Jonathan Blow

What if we looked at the productivity of the industry as a whole instead of the productivity of the average programmer? There are definitely perf and productivity taxes inherent in high-level languages but would you agree that they've increased the no. of programmers available?

It definitely takes more effort to reach a base-level of effectiveness with low-level languages. So perhaps low-level programmers appear more productive because they have to become better programmers earlier.

But that additional effort has a cost. I'd guess that there are many successful products that wouldn't exist without the economics of high-level languages. It's still unclear to me whether that is a good thing now or in the long-term.

What economics of high-level languages? Have you seen how many engineers Web companies employ? It's crazy.

There is certainly some inflection point of cost/complexity with systems written in high-level languages, and the big Web companies are all past it. But there are probably a lot of other things going on there as well (politics, lack of focus, massive investments tracking/adtech)

When I say the "economics of high-level languages" I'm talking about the cost of training/hiring n programmers to build some application. The quality of the end product is of course lower, but our economy would at least look very different if the cost of new software was higher

I agree that this is the common belief, I am just saying that I see no evidence supporting that belief.

Do you think there is a non-economic explanation for the popularity of high-level languages as the software industry has grown?

How is it very low if it hasn't been measured? Sure, development still takes ages but at the same time complexity has gone through the roof. Although, tooling and frameworks matter quite a bit there as well, I imagine.