> The biggest barrier was time

I am going to agree, albeit from an entirely different perspective.

My job has nothing to do with programming. Aside from administrative tasks, my job has nothing to do with computers. Yet I know how to program and I realize that automating those administrative tasks will have huge benefits for me. The problem is my employer doesn't pay me to write code, so I never got around to that automation.

Realistically speaking, writing that code without AI would not offer much in the way of understanding or expertise. Most of it is: "look up X, if A then B, else C, iterate" type tasks. There is a barrier in knowledge with respect to library calls, but that's pretty close to junk knowledge (it's rapidly obsoleted, and useless if I need to interface with different software).

So the choice is between doing without, or spending a lot of time picking up tasks that are of no value to me. Time was the barrier.

These arguments also remind me a lot of the microcomputer revolution: a heck of a lot of mainframe COBOL and FORTRAN developers were critical of underpowered computers and toy programming languages. The reality is that it opened up opportunities for a new generation of developers, and new types of software was created because the problem solvers were closer to the problem.