hat’s only one (valid) side of the story. I’d argue that experienced open-source developers have limited time to invest in community projects. “Vibe coding” — when done thoughtfully and backed by real expertise — can actually raise the bar: it can accelerate contributions and lead to better product quality and UX.
Unfortunately the vast majority of "developers" are using AI to create AI slop issues and pull requests without thought or consideration. We can smell the copy-paste nature of these, and I've taken to suspending offenders on GitHub for 24 hours to give them time to reconsider their life choices and read and follow your contributing guidelines.
I don't like people walking into our community, dropping their pants or lifting their skirts to drop a huge 
in the room, then leave and never return. Volto is especially getting slammed by such yahoos.
In Discord, we're now blocking posts with keywords and phrases that smell of copy-paste "hire me" AI slop without any concern whatsoever whether they're relevant to Plone. It's become a serious problem.
Interestingly enough, even Anthropic is saying that using AI for coding is not that good.
They ran a study, detailed here: https://www.anthropic.com/research/AI-assistance-coding-skills
and they say: “We found that using AI assistance led to a statistically significant decrease in mastery".
The best funniest part of the study is when they tell people to do the test without AI and people will just refuse to do it 
Continued Non-Compliance: Even when warned about the strict no AI requirements, participants continued to use AI for both coding and the quiz.
Very interesting read!! Thanks for sharing!
AI in general can be very helpful, IMO. Not talking about vibe coding (I’m actually vibe coding my new neovim config currently in hope to safe time but the results are mediocre). This comment is more on AI helping to produce better code in general. I had quite some good discussions on some details which a Google or Stack Overflow search couldn’t answer quickly or I had to read the w3c specs in detail to get the answer. For example, questions on accessibility best practices are a recurring topic in my AI conversations. Another time an AI discussion changed my view on a specific JavaScript event handling architecture question, which I didn’t think of before. I think used with care it can help one to get better.
But I see that not only vibe coding but also AI discussions draws traffic from Stack Overflow and Wikipedia.
For me it's been hit or miss. Autocompletion suggestions can be surprisingly good! More recently some things I couldn't remember all the details for adding something in Plone were very nicely brought up by VS Code's agent (though not all were correct!). It's a shame that whatever is generated doesn't come with a confidence level indicator...(not that it would necessarily be accurate!)
AI conversations helped me improve my coding style and use of repeatable patterns. Also use it to strenghten the logical structures for my packages, module names, function signatures etc. Anything more low level will often come up with “close but no cigar” suggestions, or reinventing things that are already in Plone. Or complete nonsense with sharp edges, like the example below from QWEN3 
… do not try at home.
In zopectl debug (with -O mdb):
from zope.component import getSiteManager
from zope.component.persistentregistry import PersistentAdapterRegistry, PersistentUtilityRegistry
sm = getSiteManager()
# Reset Adapter Registry
if isinstance(sm.adapters, PersistentAdapterRegistry):
print("đź’Ł NUKING ADAPTER REGISTRY...")
sm.adapters.__init__()
# Reset Utility Registry
if isinstance(sm.utilities, PersistentUtilityRegistry):
print("🧨 NUKING UTILITY REGISTRY...")
sm.utilities.__init__()
import transaction
transaction.commit()
print("âś… Done. Restart Plone.")
Interesting to hear the perspectives. I can see how AI can undermine the need for certain types of Open Source software. I'm definitely watching this space.