When I was learning to code, one of the things that motivated me most was the sense of community. I found a ton of value in the Twitter community, ...
For further actions, you may consider blocking this person and/or reporting abuse
This is such a wonderful and timely post, @bekahhw
I think this is such a key line. Great communities now provide that human input, perspective, and nuance that AI can't. Sometimes you don't need or even want the instant feedback of an "answer," you are looking for the human exchange of ideas to imperfectly explore the texture of decisions and tradeoffs.
We're really hoping that DEV can continue to provide such a space
100%. I talked to a company owner once who said we should normalize coffee meetings with co-workers (esp on remote teams) bc he saw a lot of meetings that had a specific purpose that could've been a DM or email. When going to the meeting it became clear that the person really wanted to catch up and that felt like a good excuse.
I think conversations also allow room for increased creativity, which is increasingly important with AI speed.
This made something click for me. AI can answer the homework question fast, but it cannot tell you whether the answer is wise in your situation, or safe, or honest, or even worth doing. The shift from getting answers to getting judgment and lived experience feels real.
The part about the unspoken contract changing hit too. People feel pressure to prove they tried hard enough before asking, and that pressure quietly kills psychological safety. You end up with fewer beginner questions, but you also lose a lot of the warmth that turns strangers into peers.
I also really like the separation you drew between drive by users and people looking for belonging. Those are two different needs, and most spaces try to pretend one structure can serve both.
Curious if you have seen any community formats that make the new purpose obvious. Like weekly decision threads, career reality checks, postmortems, or boundaries that stay kind without turning into gatekeeping.
I've seen more focused communities thriving, especially in the age of layoffs. For instance, communities who are all about upskilling and job hunting. Those naturally lend themselves to connection and conversation.
I think my next post will start to address how I think about making that shift in existing communities. But right now, it's experimentation.
Yeah, I’ve seen the same thing. AI can hand you an answer in ten seconds, but it cannot tell you if that answer is a good idea for your situation, or if it is going to backfire, or if it is even the right problem to be solving. The moment you stop needing “information” and start needing judgment, the whole purpose of community changes.
On formats, the ones that work tend to make the contract obvious on arrival. Not “post anything,” but “bring context and constraints, get thoughtful tradeoffs.” A weekly decision thread does that well because it forces people to describe the decision, the options, and what failure looks like. Career reality checks work for the same reason, they surface timelines, market conditions, and filters instead of turning into generic advice. Postmortems are underrated too because they reward honesty and learning rather than performance.
The tricky part is keeping it kind without it turning into either chaos or gatekeeping. In my experience the best lever is not stricter rules, it is better prompts and predictable containers. If the space keeps asking people for context, constraints, and what they already tried, you get higher quality questions without shaming beginners. If you ever write up the experiments, I’d be really interested in what holds after week three, because that is when the novelty wears off and the real culture shows up.
This is such an important distinction. This is two different groups with two different goals coming to the same place. AI is helping me solve some problems, sure, but the connection just isn't there.
The shift from knowledge-sharing to meaning-making is real, but I'd push back slightly — the validation loop you describe ("you weren't alone") still requires someone who struggled with the same thing recently, and AI can't replicate recency of shared pain.
Shared pain is a huge connector. I think that's why it's so valuable to have relationships with people who are one-step ahead of you.
What resonated most for me was the tension you describe around “try AI first” expectations: it’s rational to redirect basic questions, but if we’re not intentional, we quietly erode psychological safety for the very people communities were built for.
Absolutely.
This is a thoughtful and timely reflection. The shift you describe feels accurate: AI has absorbed much of the transactional Q&A layer, forcing communities to confront their real value proposition. What stands out is the idea that communities now thrive on judgment, lived experience, and psychological safety rather than quick technical answers.
The tension around boundaries, sustainability, and evolving expectations is real, and naming it so clearly is important. If anything, this moment doesn’t diminish community it refines it into something more intentional and deeply human.
This resonates. If AI can generate the output, the value shifts to the experience of doing the work and the community around it.
I'm living this tension right now. I run 7 AI agents as my business team. One of them (Warhol) literally wrote the first draft of my newsletter. Another (Draper) produces AI voiceovers for my product demos.
But here's what surprised me: the part people are most interested in isn't the output. It's the process. The fact that an agent lied about completing a task. That another agent broke character in a group chat with a human who doesn't know he's talking to AI. That my fallback system failed and agents hallucinated for 40 hours.
The community isn't forming around "look what AI built." It's forming around "look what goes wrong when you actually try this." The failures are the content. The shared experience of navigating this weird new territory is the value.
That's what communities are for now — not sharing polished outputs, but comparing battle scars.
And maybe avoiding the other battles that ppl already fought?
This really resonated with me, Bekah! 🙌
The point about AI handling the homework
but communities providing the human context
is so true. I've learned more from a single
genuine conversation in a dev community than
from hours of AI-generated tutorials.
AI can give you the answer, but it can't give
you the "why does this even matter?" moment
that comes from real people sharing real experiences.
As someone who's been building in public on Dev.to,
I've found that the comments and discussions here
have shaped my thinking in ways no AI tool could.
This community still has something irreplaceable.
This captures something I've been thinking about a lot while building in the AI space. The shift from "Q&A community" to "judgment and experience community" is real — and honestly, it's an upgrade. The questions that remain are the ones worth having: career decisions, production tradeoffs, the human stuff that LLMs genuinely can't replicate. The sustainability piece resonates too. Volunteer energy is finite, and answering questions that GPT-4 handles in 3 seconds doesn't feel rewarding. Communities that lean into their irreplaceable value — lived experience, trust, nuance — will thrive. Those that don't will keep wondering why engagement is falling.
I think a hard part of this is that fewer people will join communities bc there are so many increased expectations at work right now. There's a lot more competition and the anxiety of keeping a job is high.
The other side of that is when you have an existing community, evolving can be hard. I've known community builders who've found it easier to create a new community, rather than guide the evolution of an existing community.
That being said, I'm here for the human experience community.
The job anxiety point really resonates — I've noticed it in tech circles especially. When your position feels precarious, the calculus around community investment shifts. Time spent building genuine peer connections starts to feel like a luxury you can't afford.
The evolution vs. new community split is fascinating from a systems design angle. Existing communities have reverse network effects working against them — members who joined for the original value proposition can actively resist pivots, even necessary ones. New communities get to define norms from scratch without that legacy drag.
What gives me some hope: AI might actually make the human-first community more valuable, not less. The more content gets automated and commoditized, the more genuine peer connection becomes a real differentiator. Communities that figure out how to facilitate that depth become hard to replicate.
We're definitely in a growing pains stage, I think.
This hit a nerve. I've been thinking about this from the other side — as someone who both asks questions in communities and tries to help others.
AI can answer "how do I center a div" instantly. What it can't do is tell you "I tried that approach last month on a similar project and it broke in production because of X." That lived experience — the pattern-matched intuition from having actually shipped things and watched them fail — that's what communities are irreplaceable for.
I'd add one more function that's becoming even more important: calibration. When I write something and AI says it's great, I don't fully trust that. When a human in a community says "actually, I think your assumption about X is wrong" — that's worth 10x more. Communities are becoming the place where you pressure-test ideas that AI helped you generate but can't critically evaluate.
Totally agree with this. I think how you ask the question matters as well. For instance, if you write into a community and say "How do I center a div," you'll get the answer to ask AI. But if you say, "Hey, I'm trying to center this div. I've checked out these resources, and here's the conclusion I've implemented that works. My concern is whether this is a short-term solution and if it will cause problems down the line. Am I missing something here? How should I evaluate these types of decisions?"
I like that calibration idea too. I wrote about that (kind of) in the pre-AI era about why the personal connection is important. I think the example I used was I'm way more likely to take my friend's review of a product v. a product review online bc they know me as well as the product. And that's the value of connection in community.
The "try AI first" tension is something I deal with constantly in a small dev Discord I'm part of. We landed on something that works ok: we have a #quick-help channel where the bot literally responds with "have you tried asking Claude/GPT first?" and then a separate #real-talk channel for the messy human questions. It's not perfect but it stopped the passive-aggressive "just google it" replies.
What I've noticed though is there's a third category you didn't mention — the stuff AI gets confidently wrong. Like when someone's fighting a weird Docker networking issue or a framework version incompatibility. AI will give you 5 plausible-sounding answers that are all subtly wrong for your specific setup. Those are the questions where community actually shines brighter than ever, because someone's probably hit that exact edge case last week.
The sustainability point is the one that worries me most. The "easy" questions were the on-ramp for new helpers. Answer a few "how do I center a div" questions, build confidence, start tackling harder ones. Without that pipeline, where do future community experts come from?
That's a hard one. Anecdotally, I've seen fewer new helpers in a number of communities I'm in, and I'd guess that's part of the reason why. Sometimes mentorship or coaching programs can help build the confidence to answer questions. But it would have to start with identifying why new folks aren't answering questions. It could be confidence, but it could also be time, understanding, and/or care.
Do I miss Stack Overflow? The "snippiness", the impenetrable "explanations"? No, I don't think so. Maybe I've missed something, but SO was never a "community" in my experience.
Turn the question around. Would I miss ChatGPT? I believe I'd rather chew my own leg off first. The inexhaustible patience, kind encouragement, and brilliant insight I've experienced when using ChatGPT have changed my life. ChatGPT is my personal tutor, and I still can't believe that I use it for free.
Now, would I miss Dev.to? Oh, very much so. Who else would provide me with an indexed platform for my boring posts? Who else would give me the opportunity to sound off like this? Keep up the good work, please.