> Even from a business perspective, what you're working on today, if it's tech or software related, will be horribly out-dated in 3 years.
Here's how outdated in tech usually works across three years:
We gotta switch to NoSQL, it's the new new hot hot thing! It's going to save us money, it's going to make our service much faster, it's going to be easier to manage. It's revolutionary, everybody says so. Our competition is switching too.
Three years later: we're going to switch away from NoSQL, turns out we didn't actually need to use that. It caused more problems than it solved in our case.
Outdated in three years is mostly only true in a small echo chamber of the tech industry obsessed with chasing fads 24/7. It's an almost entirely self-imposed form of torture by that part of the industry. There are exceptions of course, cases where things legitimately get outdated in just three years, and they're rare.
Was the NoSQL fad that so many people chased necessary? Nope. Are there great use cases for NoSQL systems? Of course. The legitimate use cases aren't fraudulent, it's the fad that is, the mania, the hype (which always implodes, then the headlines come rolling in about switching back).
That's a microcosm of the three year fad hopping that you're talking about (and there are obviously a lot of examples of it, from big data to AI/ML, to the latest framework craze). A small segment of the tech industry is like a dog that never stops chasing fad cars.