Hacker News Reader

Comment by collabs | Hacker News Reader

I am curious what if any progress has been made with free and open source software and computational photography. I don't have any expertise on this matter. I just remember about five or six years ago the meta was that free and open source apps such as open camera pretty much suck balls when it comes to any advanced photography that allowed the user to click and forget when the likes of the proprietary Google Camera app was where the "magic" happened.

View on HN

GOS has a camera app that works pretty well, but I prefer the proprietary Google one. You simply don't install any other Google apps or Google Play services in the same profile and remove its internet permission.

Now, it could technically only communicate with other apps that have Google libraries. I doubt these libraries include any code that does that. They are not that big, and we can reverse-engineer them very well nowadays with AI.

A targeted attack is still possible, but they won't systematically exfiltrate your pictures through other apps. That's way too much traffic. They would easily get caught if they did this, for example, on my network.

I think progress in that area has kind of plateaued no? We got to a point in which the AI post processing was too noticeable and creating weird artifacts (like the ones in those hyper-zoom pictures) and companies took it down a notch. You can probably find open source alternatives that can compete with closed source ones

Reply on news.ycombinator.com