Stack Overflow It has always been one of the main consultation and discussion pages for developers of all levels. But its popularityhuge in the sector, began to fall rapidly from the appearance of chatgpt In November 2022. Far from stopping for a while and the developers of the new situation, their fall has followed at high speed, and has no signs of stopping in 2025.
Comments on the reduction of the post number on your website go from mouth to mouth in the sector. In addition to being published less, it has also fallen. The reasons for this fall, whose beginning coincides with the popularization of chatbots and large language models, seems clear: LLMS are faster and even more efficient helping developers out of a jam with their code when they do not know how Advance or solve a problem.
Apparently, until the generative AI arrived, Stack Overflow was the preferred option by developers to save their obstacles with virtually any programming language. Above Google. They used to go to her to find a solution to their problems, or in case you had not published one that came well, to ask the community a question so that any of its members answered and solved the problem that was addressed in it .
Chatgpt, apparently, changed this trend. But the reality is that Stack Overflow’s decline had already begun when he arrived. In full pandemic, in April 2020, Stack Overflow had a remarkable peak of use. But since then he began registering a decrease, slow but firm, in the number of questions published on the web. Chatgpt was only in charge of accelerating its decline.
The beginning of the end of Stackoverflow, long before chatgpt
The problems of the site began some years ago. Since 2018those responsible for Stack Overflow have had to face everything type of criticism for disagreements with their moderation policies. On the one hand, Stack Overflow depends on moderators to eliminate duplicate questions, close other publications that have nothing to do with the web issues and take care that conversations between users are carried out in a civilized way.
On the other hand, newcomers to the site found more and more problems to use it due to the approval of a new code of conduct that sought to promote a more pleasant and less toxic atmosphere. What happened is that it ended up irritating users, and the most veteran moderators did not take long to start leaving the ship.
The most affected were the new users of the page, who had more and more problems to publish questions and that a moderator did not close them soon. Therefore, less questions began to ask Stack Overflow. Far from solving, the problems with the moderation of new questions soon began to extend among the veterans, who in certain cases tired and began to go to the web less frequently.
In addition, the users of the page began to feel that it did not evolve. For example, he never started using answers or video publications, at a time when applications and services dedicated to videos began to take off. And like it or not, the younger generations prefer to see a video explanation rather than read a long text that offers the answer to their doubts. To the point where currently, most online questions are answered on video, of different length depending on the complexity and detail of the answer.
Stack Overflow managers and administrators should have noticed this trend and have evolved towards it to have a hybrid site between video and text explanations. Thus, when the great language models arrived, their relevance began to decline. Chatbots have only accelerated the process.
But this not only has to do with the video, because communities also began to migrate to other services. In the case of developers, to programming groups in Discord or Telegram. That is, to places to ask questions less, and can be done more informally.
A year ago, the fall of users still did not seem to arouse much suspicion in the company behind Stack Overflow. At least, he did not seem to the outside, and blamed his fall to SEO. But a year has since spent, and the reality is that they need to do something for developers to ask again as before in Stackoverflow.
A vicious circle: less questions and less options for the IAS to learn
The data is so stubborn that the excuse of Stack Overflow owners is not very credible, because At present, most of its traffic comes from search engines. That is why it is not likely that the descent on their visits are due to SEO. At least, not immediately.
But the website is already entering a vicious circle: if less questions are asked and answered, the content of the site loses news and relevance. Because the lower number of questions, less quality answers. Therefore, the site receives less SEO traffic, and web visitors find more outdated responses.
The Stackoverflow fall is an example of how the generative AI can contribute to the decline of a business, even stable. Stack Overflow is owned since 2021 of the Risk Capital Investor Prosus, which paid 1.8 billion on the web. Even with a moderate traffic drop, in its day it was among the most confidence websites for software engineers. But new data indicates that it is suffering an irreversible fall, and it is not known whether it will maintain its relevance in the future. About a year ago, they had to fire 28% of their workforce, claiming that everything was due to the effects of AI.
On the other hand, since it is more than likely that Stack Overflow’s questions and answers have been used to train LLMS, no matter how good they are now answering developer questions, where will the LLMS get out of the future the updated information to train in development problems?
Why The fewer questions ask and answer at Stack Overflow, Less sites and forums companies of all sizes will have To train your IAS for development support. And as languages evolve, developers will not only be enough to ask questions in a pleasant interface if they stop finding what they are looking for for lack of data so that the LLMS behind the chatbots are trained.
For now there is no answer to this question. Nor is it known if Stack Overflow will evolve and attract developers again, or will continue to walk little by little towards irrelevance. For now, they have made the generative companies another source of income for them, since both Google Cloud and Openai, among others, have been associated with Stack Overflow to use their knowledge base to train their models.
