I think the issue there, to continue your analogy, is that some carpenters are busy creating new levers, while others are busy repeating the same carpentry over

I think the issue there, to continue your analogy, is that some carpenters are busy creating new levers, while others are busy repeating the same carpentry over and over again. Same trade, different aims. The "multiplicative effect" is easy to translate to economic factors: it's what allows software-based companies to have very high marginal profit. The average net profit margin of a small business in the US is less than 10%, but in scaled, software-centered businesses it's not unusual to be above 30% or 50% due to the zero marginal cost of production. But for every company like Facebook, Amazon, Oracle, Google etc which understand this and plays to it to shoot for sky-high unit margins, there are N companies like Tata Consultancy, Accenture, Capgemini (or in-house IT departments) which effectively employ software engineers to do non-reusable "glue" tasks where the margin expectation is topping out somewhere between around 10%. These folks cannot pay as much for talent, in part because their stock isn't valued nearly as much, whereas it is a relatively cheap source of compensation for tech companies. A key to understanding the difference between US and other markets is the 1:N ratio. Software engineers in the Bay Area are gifted with very low ratios, and places where the ratio is higher in my experience have much more depressed software engineering compensation because there are not enough high margin software businesses to drive them.