[dupe]<p>Lots more discussion:<p><a href="https://news.ycombinator.com/item?id=40297027">https://news.ycombinator.com/item?id=40297027</a><p><a href="https://news.ycombinator.com/item?id=40302792">https://news.ycombinator.com/item?id=40302792</a>
Discussed the other day in a consolidated thread from a few duplicate submissions (311 comments total):<p><a href="https://news.ycombinator.com/item?id=40302792">https://news.ycombinator.com/item?id=40302792</a>
> Stack Overflow has been banning users wholesale who have attempted to delete or deface their own posts on the site<p>Its not just people who are upset getting banned for being upset, its people who are attempting to burn it all down on their way out the door in protest.<p>I get where the protests are coming from, but the cardinal rule of online communities is "once you post it, its out there". Other people have reacted to it, replied to it, quoted it. You break not just your own content but entire discussions if you mass-delete your contributions. They stopped being exclusively yours to take back once you contributed them to a broader conversation.
It's not clear to me what the deal is supposed to be about. Isn't it expected that Stackoverflow questions and answers can be used by anybody including to train models?<p>If Stackoverflow is trying to make exclusive deals with Openai, that is against the collaborative spirit of the platform, and I will stop contributing. After all, Openai is charging people for service. If Openai are the only ones given access, Stackoverflow becomes a gatekeeper, peddling my contributions. It'll beget a fork.
I don't get it.<p>Like, it's one thing if a fact-collecting business like the NYTimes doesn't want its stories to train an MML. I think under current law they don't have much of a case, because <i>facts</i> aren't copyrightable, but there's a reasonable argument that the law should be updated somehow in light of technological change.<p>But the work produced by all StackExchange users is <i>explicitly</i> released under a CC BY-SA license. The whole <i>point</i> is to collect and publish facts/ideas/understanding for anyone to see and use for any purpose, including running a business. Yes, the "SA" (share alike) part means if you want to use and modify the <i>words</i> then you need to release them under license that is at least as permissive, but LLMs aren't using the words; they are clearly digesting the facts and expressing them in their own words. And, unlike the NYTimes, there is no issue of "couldn't new tech undermine society's current method of economically incentivizing fact-collection?". The StackExchange users are not being paid, and the fact that the license is <i>not</i> NC (non-commercial) explicitly means that using their hard work to make money is allowed (and encouraged!).
StackOverflow is destroying its brand one step at a time. ChatGPT is good for boilerplate and generic errors. I still use SO for complex errors that are still getting updated with new ways to solve that said error, etc...<p>Without the community posting and answering questions, ChatGPT won't work anymore. They (SO) have to block any scrapping, wait until GPU prices go down, or use open-source LLMs with their data and figure out how to monetize that service (maybe giving some percentage to devs, etc...) or even check if that approach (add a chatbot service) makes sense.<p>They just caved to the FOMO mentality without considering that tech is always evolving and they need engineers, dev, etc... to keep finding out bugs, writing about their experiences on how they solved those errors, etc...<p>We'll see if a different contender figures this out and comes out with that solution if SO doesn't change its course.
Fascinating. Perhaps Stallman's greatest innovation is copyleft. The idea of required reciprocation seems to tie deeply into people's views. For the little OSS I have I picked it by default but would gladly use BSD or some safe PD license on the other hand.<p>The idea of Pillaging the Commons is interesting. I wonder when the mainstream opinion started shifting. It wasn't quite sudden but it seems to me that even ten years ago the dominant Internet visible position was that much of copyright was bogus: information wants to be free / if a pirate copies your stuff, you still have it<p>But now there's a stronger sense of "this information is ours". Perhaps that subculture moved somewhere and this one came here or perhaps I moved from where the former was to where the latter is.<p>I find it an interesting sociological phenomenon.
Ultimately they are digging their own grave (who needs SO if you can ask ChatGPT and shell out a few bucks to Microsoft eh?)<p>If anyone else thinks it's a bad move, the most efficient way to boycott it is by adding senseless questions and answers and upvoting them. Us people is how they got big, us people is how they go down.