I'm not sure that this is true even in software. "Professional" skills, interactions, leadership and other skills can indeed be taught, but are difficult to actually test for. There are a lot of skills in software that are invaluable and come only with time, experience and self-reflection.
Also, I find a significant divide in working with people who have come out of bootcamps, etc at a junior level. The divide is really down to those who did work outside/beyond their assigned work, and those that didn't. It's an indicator of passion and aptitude that isn't really testable because then that would become the new baseline.
I'm not saying that reform isn't needed, I think there's a significant space for it. For that matter, changing perceptions of what higher education actually is and isn't good for would be a start. In my own experience, the top performers out of a vocational program or self-taught will outperform and do better than the vast majority of those traditionally educated. The rest will stack in the middle, and those that didn't do well in the vocational setting will be at the bottom.
I think the start is to recognize that procedural differences can and do have overlapping outcomes, and to look for those secondary indicators and try to acknowledge them.