It's an absurd claim, of course (not that capital and labor are heterogeneous - but that this fact is ignored).
One of the things that a lot of commenters on his posts seemed to get caught up on was the use of aggregates like K and L producing an aggregate Y. Aggregation somehow meant homogenization to them. It was never explained. I'm not sure why they drew that inference. But it was there.
I lectured at my first undergraduate macroeconomics review session tonight, and this issue of labor and capital heterogeneity actually came up when we were discussing an issue addressed on the second day of a freshman macro class: the production possibilities frontier (PPF). One of the things the students didn't quite get was why we assume the PPF is usually concave.
The answer, of course, is that different labor and capital is better at producing some things than others, and you can't just switch them around and expect them to perform as well. Labor and capital are heterogeneous. I didn't use the word "heterogeneous", but the point comes across very clearly when you start at one of the corner solutions and verbally walk through the concave PPF.
The goods the students settled on were beer and mountain bikes. Fair enough. Why would we expect a PPF of an economy producing beer and mountain bikes to be concave? Well we assume that there's a range of brewing productivity and mountain bike manufacturing productivity across labor and capital in the economy. Let's first imagine we're at a corner solution: all workers and all capital produce only beer. Some workers and some capital, though, are ill-suited to this work. So if we - sitting on this corner solution - decide we actually want some mountain bikes and we're willing to give up some beer to get it, we'd naturally release the labor and capital best suited to make mountain bikes and worst suited to make beer first. That would result in a relatively low sacrifice of beer (since that labor and capital wasn't producing much anyway), for a relatively high increase in mountain bikes. As we continue along the PPF it becomes harder to find people with such a specialization in mountain bike production, and to get more mountain bikes we have to use people that are pretty good at producing beer. The opportunity cost of producing mountain bikes gets higher, and the slope of the PPF gets steeper.
The point is these ideas about capital and labor heterogeneity are implicit in all of economics. You can get a concave PPF in other ways, of course, but this is the most obvious thing that people turn to. Perfectly interchangable workers and capital result in a linear PPF, and nobody thinks that's what we face. The same with specialization and trade. How do you teach specialization and trade without heterogeneous labor and capital? I just don't know.
For the life of me I have no idea why Steve, Troy Camplin, and others insist that this is such a mystery to other people. Specialization has been the core of economics since Adam Smith. I can't even conceive of how economics is done without capital and labor heterogeneity. I know at American University we teach it to freshman in the first month of class - and we're known to be a fairly left/Keynesian department.
Keynesianism is also simply inconceivable without capital heterogeneity especially. See why here.