@fogti @mhoye I typically speak in melodrama. "nobody" is of course not accurate. but in my experience it's very very close to nobody

@mhoye nobody new wanted to learn before either. you saw less of these people because they had no way to submit plausible work. So AI has certainly enabled a new kind of spam. But this doesn't indicate the pre-existing, tiny pool of new people who genuinely want to learn is any different. there have been so few people like that for a long time. the "genuine learners" more often want to do their own thing, not work on *your* project. This is why AI is making such an impact on Linus and others.

@glyph people who are using LLMs to save themselves time and effort in this case are being asked to document all of their LLM use and all the time spent dealing with an LLM and then somehow produce equivalent data from before they used an LLM in order to prove what to them seems obvious. Until they can do this, you'll continue your assertion that they're just "wrong". Even though you dont have that information either.