Comment by dan-robertson

It seems kinda useful to me. Like, the normal problem looks much less like ‘we took competent professionals over a long time’ and much more like ‘we took 10 undergraduates over a month or two’.

If you add up employed engineers at google-like companies (by which I mean large tech companies), you’ll quickly come to hundreds of thousands, and there are plenty of other companies who hire similar people. I think Google employees are reasonably uniformly distributed across that population (that is it isn’t like the Google has the best X% of them) and I think a study of such people is more interesting to me than a study of undergraduates (or indeed a study of programming language enthusiasts). I think an unrepresentative thing about Google is that they may use C++ more than many newer tech companies, and this might help for learning that.