I simply do not care about deep learning and NLP advancements. They invariably end up used to recommend me more garbage that I do not want. The only "ML" feature I can think of that's made my life any better? Facial recognition in Google Photos, because it's easy to find pictures of a specific person or animal. But that still mixes people up routinely, and is by no means comprehensive. Not to mention the fact that it's likely used to profile and track me behind the scenes at Google. I could take it or leave it.

I prefer DSLR pictures to smartphone photos, so I don't really care about ML image processing pipelines in phones much. Maybe night sight counts as an advancement, but I don't use it. Again -- DSLR, if I care about the photos at all.

The Pixel phones have some neat "ML" features, like instant translation and mostly-good transcription, that are definitely neat. But I'm not convinced that those matter much to me.

Cloud orchestration innovations have surely happened, but again, I'm not convinced they're a net good for the world. Kubernetes? Sure, it's impressive. In some cases it's damn useful. But it's also a great way to overcomplicate cloud machine management. Most smaller companies would probably do better with physical machines, cutting the cloud out entirely.

TL;DR: I can feel and see Apple's innovations. Google's are usually more indirect... and even when I can see them, I'm not convinced they're true useful advancement. They often feel like gimmicks.