对于全世界码农而言，Stack Overflow（SO）可以说是最有用的网站之一。基本上所有开发相关的问题都能在上面找到答案。如果没有答案的话，你提出的问题大概率会有人给出答案。
图源：Pixabay
但2026年7月的一个报告直接让很多人破防了：整个一个月的新提问数量定格在1304个问题上，而在2014年的时候，单月提问量则是20.7万个。那个曾经被全世界程序员尊为“救命圣经”的黄金资料库，正以一种不可逆的姿态走向落幕。
前两天，Daniel Lockyer 在 X 上晒出一张陡峭下坠的折线图，叹息这是一个“时代的终结”。随后另一位开发者 Pieter Levels 转发了这张图，同时抛出一个更令人窒息的问题：当没人再公开提问，互联网上还剩多少真人写的内容？没有新数据，下一代 AI 模型靠什么训练？
有的人认为SO是被近两年的AI大模型精准打击，一枪爆头的。这个分析可能是有道理的，但似乎并不全面。
近两年AI的兴起，确实起到了一个“截胡”的效果。
当程序员遇到问题的时候，如果在浏览器中搜索，首先会被搜索引擎底层的AI获取问题，并生成相应的答案，完美截断早期到SO的流量。
甚至都没有到这一步。很多程序员已经习惯了类似VS Code这样编码工具中的AI助手。有啥问题直接问AI助手就能得到答案，而且体验更好，不用自己筛选，直接拿到最终结果，是不是更香？
对于使用类似Claude/GH Copilot CLI的程序员而言，似乎更近一步，你连问题都不用问了，直接把想要做的东西告诉AI，然后AI直接帮你把代码写好了。至于SO，对不起，新生代的程序员一直和AI协同工作，似乎不知道，也许早就忘记SO是个什么东东了。。。
但如果把SO的衰落全部归结到AI上面，似乎也不全面。SO本身也一直在做。可以说是No Zuo No Die的经典案例。
被吐槽最多的就是其拒人千里之外的“精英式傲慢”。如果你曾经用过SO的话，肯定遇到过类似场景。你的提问被秒关，打负分等等。整个社区似乎对新人非常不友好。长久以往，新用户大批量流失就成了不可避免的趋势。
还有一个不可避免的趋势就是，既然大家都不在这里提问了，那么原来回答问题的专家/资深人士们也就失去了来这个社区的动力。以至于造成了一种趋势，当AI给出的答案是错误的时候，人们似乎已经失去了一个获取答案的途径。
其实不仅仅限于类似SO这样的程序员社区，前段时间我也提到过AI对靠SEO生存网站的降维打击。还包括维基百科的真实人类访问量持续下行。。。
哎，真的是所有内容网站的末日黄昏了。。。
I find this thought really interesting, although I wouldn’t attribute the decline of Stack Overflow to AI alone.
I used to use Stack Overflow a lot myself. In the development department I worked in back then, it was quite normal for us to discuss problems there and work together on finding solutions. That was one of its greatest strengths for me: you didn’t just get one answer, but often different approaches, experiences, and insights from other developers.
At the same time, I also noticed the other side. Especially for newcomers, the way people interacted with them could sometimes be rather unfriendly. A question would quickly be closed, someone would give it a downvote, and only afterwards might you get an explanation of what was actually wrong with it. For someone just starting out, that isn’t exactly an inviting environment.
And then AI arrived.
Today, I no longer have to find the perfect search query, work my way through ten different answers, and figure out which one actually applies to my problem. I can simply describe what I’m trying to do and get an explanation, a possible solution, and often even the appropriate code.
So AI has certainly taken over a large part of what made Stack Overflow so valuable in the first place. But perhaps it has also exposed a weakness that was already there.
What I find particularly interesting is the thought you raise yourself at the end of the article: What happens to the internet if, eventually, hardly anyone documents their experiences publicly anymore because most people simply ask their AI?
Because AI can explain existing knowledge wonderfully and combine it in new ways. But someone still has to create that knowledge and write it down in the first place.
Perhaps that is the real challenge of the coming years: not access to knowledge, but maintaining the incentive to continue sharing new knowledge with others.
Edit: I actually only meant to write a short comment about how much I like the picture in your post. Somehow, it turned into a small reflection on Stack Overflow, AI, and the future of shared knowledge. 😉
More information - Support - Reports - Discord
Support engagement with a delegation to topcomment:
50HP - 100HP - 250HP - 500HP - 750HP - 1000HP - 1500HP
You hit the nail on the head. The ironic part is that Stack Overflow's harsh environment made AI look so much better, but now we're running out of the very human knowledge AI needs to learn from.
If developers stop posting their real-world fixes publicly, future AI models will just keep recycling old data from a few years ago. The real issue down the road isn't just getting quick answers - it's losing the places where new solutions are actually built and shared in the open.
Thanks for sharing your thoughts
AI的出现确实颠覆了部分行业。。
确实这样