I wrote not that long ago (see article in Diginomica)that the future of serverless is not Lambda, but is technologies such as AWS Fargate. I truly believe this is. Lambda is very useful for some kinds of computing needs, but it is not suitable as a general serverless solution to replace standard programming methodologies for building services and systems.
I was therefore very happy to see this article, written by Brendan Burns, come out and talk about the future of Kubernetes. He makes many of the same arguments and he is right on in his analysis. I highly recommend anyone interested in this subject to read his article: