My personal read: there are lots of people that didn't "fit in" as a teen/young adult. Instead of learning that excluding people for being different is ignorant they learned that one day they would be the one doing the maltreatment, at which point it would be totally cool.
I think part of the reason is because in software engineering there are several "big problems" that come up time and time again, but in different flavors. For example, when AJAX was the buzzword of the day, there were many engineers I knew who worked during the 90's cynically complaining about how we were just revisiting the "thin vs thick client" issue once again (much like their 'ancestors' felt like they were revisiting the "dumb vs smart terminal" debate in the 80s.)
The problem is when these problems re-appear, problems that you personally grappled with for many years, it's easy to just want to throw up your hands in frustration because you thought the issue was settled. But of course, in reality its not about the issue being settled or not, it's about the capability of technology evolving and macroscopic design patterns falling in and out of favor due to the particular tradeoffs available with the state-of-the-art. But it's hard to see this objectively when you had a lot of emotional investment into any problem, developed some expertise, and see your entire way of solving problems being thrown out for what you perceive as a fad.
A contemporary example that I've mentioned already is the recent abandonment of the relational model/ACID in databases. A cynic will tell you the people doing this are stupid because it's an inferior solution. But they fail to mention the reason people are doing this: it's because the relational technology has not caught up with the scalability requirements of many problems faced by engineers today. An experienced engineer with the right attitude can have an excellent perspective and hedge their technology bets accordingly: the people deploying NoSQL are not stupid, they are just getting their jobs done. But at the same time, this too shall pass, since the relational model and ACID are fundamentally good ideas, and eventually someone will swing the pendulum back that way once other problems are solved. So, for the experienced engineer who takes a longer view, certain types of NoSQL databases can serve a role but are likely going to be considered a transitional technology and systems should be architected accordingly. The inexperienced engineer will not recognize this and will go all-in with NoSQL and fall right into the traps that relational modelling and ACID are designed to solve. Lo and behold: Google Spanner paper is published -- but we're not on the other side of this cycle yet until this tech is commoditized.
This tendency of cynicism is particularly worsened because engineers are naturally repulsed by the idea of reinventing the wheel, particularly a wheel as giant as these "big problems I thought we solved already." It's also worsened by the fact that within any new technology cycle, there is often more noise than signal, and so it's easy to latch on to the noise as evidence that there is no signal if you're biased to believing it's all noise. It's also much easier to tell yourself that there is no new innovation happening, it's all noise, when the people building the new stuff are all younger than you and many seem to be unaware of history.
