Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

No need to "pass it through OCR" if you're viewing it in Safari :-)

(because it OCR's all images anyway, mobile and desktop both—I browsed to the page and select-all'd the image text just like it was HTML—still, they should have just made it HTML)



LiveText on iOS/MacOS is one of the best features they’ve ever added. So incredibly useful.


The Android activity switcher has been able to OCR whatever is displayed on the screen since 2018. Nice to see iOS added something like it last year. It's come in handy loads of times.


I wasn’t aware of that but it doesn’t surprise me given some of the other image intelligence stuff Google has had that Apple is slowly catching up on like recognizing animals/objects.

Already it’s saved me a bunch of time numerous times.


Apple tends to be a bit behind Google on AI since they like to do things on-device for privacy, etc..

Also sometimes they take their time in implementing something if they don't think they have the right UI for it (such as copy & paste, which was missing on the original iPhone.)

But I agree: Live Text is nice and it works just the way you'd expect it to.


Note that speech to text and voice assistant functionality happens on device on Android since 2019 for privacy (https://blog.google/products/assistant/next-generation-googl...), while iOS continued to send data to Apple until 2021 (https://www.theverge.com/2021/6/7/22522993/apple-siri-on-dev...). Neither privacy nor UI were an issue. Android used the on-device speech to text to transcribe phone calls, so I never have to listen to menu options repeat (though it doesn't yet automatically enter my SSN or account number), automatically caption arbitrary audio playing on the device, etc. These are other features I use frequently that Apple will get to eventually, but it's not because of privacy or UI.


iOS added live video captions this year. Would love to have it added to the phone next year.


I just went googling for it, wondering if the Android version is on-device or sends data off to servers to do it, and couldn't find anything about a device-wide OCR on Android. Guides to developing with MLKit, and something about the feature specifically in Google Docs on Android, but nothing about the device-wide feature. Bizarre.

Same search but "apple" instead of "android" (not even iOS, which is what I meant to search but messed up) turned up tons of hits about the feature on iOS and macOS, and confirmed what I thought, which is that it's on-device.

[EDIT] Mind, I'm not saying it doesn't exist, just that it's weirdly-difficult to find anything about with searches on Google's own search engine.


Apple marketing likes to name every little feature. This feature does not have an official name on Android. https://www.thurrott.com/mobile/android/165834/android-9-pie...




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: