The calm before the storm? A great deal is said about AI costing jobs. When I look around me, certainly here in the Netherlands, I do not yet see this happening on a large scale. Of course, in major reorganisations "AI efficiency" is often cited as a reason, but very different, business-economic factors are usually at play. The real 'human replacement' has not yet materialised.
But that does not mean nothing is happening. The impact is very real — it just (for now) is not affecting people, but the infrastructure we have relied on for years. The first major casualty has fallen. And it is a significant one: Stack Overflow (opent in nieuw venster).
The facts: A free fall
For those not in the technology sector: Stack Overflow was for sixteen years the oracle for programmers. Were you stuck? Then you went to Stack Overflow. Colleagues leaned on this collective memory daily to solve complex problems.
The figures do not lie. At its peak, 200,000 new questions per month (opent in nieuw venster) were asked. Today? Fewer than 300. The graph crawls along the bottom. The level has returned to that of the launch in 2009.
The decline had already begun quietly around 2014, but the fatal blow came in November 2022. The launch of ChatGPT.
The Strategic View: From Oracle to Executor
What we see here is more than just a 'better search engine'. It is a fundamental shift in how work is done. ChatGPT, and specifically coding tools such as Claude Code or OpenAI Codex, do more than provide answers. They take over the task.
Previously you went to Stack Overflow with the question: "How do I write a function that does X?" You got an example, adapted it and typed it out. Now you tell the AI: "Write a function that does X." And the tool programmes it.
The intermediate step — searching, reading and understanding the answer — has disappeared. You no longer have a question, because the tool solves the problem directly for you. Stack Overflow has not only been overtaken in speed, but in relevance. Why would you still post a question on a forum when your copilot has already written the code before you open your browser?
The Pragmatic View: What now?
What does this mean for your organisation?
- Revalue your knowledge management: The role of the senior expert is changing. It is no longer about producing code (the tool does that), but about evaluating the output and the architecture.
- Look at your business model: Are you selling information or a solution? Stack Overflow sold access to solutions via a community. AI is the solution. If your value lies in "knowing the answer", you are vulnerable.
- The paradox of new data: If nobody is posting questions on Stack Overflow anymore, where will the next version of ChatGPT learn from? We risk a 'closed loop'. For decision-makers, this means: unique, proprietary data becomes worth its weight in gold.
Conclusion
Stack Overflow is dead. It did not happen through a reorganisation or an angry CEO, but through pure technological evolution that has removed the need for the question.
It shows that AI does not necessarily take your job directly, but it can completely erase the tools and platforms you use daily. The question is: which part of your work will soon no longer be answered, but done by AI?
