With beanstalk, you're paying in terms of multiples of server instances at all times, even if nothing is hitting your servers. With Lambda, you pay only for the execution time of your function, which is only when a request is made.
I don't think those are programmed into the driving software as much as it has learned them through machine learning. Most of these systems have gone through the equivalent of millions of years of humans driving so I'd trust it over a human driver any day.
I've been using Docker in an end-to-end development to production environment for over 6 months now and was hoping to get some valid insight from your review. However mentioning something like that makes the article seem less credible (even though the rest of the article might have some valid points).
I find it bewildering that you ran docker in production for 6 months and yet you find that running your own docker registry a "complex" operation. Docker has its flaws, like any other technology, but running a registry took me all of 2 minutes on an EC2 instance. Kinda makes me wonder how much you really understand about a real docker workflow.