Why we love ChatGPT: No question is stupid
Being able to ask questions without fear of judgement is powerful. So what happens to answer sites that used human generated content?
An underlying element of the love for ChatGPT/LLMs is that no question is too stupid.
Ask whatever you want, rephrase it multiple times. Ask the most basic things, that everyone already asumes that you should know.
Stackoverflow is a cherished website for developers looking for help with code, with multiple group sites on non-code topics. But these sites were a pain to use. In an effort to keep the site tidy, new questions would constantly be closed as 'duplicate', 'off topic', 'not formated correctly' etc. You might get a great answer to your question, or you might get ridiculed. It was a steep learning curve for new users.
Over the last 1.5 years, Stackoverflow(SO) has lost half its traffic, and the number of new posts are down a similar amount. Many of the 'easy level' questions that developers would go to SO for, they can get an answer easily in ChatGPT or anotehr LLM.
Was there a space for SO to be an early adopter of LLMs? Maybe autogenerating answers and then incentivising community members to vet them?