Scott Alexander on the idea of a “dualized” field, where opportunities are few on the ground, but the rewards for “making it” are huge. This principle applies to gangs, colleges, and medicine:
This concept applies much more broadly than just drugs and colleges. I sometimes compare my own career path, medicine, to that of my friends in computer programming. Medicine is very clearly dual – of the millions of pre-med students, some become doctors and at that moment have an almost-guaranteed good career, others can’t make it to that MD and have no relevant whatsoever in the industry. Computer science is very clearly non-dual; if you’re a crappy programmer, you’ll get a crappy job at a crappy company; if you’re a slightly better programmer, you’ll get a slightly better job at a slightly better company; if you’re a great programmer, you’ll get a great job at a great company (ideally). There’s no single bottleneck in computer programming where if you pass you’re set for life but if you fail you might as well find some other career path.
I have more thoughts coming on this, but I believe Agile (and Scrum in particular) causes dualization in engineering departments that adopt it. It collapses the number of different ways that engineers can measure their own success relative to their peers–Bob might be into databases, while George likes type theory. In Agile they’re just team members that rise and fall–or more accurately, don’t go anywhere–together. At the same time, the system forces them into a generic engineering role that ignores their specialization and preferences.
…[D]ualized fields are a lot more likely to become politicized. The limited high-tier positions are seen as spoils to be distributed, in contrast to the non-dual fields where good jobs are seen as opportunities to attract the most useful and skilled people.