I read the article as a general philosophy to spend more time on the product instead of premature scaling optimization. If e2e testing + ci/cd is cheap/easy and helps in your situation, then that's what you should do.
I feel the author is more criticizing the CTO who prematurely decides to rewrite the infra to handle scaling to millions of concurrent customers before they have those customers.
I feel the author is more criticizing the CTO who prematurely decides to rewrite the infra to handle scaling to millions of concurrent customers before they have those customers.