Submitted 3 days ago by HubertManne@moist.catsweat.com to mildlyinfuriating@lemmy.world
I find the most common thing I use ai for is to give me the text from a screenshot due to websites putting important information in a manner that does not allow copy and paste of it. So I should completely not need to do this but it comes up a lot.
If you are using Windows: learn.microsoft.com/en-us/…/text-extractor
Yeah, AI text extraction can be very powerful. It’s arguably the thing that neural network systems are best at, and one of the first things. We got them to do.
There’s alot of promise in using this kind of tech for quickly digitizing massive swathes of books. The more complicated part, however, is converting any images in-layed in the text.
There are nots of use cases for LLMs. The thing to remember is that the output can be seriously flawed and is not your property.
Man, before “AI” became a marketing gimmick,
Samsung Note devices allowed you to use the S-pen to circle text and extract it.
I remember using this feature at least a decade ago!
Can’t you just copy it as normal when holding alt or control (can’t remember)? There is an option in firefox about:config to ignore what sites want and re-enable copy.
If you are using a chromium based browser, project naphta allows you to select text in images directly in the browser.
Presi300@lemmy.world 19 hours ago
Main use for ai: glorified email template generator