Previously it was a work-in-progress policy change they were still unsure about, and now it's decided that this is the way going forward.

I reckon sadly at least part of the reason will be that they are in a partnership with OpenAI, and feeding generated stuff into a GenAI breaks the model, so they need to keep SO as non-AI as possible.

[–] [email protected] 10 points 2 months ago (2 children)

Yeah this is just them trying to prevent model collapse (stemming from pollution of training data by ai garbage). The moral thing to do here is to not give a fuck and upload as much AI garbage as possible to poison their dataset. Fucking assholes.

[–] [email protected] 3 points 2 months ago (2 children)

Does the harm that does to the community not factor into your moral calculus?

[–] [email protected] 3 points 2 months ago* (last edited 2 months ago)

I does factor that in and my answer is to abandon ship or mutiny. By contributing to their ecosystem you validate their sense of not having done anything wrong.

As we have seen with reddit, the next step here is closing off API access and banning searchengines and crawlers from using the sites resources. At that point there is little value left in the platform just like with reddit.

[–] [email protected] 3 points 2 months ago* (last edited 2 months ago) (1 children)

Stack Overflow only pretends to care about their community and I'm tired of contributing the greater good of the technical community at large (edit: by contributing to Stack Overflow). Stack Overflow has a track record of ignoring Meta. Why even have it if you're not going to listen?

[–] [email protected] 1 points 2 months ago (1 children)

Ok, but why would you know that and then go harm the community yourself too?

[–] [email protected] 1 points 2 months ago (1 children)

I'm not advocating for maliciously using AI generated content to make Stack Overflow worse, I'm just saying I feel no motivation to seek questions to answer there nor to waste my time posting good questions there that I research and find no duplicates of only for them to be closed.

[–] [email protected] 1 points 2 months ago

Ah, I confused you with the person I had originally replied to

[–] [email protected] 3 points 2 months ago (1 children)

“Avoid model collapse” aka “avoid giving garbage answers”. If they’re gonna use a lot of energy to train an AI, why not at least make the product good-ish?

[–] [email protected] 6 points 2 months ago (1 children)

If you allow them to make money with their garbage, they will be incentivized to keep going. I want them to go bankrupt.

[–] [email protected] 0 points 2 months ago

Until we manage to unite the workers in revolution, groups shouldn't be attacked solely because they're trying to make more profit for a little amount of harm. It's nearly the only way groups can survive if they're bent on being a company.

[–] [email protected] 12 points 2 months ago (1 children)

Good. I had a couple of answers to one of my questions that just wasted my time before I realised they were AI. The authors didn't get banned annoyingly.

[–] [email protected] 4 points 2 months ago

The authors didn’t get banned annoyingly.

Did they get banned kindly? Happily? If they weren't banned annoyingly, how were they banned?

[–] [email protected] 10 points 2 months ago (2 children)

But didn't they say they were going to use all questions and answers to train ai?

[–] [email protected] 26 points 2 months ago* (last edited 2 months ago)

That's probably the main reason to reason to ban AI. They want a mostly clean training set and they will probably add their own AI answers to each question as well.

[–] [email protected] 17 points 2 months ago

Yes. They just don't want the AI being trained on its own excrement.

[–] [email protected] 10 points 2 months ago

Fuck Stack Overflow. They've created a culture more interested in finding reasons to close questions than anything else.

[–] [email protected] 5 points 2 months ago

Anyone who’s unclear on how the whole “AI transformation” is going: a major Geek House just banned it because it sucks so bad.

Big Tech Corporations: your naked attempts to lock out workers from IT profits by laying them off in the hope that AI will replace them is starting to fail catastrophically. Here’s to the dim hope you get bounced to the curb by a vengeful board.

[–] [email protected] 4 points 2 months ago

“…may be revisited…” see you in 2 years of reverse action.