They're not the fundamental constraint in work output differences. IQ/ability gaps matter way more than they do in programming than they do in construction work.

Let's not forget that there are unions for highly skilled labor as well. When I was a classical musician I belonged to a very robust and very helpful musician's union. Everyone I know _always_ had a preference for union gigs as they usually paid better and ensured that you didn't have to deal with all sorts of nonsense from your employer (like rehearsals running overtime, not getting paid in a timely manner, blind auditions, etc).

Professional classical musicians require a great deal more training (and talent) than programmers to get to the level where they can make a living doing it.

Blind auditions are bad? I thought they would be preferred by the musicians.

My mistake, I mis-edited that. Blind auditions are preferable. The musician's union helped ensure that auditions were blind.

Whew. Thanks!

It does, but only at a very low level. Most in the profession are several layers removed from directly worrying about how fast the bits are moving on a CPU, and are more concerned with adhering to sound design and algorithm patterns.

The poster you replied to never said that. He said that the laws of physics matter far more to traditional engineering than software engineering. Unless a software engineer is working on algorithmic complexity, low level hardware/security, or distributed systems, real world physics usually don't matter too much.