I continue to not really be sure how best to get started developing Jedi. Like, there's tests, so I can't break anything, but it's also not obvious to me how to correctly add functionality.
This kind of blockage is one possible explanation for why I'm circling back around to random old projects. I've got some old machine learning code that I put together in an attempt to put the underlying techniques into terms I can understand. It's really slow, but I think I've worked out a tolerable way to make it faster. I'm kind of leery to write these ideas up, because I'm not at all confident that what I did is "correct" or "sensible", and the right answer to what I was trying to do is probably "use libraries". I'm also following along with tutorials, and I'm kind of like, wow, Anaconda impresses me less and less the more I use and hear about it. Like, is there some way to set it up so it won't break Homebrew packages that depend on Python? I'm currently getting by with pip, because I really don't see how Anaconda is supposed to be worth it.
Oh well, that's it for today. Good night.