Intelligent Idiots For a time, I thought intelligence were sufficient for creating great things; I've learned it's not. The basic pattern I noticed was ostensibly smart people doing obviously stupid things, building work which needn't be built, creating things which can never truly work, designing social conventions for working around issues which can be eliminated entirely. The average idiot would find himself unable to create the grotesque and useless software that exists; intelligent idiots are how I explain this. Being base, intelligent idiots could be explained by a dearth of aesthetic sense, and introspection. An intelligent idiot could work for decades on a tiny part of an operating system, that itself rates at millions of lines of code long, and never once question why this must be so, and only ever making minor and inconsequential improvements to the process, rather than one large but fundamental change. An intelligent idiot can make a program faster, but he can usually never find a way to eliminate it. There was a recent bout of drama regarding hash tables. There was some manner of nonsense regarding the ability to cause the worst-case performance to repeatedly strike, and there were changes made to stop this. It requires an intelligent person to analyze such a problem and create such a change; it takes an idiot to consider the worst-case performance of a data structure to be a ``vulnerability'', in need of ``mitigation''. One not inflicted with such idiocy would use a different data structure, but this requires knowing of more data structures than the array and hash table, which idiots don't. One subject the intelligent idiot is particularly good at considering is excuses for his mediocrity. To this end, there are a great many ``laws'' created, justified with insufficient or false evidence, to explain away anything the intelligent idiot is smart enough to notice, and yet too stupid to fix. These are the programmers who create ``mitigations'' that only serve as pathetic obstacles; who moan in pain at the idea of getting the program and its interface design correct the first time; who find any reason to learn something new, so long as it be exactly the same as something they already know; and who don't know for why the machines exist, which is to correctly automate work for human beings.