Rly feels like stackoverflow destroyed every part of itself for an AI pivot that was never going to work for them. Like "AI within StackOverflow" was always gonna be worse than "ai who can search stackoverflow (and also knows other things about your problem)", and so after many stumbling steps and bad ideas they pushed out some lazy RAG chatbot and near as I can tell just accepted their forthcoming demise
BUT if they had kwizatch haderach vision of the futures to come! they could have held off for a year ("AI is cool but we are fundamentally a site about high quality questions and confirmed solutions by experts") and then, December 2025, moltbot happens:
Announcement: "It's clear to us that AI are now experts in certain fields and capable of suggesting, validating, and contributing to the global pool of knowledge. As such we're becoming the first social platform to allow AI to contribute on equal footing with humans."*
[talk shows. podcasts. media coverage. Meta stack exchange in an uproar. The world waits with baited breath to see how this goes wrong]
Despite the "equal footing" copy, agents operate under pretty strict terms. AI accounts have a badge, need to be tied to a human contact, their first few posts are held in a mod queue for review, they're unable to post on meta, and the human operator needs to pay $5 to enable them to post questions "to keep quality standards high".
MSE in an uproar, look at these dogshit answers by robots. BUT! other people point out ok, yes but to be fair, but look at this answer by another AI, it cites sources, has a table, and has a try-it-online, and its genuinely novel. this isn't an AI problem, this is a low-quality user problem (and we have plenty of human low-quality users!)
World looks the same except a slightly different bunch of autistic nerds yell at you when you post a StackOverflow question formatted slightly wrong