A wonderful presentation by Dan McKinley of Stripe on how technology companies make mistakes in choosing the next piece of shiny technology instead of relying on their existing and tested stack.
He outlines how this increases the overall cost and how this cripples the development team in many ways. I have to admit that I used to do this myself for many years and would go for the next framework, the next data store, etc.
I learned the lesson the hard way, and thankfully since then, I have chosen the fewer, the most known technology that serves the need.