You must log in or register to comment.
Will Chrome send all your PDF/images/browser history to Google cloud to provide that service?
Or will the processing be completely local to preserve privacy?
You already know the answer
Yes, but it’s sad this isn’t even addressed in press coverage.
Honestly if they said it would be local, would you trust them?
This is just ensuring that companies are forced to blacklist Chrome if they want their secrets to stay secret. It’s already happened at my partners workplace (power industry, federal regulations on security) - hilariously, all google cloud services are blocked, but Bing is fine (w/ automatic ChatGPT integration).
It will be very interesting to see how companies handle this type of practice in the long run.
And I’m so happy to see Firefox doing the exact opposite, with the recent inclusion of an offline translator. Though not as good as Google’s yet, it’s already usable and will only get better.
The issue with OCR’ing pdfs is typically that it doesn’t understand the document formatting. So if you’re reading a document which is formatted as two columns per page, the OCR text will be a mess.
I’m willing to bet that given that most scientific papers are in that two format column, this ocr will take that into account or it’s dead on arrival.
I have this trouble a lot when trying to digitise or refresh some of the older documentation our company uses. It often becomes easier to just recreate it, by the time you’ve fucked about trying to restore the formatting.
So this is confined just to Chrome on ChromeOS it sounds like. I mean, ChromeOS was already an anti-privacy hellscape… anyway I sure hope this doesn’t spread beyond ChromeOS.
So it’s all about gathering content for training AI these days, isn’t it?
This is exactly what I thought clicking on this post.
