New technology has a learning curve and by adopting it, it slows us down.
If we adopt more than 1 new thing at a time we have a multiplier effect. They exponentially slow us down. Only do one new thing at a time.
When exploring new things be sure to fail fast. Don't go deep and then realize it didn't work.
Given the cost we have to have a measurable ROI on why we would do some new thing. We don't adopt tech because we like it. We adopt it because it will give us X within Y period of time.
Make sure the ROI takes into account the “old thing” legacy deprecation, upgrade and conversion cost. You are now maintaining two things. The old and new. They have to be compatible with each other or have some migration path, and you have to realize you are now doing two things until the “old” is gone.
And given all of the above we make sure that we fully adopt, perfect and realize the ROI on the "new thing" before we consider another "new thing". Don’t let shiny object syndrome kick in, no "squirrel" allowed.