Back

@omgubuntu I wonder how much of that is attributable to actual ML projects, and how much is slop generated by models with a bias towards Python output

@jmorahan I don't have any stats, but I'd wager you're probably spot on. I kind of alluded to that in the post by mentioning that Copilot integration with Github. Imagine quite a few hello world/demo projects being spat out that default to Python, and i'd wager - i haven't used copilot on Github - that it offers to create a new repo for the project too...