Is AI changing engagement across the CKAN ecosystem?

I can’t speak to CKAN-specific numbers, but the broader trend is well documented. Stack Overflow’s own data shows monthly questions have collapsed to levels last seen at its launch in 2009, and the decline accelerated sharply right after ChatGPT came out in late 2022. Gergely Orosz covered it well here: Stack overflow is almost dead - The Pragmatic Engineer

Interestingly, one academic study found the questions that do still get asked publicly are shifting toward harder, more complex problems. The easy ones get absorbed by AI, and only the genuinely gnarly stuff surfaces in public. Maybe that’s the pattern forums like this one should expect: fewer questions, but denser ones.

Honestly, yes, quite a lot. When I hit a brick wall, AI gets me unstuck most of the time. But when it’s a niche issue I’ve already learned the hard way (from experience, or from seasoned devs), I usually know the fix myself. Sometimes I still ask AI just to see what it says, and I’ve watched it spiral into token-burning loops of confident solutions that fix one thing and break ten others.

It’s a lot like Google Maps for me. In unfamiliar territory, I’d be lost without it. In my own neighborhood, I tell it to be quiet :slight_smile:

I think we have to accept a shift in the default: many users now expect to be served answers rather than go find them, exactly the pattern @jqnatividad described. If public Q&A is drying up as a discovery channel, the thing AI can’t replicate is the human side of a community: the weekly open calls, real conversations, people knowing each other. Might sound vague, but keeping the community stubbornly human may be the actual moat :thinking::thinking:

3 Likes