The main thing is the difference between the standard deviation person over the other five. If I ask six people questions about lambdas, or concurrency, or gen

The main thing is the difference between the standard deviation person over the other five. If I ask six people questions about lambdas, or concurrency, or generics, or language questions, or framework questions, generally a person will respond the same no matter the topic. For instance if you ask what generics in Java are from people who have been programming Java for two or more years, one in six will have no idea at all what they are. Four in six might say "well Java generics...that is when a class or method uses angle brackets in its definition". OK, what is in those angle brackets? "If you're initializing a list of strings, you send String as a type argument to List - so when you're initializing a list, what is between those angle brackets is String - the type argument for the list. If you have Integer as a type argument the list will contain integers". So they have this level of knowledge of being able to utilize existing generic classes, but they don't define their own generic classes, and can't go in-depth on generics. The standard deviation above one in six can clearly define what Java generics are, knows about type erasure and that generics are a front-end conversion at the front of a Java compiler, can give a decent explanation of what an upper-bound wildcard is etc. A long time ago I was interviewing proto-SREs, and one question I asked everyone was what are the facilities of syslog, and what are the priorities of syslog, in order if possible. I must have interviewed at least two dozen people who were working for big and small companies running Unix servers for years, none of them ever came anywhere close to being able to answer that question. You would think they might have to edit the syslog configuration file to log different facilities at different levels to different places (files or otherwise), but I never got anywhere near an answer to this. These were people all making $80k-$120k TC (eroded by inflation since then) for a position with similar TC. It's a hard thing to see if you haven't interviewed dozens of people, it's why it's so enlightening, especially if I go out and interview. Generally, one in six is a complete dud, one in six is a standard deviation above the others, and four of six are muddled. It's kind of like a classroom where some kids sit in the back and fail or get bad grades, a few sit near the front and always seem to know the answer to the teacher's question, and most of the class is kind of muddling along, getting Bs and B-'s, with a few A's and C's sprinkled in. For pretty much any question in the rabbit hole of topics - one person knows very little, one can answer most things in some depth, four have a surface knowledge of many things but can't go into depth on them much.