When I was teaching at Los Alamos National Laboratory in the late 1990s, a team of archivists was going around recording old-timers so that their tacit knowledge of how you actually do nuclear testing wouldn't be lost.
A few weeks ago, Bruce Schneier pointed at this Slate article
about pickpockets disappearing in the US because the mentor-to-apprentice chain had broken down. I think the same thing almost happened with small memory devices in the 1990s and early 2000s: most people were programming desktop machines with oodles of RAM and CPU that drew power from wall sockets, so they didn't have to worry about packing data structures to save a few bits. Now that mobile devices are all the rage, it's kind of fun to watch 20-somethings rediscover tricks I learned on a PDP-11/03 in 1981.
s/get off my lawn/hey kid let me show you a trick/