wagesj45

I know it’s pandering to my millennial nostalgia, but they’re doing it so well.

Looking at the documentation it looks like it relies on Mistral’s python tooling to work. I’m fairly dumb, so I don’t know if the tool suggestion coming from Mistral is from some kind of separate neural net or as some kind of special response you have to parse (or that their client parses for you?).

Mistral Instruct v0.3 added in function calling, but I don’t know if its method for implementation is the same/compatible. Also, it is fairly new and wasn’t released all that long ago. Hopefully we’ll get there soon. :)

I don’t think that Libby itself is. There’s DRM and while there is probably a way to strip it, I don’t think that is easy and/or publicly shared. But Overdrive, which is Libby’s predecessor, allows for DRM free MP3 downloads. But they’ve been trying to sunset Overdrive for a long time. The Windows desktop program needed to download the MP3 files is no longer linked on their site, for example (but is still downloadable if you know the exact link). I’m honestly not sure why it even still works unless it’s to comply with some ancient contract they have with a library somewhere.

Unironically the library. Then just use something like Audiobookshelf to organize your collection.

The courts get things wrong all the time. I may have to legally abide by their decisions, but I am not morally bound to agree with them.

Courts guarantee an answer, not the right answer.

I don’t buy that legal theory. AI models as they’re trained are the very definition of transformational. It’s fair use.

My internet experience has been slower since switching to Mastodon and Lemmy/Kbin. And it’s so nice. The things I see are more interesting. The conversations are usually more well thought out. And lowest common denominator dopamine content isn’t being driven into my eyeballs by Algorithms. I’ve legitimately been happier since the Reddit API debacle.

Long live the Old Internet.

As always, there is an XKCD for that.