Just think about your typical polymath like da Vinci. Sometimes people can be really good at many things, and if they assume that they can only be really good at one thing, they may not even try to be really good at many things.

So my main point is: don't assume from the beginning that you can't do many things well, or it will be a self-fulfilling prophecy.

It really depends on what you mean by really good. I agree that you can't be really good at many things - but only relative to yourself. Meaning that if I (personally) practice Skill A for a week versus practicing Skills A,B and C in the same week I would almost certainly be better at Skill A.

But if you work on Skills A,B and C for a week and another dude works on Skill A for a week there's no guarantee that you're going to be worse than him. Depends on smarts, I guess.

I guess it's academic but inexplicably interesting to me.

I think your objection is a good one. Horizontal scaling in humans has low efficiency, so it's usually best to max out the benefits of vertical scaling first.

That being said, Scott Adams' career advice[1] seems pretty solid. In short, it is usually just as useful to be competent at three different things as to be extremely good at one thing.

[1] http://dilbertblog.typepad.com/the_dilbert_blog/2007/07/care...

Carl Djerassi, Cosma Shalizi, Tyler Cowen, Richard Feynman, John von Neumann. Djerassi was a Stanford chemist and has multiple works of fiction published, Shalizi is a physicist/statistician/sociologist, Cowen is an economist/public intellectual, neither Feynman nor Vonneumann need introduction.

http://moreintelligentlife.com/content/edward-carr/last-days...

Michael Crichton comes to mind too.