Yeah, the newcomers asked worse questions than average. There was a huge flood of people presenting like they thought they had a difficult question about how neural network simulations work, but actually had code full of multiple elementary errors (completely unrelated to the domain, just learning the programming language kinda stuff) and no apparent willingness to read error messages even to the extent of "it says TypeError, I guess that must mean something is wrong with a type".
replyRight? It's not a very subtle spike, is it? Even then, if you erased the bump, the graph is pretty close to a straight downward-sloped line from 2016 to late 2022.
replyIt's kind of sad: I bet there was some excitement around the SO office. Is this it? Are we on the way back? Have people remembered that we're here and now they're returning? Alas, no, they weren't, and they weren't. That must've been rough.
“Has our strategy of being openly hostile to anyone asking questions finally resulted in the increase in questions we always dreamed of? Did our long term bet on alienating the community finally pay off?”
replyThat's probably why the number of new questions asked dropped off. Once someone has asked "how to fizzbuzz in python", that's pretty much it. You'll get the textbook answer, the "why are you fizzbuzzing in the first place" answers, and then an answer a couple years later using the new python fizzbuzz built in library. You have a massive compendium of questions and answers and for most people, that's good enough. I don't think that stack overflow was necessarily less popular in visitors pre-LLMs but they certainly killed it off quick.
reply[flagged]
replyThis was all too common across tech. I know of many products at the time that, say, had their usage quadruple in Covid (from a very large baseline) and nobody in leadership seemed to ask the question "is this permanent or just an artifact of people being stuck at home?" Clearly the answer was "no".
reply