StackOverflow and OpenAI are getting together

This page has 0 threads | Add post

Blogs

Finding our blogs useful? Please consider helping us to share more!

StackOverflow and OpenAI are getting together - not everyone likes the idea
StackOverflow and OpenAI have announced a partnership, but I think it's fair to say that not every StackOverflow user is sold on the idea ... !

Posted by Andy Brown on 05 June 2024

It was probably bound to happen, but it's made a lot of people very angry!

This partnership was announced in early May, with more detail on how (whether!) OpenAI will attribute contributions here.

Here as far as I can see is how this deal will work:

Party	What they get out of it
OpenAI	They will be able to include the vast (and useful) trove of StackOverflow data when training their large language models.
StackOverflow	Less clear: OpenAI have said that they will include attributions to StackOverflow, but it's not clear yet what this means.

I'm not sure what response StackOverflow were expecting, but surely it can't have been this. Some people have protested by changing their user names - for example:

This angy former contributor has changed his or her user name to point to the alternative Q & A open source community-based site Codidact.

Here's another angry user, and name change:

It takes a lot to stir up angry reactions like this.

A few people have pointed out that StackOverflow aren't the only ones to break their contract - Darth Vader did it too:

To explain why people are so angry, here's one person's experience (mine). I've been a member of StackOverflow for over 13 years now, and my answers and questions have put me in the top 10% of users:

I've asked 34 questions and provided 108 answers.

I provided answers to help other people like myself. I was vaguely aware that StackOverflow was privately owned (it turns out it was sold for $1.8 billion to Prosus, a Dutch company, in 2021) but I was happy to share information knowing that it would benefit others.

I'm considerably less happy that my answers are now being fed into large language models being created by OpenAI, which will mean ChatGPT is using - and probably distorting - my answers without offering any attribution, acknowledgment or recompense.

However, I definitely don't feel as strongly as the many angry people who have commented on the posts, a couple of examples of which are shown above. I will continue to use StackOverflow, but I suspect I will be turning to Codidact to share any insightful comments that I have. We shall see! I'd be interested to hear any other readers' views on this.

Some other pages relevant to the above blog include: