For certain technologies, that can be true.
Although, even for those, we still have people in the legislature who are in their 60's, and will likely stick around another 15-20 years if they had their way. And they're likely behind that curve.
AI appears to be a different beast in terms of how quickly things go from "new to obsolete".
Being in the field, I go to the major conferences every year, stay on top of changes and new releases as much as one can, etc...
The features/toolsets change very rapidly.
AI-based tools - and the knowledge required to use them proficiently - have changed more in the last 2 years (and will change more in the next year), than Microsoft's Visual Studio offering (an IDE for programming for those who don't know) has changed in the last 10-15 years.
A person who was familiar with building solutions using Visual Studio 2012 (and haven't touched it since) wouldn't have that drastic of a learning curve if they cracked open VS 2022 and started tinkering around with it, there would be a lot that felt familiar. The same will certainly not be true for the various AI toolsets out there.