I guess it depends on what scale you are talking about. I used GCP at a previous employer, a small saas startup, that was dealing with thousands of api request

I guess it depends on what scale you are talking about. I used GCP at a previous employer, a small saas startup, that was dealing with thousands of api requests per second, and lots of big(ish) data processes on the back-end. We made use of GKE, Bigquery, Pubsub, Stackdriver, GCS, Compute, Dataproc, Cloud functions, Data studio and Cloud SQL mostly. I had a really positive experience with GCP - I think for every one of those services, the workflow and developer experience was far more pleasant and easy going than my experiences have been on AWS. GCP just clicks with me more, for some reason. To me, it feels like they put more thought into their suite of offerings and how they all work together. Lots of the services on AWS have a great deal of overlap. That alone can waste a surprising amount of time, just figuring out which you ought to use for your use case - which is not always easy to discover. I'm constantly underwhelmed and disappointed by most AWS services, and keep finding that I'm better off just using AWS as a glorified ec2 VM hosting service.