Egregoros

@lain wait, finally something better than calibre for managing ebook conversions

@cell yeah it's a lot better. muse spark also has vision so it can just look at the page if the first OCR run produces nonsense.

@lain this is very cool, PDF are a pain to browse on my old Kobo. Muse Spark is definitely a really capable model

@foolghool really an amazing model, especially for the price. happy that meta is back in the game

i used to not read a whole bunch of books i found becuase they were available only as pdfs and those are bad to read on an ereader, but you can just use muse spark for free on opencode to convert and proofread and the result is better than most epubs you pay money for

@lain I'm currently using it to replace the language/client part of SuperCollider into an Elixir library. A bit of dancing around at the beginning but I had the thing in a functional state after a day of work. Dream territory 2 years ago, and here I am rewriting and live-testing old patches within Emacs

I grew up researching using microfilm and micro fische so pdf reading is only a problem if it was scanned by a half blind lepper or my hardware is only purpose was reading a handful of niche file types.
Holding onto a nice reader tends to fall in that rank and file, so I get that.

@HuWhite_Wall

@foolghool yeah, even if they don’t one shot it, the sheer productivity gain is insane