Your Pokemon remark seems in the right direction, I think disagreements here are perpetually going to revolve around what one's individual experiences have been and what other experiences one reads or hears most about and how much one lets all that impact what one believes about the "average everydev". At least until the CS departments worldwide take a break from the applied math, walk over to the social science departments, and ask how to start conducting scientific studies about subjective things so that we can collect some stronger evidence one way or another.
In the meantime, how is the (Bayesian) evidence from one's experience to be properly measured next to the existence of books, papers, journals, conferences, videos, some with easily accessible and fairly accurate sales or views numbers to prove a level of popularity? In my office I doubt more than 20 people, perhaps no more than 10, out of the thousand or so spread across the floors (which includes many non-devs) could name over 200 Pokemon. This doesn't negate the existence of millions of people who can do that and more elsewhere, and the millions of dollars in that market. But I think it is at least suggestive about the type of people my company, and companies like mine, tend to hire. Is it really a stretch to extrapolate and make a bet that "perhaps only 2% of salaried devs can name over 200 Pokemon"? How about something just as trivial, the names of all 50 united states?
Pokemon is Pokemon, it doesn't matter much for our field, but where the widespread self-hating of one's profession comes in is that depending on one's experience the bets might not seem to be that different for things more programmers (or other roles like managers) probably ought to know about, their field's own short history just one of them. I recently quipped to someone that I doubt more than 5% of developers are even familiar with the SOLID principles thus it's not very fruitful to ask about them in an interview unless knowing about them is a hard requirement. Even fewer have a good grasp of what each principle means, and even fewer know the trivia that the L is named after a woman (with even fewer still knowing the less trivial nature of what else she contributed to our industry). Maybe I'm over-extrapolating, or a bit pessimistic, maybe I'm letting my enterprise day-job and my conversations with others at places (both enterprise and not) jade me in my expectations for what I think the "average everydev" is like.
My underlying theory is that power laws are everywhere, which means if the distributions on facets like "knows Pokemon", "knows about SOLID", "frequents Hacker News", "knows important bits of computing's history", "writes/buys/reads technical books", et al. both trivial and non-trivial are power laws then the average level would be terribly low compared to the narrow peak on one side of the distribution. The mode, which is probably what influences a lot of personal-experience feelings about the average, is lower still. If your experience is mostly around the peak it may seem otherwise.
Who's to say how much it matters anyway, it's often a pointless discussion / complaint to me. Normal laws are also everywhere, and we know that at least IQ is normal. "Smart but ignorant" is acceptable for many things, and besides, ignorance need not be a permanent condition. One doesn't need to know that Lisp machines in 1990 could do full 3D rigging, editing, and effects (https://www.youtube.com/watch?v=6VmJVNYfxDc) in order to be effective for their Java shop employer.
