Post

Holy shit.

Hacker News Top Stories@hackernewsbot.bsky.social

Total monthly number of StackOverflow questions over time | Discussion

data.stackexchange.comdata.stackexchange.com

Peace

@techno-hippie.bsky.social

4 Jan

Why would you go there and endure the abuse of drunken sweaty old Coders when chatGPT is so respectful and correct?

Bill Criswell

@crswll.dev

Totally understand the criticism but without the sweaty old coders ChatGPT would have a lot less to work with. I always thought of it as a tax haha.

4/1/26, 1:31 am0 reposts10 likes

Peace

@techno-hippie.bsky.social

11 Jan

Seems like GitHub is a better resource. Perhaps why copilot is as good as it is for vibe coding, though Claude really seems to be the stand out for me right now.

Lawrence Jones

@lawrencejones.dev

4 Jan

I do wonder how much SO actually factors into the training set. Feels like not very much, especially with the new synthetic data training approaches.

Sam Rose

@samwho.dev

4 Jan

As a percentage it’s almost certainly ~0%. But I suspect it’s probably a meaningful source of much-higher-than-average-quality data.

Lawrence Jones

@lawrencejones.dev

4 Jan

Yeah hard to know. I would imagine a negligible impact nowadays but may have been a lot more important getting things off the ground initially. Expect it’s not something any foundational provider would be willing to share 🙈

Sam Rose

@samwho.dev

4 Jan

I really need to spend some time looking at the history of synthetic data use in training. I’m vaguely aware that it’s become standard to use it but I’ve not tracked when this happened or how it was evaluated to be beneficial at all. I remember lots of fear of model collapse a year or so ago.