Hacker News Reader

It's likely to avoid crawling, especially to train competitors to GitHub Copilot or other LLMs.

Why can’t you train on the git repository itself? Clones are not behind a login.

How do you discover the repository, especially in an automated fashion?

The way AIs are evolving, I wouldn’t be surprised if they have the ability to go to GitHub and search on demand based on a users query.

This seems like the most probable reason of any I’ve read so far.