Clojurians Log v2
ai-skeptics 2026-09-07
Did increased AI use result in a much lower activity in programming Q&A communities such as Clojurians Slack? My hypothesis is YES, but I wonder if this is something that is being discussed and what the potential ramifications are down the line. We already know that there is a risk of "inbreeding" with the amount of AI-generated content that is available online and being sweeped up as training material, but lately I am wondering more about this perceived decrease in human activity. It may also just be a local Clojure thing, not sure.
It certainly makes much more sense to ask AI about language/ecosystem specifics, beginner or not. I'd imagine that to be a significant contributor.
> My hypothesis is YES The data is only available for the past year, and I wouldn't call it conclusive by any means, especially when we have nothing to compare it against (e.g. large dips could be stable across years due to specific seasons). > if this is something that is being discussed I haven't seen such discussions. > what the potential ramifications are down the line If the hypothesis is correct, there will be a drop in active users. But there won't be a mass exodus and a dead platform.
I’d say that it has significantly reduced my own activity, at least. I can ask an LLM ELI5 some library usage for me, for example, rather than going to the relevant channel and asking a question. It hasn’t eliminated it, but reduced it for sure.
https://meta.stackoverflow.com/questions/437921/how-does-the-continued-decline-in-posts-since-may-25-influence-our-interpretati StackOverflow has been undergoing a huge decline, and, all else being equal I would assume somewhat similar dynamics are happening here. Compared to waiting for a reply in a forum, the LLM offers basically instantaneous feedback, so it's not hard to see why people would turn to it for daily questions.
Without having data, I always wonder what happens to ‘water cooler adjustments’ in this llm world. Like a coworker asking about some sort of a change and together realizing that the wrong problem is solved
Probably, but I also think Covid had an effect.
I get it, but it also makes me a bit sad; there are so many people who really love to teach others.
StackOverflow was declining before LLM, but for sure the friction of dealing with obnoxious moderators was no match against LLMs.
> Without having data, I always wonder what happens to ‘water cooler adjustments’ in this llm world. Like a coworker asking about some sort of a change and together realizing that the wrong problem is solved we simply have much higher level discussions, and more time for watercooler discussions in-between prompts 😄
Hi @jean.boudet11 Thanks for the link. Would you provide some editorialization on how this video is important to you?
I feel the same actually about the mutation of the job of dev. Where is the fun if all the good thing can be done by the AI?
Here is a similar view, but with a different ending, not sure if that helps: https://notes.erlend.sh/3muwdtclths2d