[jfim]

Right, that's why for example it made sense from Google's perspective to kill things like Google reader because that doesn't really move the needle.

The biggest issue with that approach is that they're burning customer goodwill, and that affects their odds down the line of getting customer acceptance. Stadia is just another project in a long line of cancelled projects that leaves customers screwed over and less willing to use Google products in the future.

Your point about Amazon is pretty interesting, IMHO I think that's why in the long term Amazon will keep on beating GCP in the cloud space. Amazon having such a large set of offerings is that it makes their cloud offering particularly compelling, especially considering they seem to keep old services alive and have relatively stable interfaces. Their offerings might not be as polished as GCP's, and some of them are laughably awful, but at least I feel confident that Amazon won't change the API or deprecate/retire the service.