It’s plainly false for all sorts of reasons, but that’s the source of the disparity.
I think the implicit assumption has it the wrong way around. It seems like the industry takes about 30 years to catch up with the mathematicians who had discovered the solution to our problems.
We create this problem for ourselves. Some hot new web framework rediscovers the continuation monad and treats it as a novel invention. Giving it a slapping new name, a slick logo, and talks at the hottest industry events. A whole generation of developers learn to talk the talk, walk the walk, and hit a wall. The next framework comes along and suddenly they're, "too out of fashion."
Computing hasn't changed radically in a long time and certainly not programming.
This might depend on what you consider radical.
Web forms vs MVC vs react + restful API are fairly different coding paradigms.
Trying to create a windows service in net framework 2.0 is totally different than net core 5.
Or using SQLDataReader to read executed sql strings vs entity framework?
Even just dependency injection and scope definitions (transient, singleton, scoped) is very basic yet is something I’ve seen older developers tripping over. At least it’s an easy to identify run time error as it validates the container when running in development.
I’m not saying older developers are bad, but developers with the mindset that coding now is the same as coding 20 years ago are either using a language that never developed itself or are going to stumble over the changes that did occur.
You're still trying to accomplish the same goals, and there are only so many building blocks, so yes the big picture stays roughly the same and its only the details that change, but the devil is in those details. I find working with people who are not only ignorant of the details, but proud to recognize that they are "just" details, to be a nightmare.