I get the reasoning, I just don't think it works. By the time you're 50 only a fraction of your knowledge is still relevant. Sure the fundamentals are the same, but the tech you're working with is probably fairly new. I've worked with a few 50+ guys who have been in the business for almost as long as i've been alive. One of the common traits i've seen is they've learned to not trust certain levels of abstraction, or techniques. Which is fine, but tech has progressed, and now many of the original reasons are no longer valid yet they're still resistant to using them.
The good guys are up to date in the state, they've learned the new features, they trust them, and they use them.. but staying up to date for 20 years is an exhausting race.
Personally I hope i'm only writing code as a hobby at 50.
I solve problems. My experience in problem solving will never become irrelevant. It happens that I solve problems with code. With my experience in problem solving, I can pick up "new tech" quickly-- primarily because there's really nothing new. Things get smaller and faster; the real technological advancements are in the physics of storage and processing. The "advances" in software seem to follow what The Graybeards Of Antiquity dreamed up in the 1960s, now with a slicker, graphically enhanced user interface.
I see so many less experienced software creators going nuts over something "new" when it's just repackaged from a decade (or three decades) ago. It's interesting to see reactions when asked "how's that different from Q?" or "but what about problem J?" and even "ah, that's over here in TAOCP ... you know Knuth, right?"
EDIT: relevant quote from another headline on HN-- 'Calling a hierarchical directory a "folder" doesn't change its nature any more than calling a prison guard a "counselor".'
The good guys are up to date in the state, they've learned the new features, they trust them, and they use them.. but staying up to date for 20 years is an exhausting race.
Personally I hope i'm only writing code as a hobby at 50.