Why is local on device ml a pro when compared to google photos?
Part of the value for photos for me is the fact that they do all the ml work ahead of time on their servers and not my tiny battery powered gadget. I can open it in on any device and search any photo instantly.
I am aware of the why, it's claimed to be e2e. But then it shouldn't be touted as a "pro" when compared to Google Photos which the site's home page is asking me to do. Hence the question.
As for iOS, Apple controls the whole stack. Whatever ML they do, they do it only on devices capable of that level of ML.
I can buy the cheapest android I can find with a 10 year old mediatek processor and search for pictures of my dog and it'll show it. There is a value to doing the heavy things on a heavy hitting server.
We have to pick our battles, and we currently aren't looking at serving customers on low-end devices. But if they access Ente on a laptop, the indexing will run there, and the computed indexes will sync to their low-end devices, e2ee.
Also, given that compute on smart phones is getting better and cheaper with every iteration, we believe it's best to bet on Edge ML for the long term.
Thanks for clarifying; you guys have built something amazing and also made it open source, which you didn't have to. Please don't take my comments as an attack or as demotivating.
I wanted to share my view as a regular person, sadly for whom absolute privacy is a feature and not a necessity, which I think is the vast majority.
You’re comparing apples to oranges. Google Photos is not E2E. Explain how you can do ML without having the key. Either you hand them the key or you don’t.
Of course they mention it as an alternative to Google and Apple Photos, but that doesn’t imply that they have 100% of the feature set of each.
