This is one of those days when I'm glad I don't depend on AWS (or GCS) at work. At a large enough scale it can be hard to avoid using other people's computers, but many websites really don't need to be distributed across timezones, and thus tolerant of faults except those introduced by your cloud provider. I have the same feelings about cloudflare, which is a brilliant service and has some fascinating people working there, but centralises a huge number of sites behind one point of failure. Then again almost everything on the internet relies on other people's infrastructure at this point - this site depends on Digital Ocean, and the CoreOS servers not being hacked (updates are applied automatically). So nobody can really gloat about this event and feel safe about their own infrastructure. Even the most reliable services go down.
Minio is a nice alternative to AWS though, and you can easily self-host. My biggest concern with services like AWS is not the potential for downtime but the lock-in. If you run your own code you can easily move it anywhere, if you depend on AWS semantics and silo your data in Amazon services, it's very difficult to move on if you ever need to.
I totally agree with you on this. This is the reason I posted this, we should protect new devs and organisations. We all seen lately (the last year even more) new developers are so enthusiastic about 'serverless' and they don't think it clearly. Yesterday(or today, it depends xD) was a good lesson for all of us.