I’ve got very mixed feelings about computational photography. I like that it can improve image quality with zero user effort. But where does that end? Maybe the image would improve by reframing it slightly - or reframing it a lot. Maybe the overcast sky could be swapped for clear weather. Add a few dramatic hard shadows. Tune up the depth of field. Heck, let’s just look up the GPS coordinates and auto-download a photo of the same location from Flickr that was taken by a better photographer.

I would honestly pay for a database of inages that could give me professional shots of locations I have been to.

Is the problem that the photos need to be photoshopped to hell by a professional rather than an algorithm?

Google Tensor SoC debuts on the new Pixel 6 this fall - https://news.ycombinator.com/item?id=28038807 - Aug 2021 (96 comments)

As an iPhone user, I hope it's good! Apple could use some actual competition, and Qualcomm does not seem at all interested in providing it (Huawei did, but we saw how that went…).