"What can one do in the face of a relatively shrinking population?" is the more interesting question to me.

As someone whose managed a team before, there is a minimum population of people practically required to sustain a particular corpus of actionable information without suffering severe degradation in terms of said information's application.

Once one ends up below that point; things tend to go the way "from scratch rediscovery required", until such time as the population of people capable of acting on it is restored.

Whether that actually happens is a prioritization decision balancing against everything else that still has to be done.

Have three boys! All successful engineers. One even knows hardware and assembler!