We are pretty close to replicating compression in AI. At the end of the day, you’re better off capturing more info/detail and post-process. You also have to think about re-processing images down the road when tech improves.