This thing is super cool! One other use case I can think of, is to use it to check reliability of the fonts that I chose in my designs, only if the option to choose a custom font was given.
Some features took a longggg time to do, such as table extraction, text editing, and (surprisingly) preserving positioning of elements (text, images etc.) when rotating the page in the downloaded file - PDF specification has a different orientation system than the web, so this was very intricate to get correct.
A lot of PDF editors have tools that all work independently, meaning you have to use each tool separately. My decision to add all the features I did while keeping it in one editor was because I felt that was a better user experience, but I means that all features become intertwined, which added a ton of complexity managing that.
Considering the rate of improvement of these LLMs, wait for a month or two and then you may not even need an os, let alone some obsecure piece of software (shell).
Except the models don’t actually compute anything other than text generation. The entire way they interact with computers is through the shell or other api layers on the OS
reply