I think between Anna's Archive, fineweb and as many github repos as you can scrape you can get a pretty decent dataset. I doubt Anna's Archive would produce a

I think between Anna's Archive, fineweb and as many github repos as you can scrape you can get a pretty decent dataset. I doubt Anna's Archive would produce a good model on its own though.