So Stack Overflow cut about 28% of its staff last week, somewhere around 100 people, and I keep coming back to it even though the news cycle has already moved on to other stuff. Prashanth Chandrasekar (still their CEO, in case you lost track like I had) framed it as the usual "macro headwinds" language, plus a pivot toward selling their data and tooling to enterprises through this OverflowAI push they announced back in the summer. Translation: fewer humans answering questions on the free site, more effort going toward licensing that mountain of Q&A data to companies building coding assistants.
And yeah, I laughed a little at that, because the thing eating Stack Overflow's lunch is exactly that kind of tool. I opened my browser history from the last two weeks looking for stackoverflow.com hits and there are maybe four. Four! Three years ago that number would've been in the dozens, easy. Not because I got smarter. Because Copilot just answers the question inline now, or I paste an error into ChatGPT and get something usable in ten seconds instead of scrolling past six "this is a duplicate, closing" comments to find the one actually helpful reply from 2014.
I want to be clear that I don't think this is entirely good. There's a real cost to it that doesn't show up right away. Stack Overflow answers exist because someone hit a weird bug, fought with it for an hour, and then typed up the fix for strangers, and that whole corpus is what a lot of these models were trained on in the first place. If the site withers because traffic moves to tools trained on its own archive, eventually there's less new material to train the next generation of tools on. Nobody's incentivized to write the 2027 equivalent of that 2014 answer if nobody's reading it. I don't have a tidy fix for that, I just think it's worth sitting with for a second instead of doing the usual "well technology marches on" shrug.
Also, small complaint that has nothing to do with the layoffs really, but every time I see Stack Overflow in the news I remember how brutal that site's moderation culture could be to actual beginners, and part of me wonders if some of this decline was self-inflicted years before ChatGPT ever showed up. I've watched people post a perfectly reasonable question and get it closed as "not constructive" within four minutes flat. A forum that treats new users like that shouldn't be shocked when those same users find a tool that never rolls its eyes at them.
None of this is a prediction that Stack Overflow disappears. Reddit's programming subs are still full of screenshots pulled from SO answers, and honestly for anything niche or old (weird Delphi behavior, ancient Postgres version quirks, whatever), it's still the best place on the internet, better than any chatbot, because a real person hit that exact wall in 2011 and left a trail. I just think the balance of where people go first has permanently shifted, and this round of layoffs is the company admitting that out loud with a press release instead of just letting us all infer it from the traffic graphs.
Separately, and only tangentially related: Meta's Ray-Ban smart glasses started actually shipping this week, the $299 ones with the camera and the speakers built into the frame. I haven't touched a pair yet. I remember the original Ray-Ban Stories from 2021 being a punchline within about a month of launch, so I'm withholding judgment until I see whether these get returned in bulk the same way or if the battery life and camera quality are actually different enough to matter this time. Early impressions I've read seem more positive than the first round got, for what that's worth, but "more positive than a product everyone made fun of" is a low bar.
Anyway. Going to go dig through some old bookmarked Stack Overflow threads while they're still there, mostly out of nostalgia, honestly. Some of those old answers have comment threads that are basically a museum of how confused everyone was about async JavaScript in like 2015, and there's something oddly comforting about that.