Very roughly speaking, fast and cheap things can be great end products, but bad ingredients. If you want to buy a cabinet for normal use, fast and cheap is pro

Very roughly speaking, fast and cheap things can be great end products, but bad ingredients. If you want to buy a cabinet for normal use, fast and cheap is probably better, as long as it doesn't fall apart. But imagine that you would want to buy hundred cabinets and build a huge block out of them, or something like that. At that moment the weaknesses of individual pieces add up, and the whole structure is likely to collapse. To make the structure not collapse, you need to build it from reliable and thus more expensive pieces. (And, as you said, economies of scale can sometimes do something about the high price.) I think the frustration many programmers feel at their jobs also has a similar origin. We use dozens of "fast and cheap" tools, and each of them individually is better than nothing, but using them all together means you spend a lot of time angry about this or that problem. (And when finally most of the annoying problems get fixed, a new programming language appears, and everything needs to be rebuilt again.)