I recently claimed that we might look back on the 2010s decade and early 2020 years as a golden age of the internet, collaboration and information exchange. The good old days when tools had evolved, users were still ambitious to share and collaborate online for free and knowledge didn't get lost in AI chats ending in dubious almost-good recommendations below the quality of the once-good-now-outdated StackOverflow answers and crafted tutorials that the LLM had been trained on.
Why DEV?
I discovered the DEV community when I got disappointed about Medium's lack of quality and StackOverflow's gatekeeping. I had criticised DEV's badges as rewarding quantity over quality, but DEV's challenges focus more on quality.
No good questions anymore?
Apart from developer-focused forums and communities, more general platforms include SubStack, Reddit, Quora and LinkedIn globally. In Germany, Gutefrage.net ("good question network") has acquired several contenders and its desperation for clicks and content attracted dubious propaganda and adolescent Q&A to an extent that its TrustPilot score dropped below 2 which is very sad for a platform built upon a good idea and an active community providing questions, answers and ideas for free for decades.
Is StackOverflow "almost dead"?
Image source, inspiration and further reading: Stack overflow is almost dead in The Pulse by The Pragmatic Engineer Gergely Orosz on Substack.
On a much higher level, StackOverflow has built its reputation on the free contributions of skilled experts since it obsoleted Experts Exchange nearly twenty years ago. However, even without contemporary AI and Google's zero-click summaries, it was seen past its best days with decreasing contributions and with upvoted answers getting outdated as time went by.
Blaming AI for a Self-Inflicted Decline?
It's easy to blame AI for lack of user engagement, but bad UX and wrong business decisions play just as big a role in my opinion. Toxic social media algorithms have destroyed Web 2.0 and Web3's ideas of a decentralized non-commercial Web 1 revival never lived up to its promise, at least not yet.
Although AI can help naive users publish a lot of content without learning and researching before publishing, AI can also help to detect its own slop and distinguish quality content from spam. But it needs constant human curation and high-quality authored by human beings based on their authentic experience to keep up a high level of AI output. Otherwise, that will deteriorate just like prior knowledge management strategies started out promising but deteriorated without continuous maintenance.
Curiosity and Curation: Growth without Losing Quality
DEV has grown a lot, and it is full of dubious content both by beginners who are welcome to share their stories and learn in public, and by spammy marketing authors often mixing actual value with sloppy filler content and opinionated links to their clients' services. DEV's tag filters don't really work. I used "negative subscriptions" to ignore content with the #ai hashtag for example, but I keep seeing those posts everywhere every day, and we keep discussing it. Nevertheless, DEV shows me relevant and inspiring high-quality content by developers that I follow and by those trending or getting featured in curated best-of-last-week's lists.
DEV vs. StackOverflow, Wikipedia and Blog Platforms
In conclusion, what's different here? DEV grows, while others decline. DEV has more content, more users, more engagement at the same time that users turn their back on StackOverflow. More general blog platforms might have more active users, but also more slop and they're not much fun anymore. Wikipedia, the most popular general knowledge platform, has managed to maintain high quality through gatekeeping, like StackOverflow risking to turn away potential contributors who don't easily fit their quality criteria or don't have enough time or priorities to spend unpaid time and effort contributing to a knowledge base which, in Wikipedia's case, is entirely nonprofit and succeeded in staying relevant and up-to-date enough for decades while other platforms rose and fell.
A Medium without a Message?
Medium has hidden most of its content behind paywalls, even content contributet for free by volunteering hobby authors, and it has encouraged writers to start ever post with a huge poster image usually taken from Unsplash or other free stock image libraries. DEV's integrated Google's Nano Banana AI image generator has created dubious artwork mostly following a neon retrofuturistic aesthetic that at least sets its post apart from other blogging platforms at first sight. I still use Medium to publish German versions of my Substack posts.
A Stack without Overflow
Substack is praised as an independent alternative, but I totally dislike its pushy user experience, constantly asking me to support, share, upgrade, recommend or post, contribute to a twitter-style comment timeline and accept cookies on every single one of their distinct subdomains. I don't get many views or followers there, but I continue to use it for its alleged search engine marketing value.
IndieWeb Posse: The Blogosphere Web
I also consistently publish on my own weblog again, which is even more disappointing regarding regular visitors, but then again, I don't need quantity when I don't sell ads. Open Mind Culture is an independent, personal, ad-free blog that does get feedback and domain authorit without striving to attract the masses.
Conclusion: Last but not Least
Last, but not least, DEV. I get tired of reading and writing on DEV from time to time, but I keep coming back. I was critical to DEV's badges and voting system. I was skeptical about their Google AI collaboration. I got annoyed by spammy listicles. I don't like all of the posts that I wrote in the past, and I did edit some and deleted others, but I kept most as legitimate historical documents that might feel helpful or entertaining in the eyes of future readers.
Top comments (46)
I think you're pretty spot on with this evaluation.
If DEV has had a decline in community quality and content it is not now.
We actually did suffer a decline, and you know what: I believe almost every reason for that was self-inflicted.
Our current trajectory post-decline has been really great, and to start 2026 we are actually seeing a huge improvement in overall quality.
Slop, and lots of other issues, still need to be addressed, but the current reinforcing cycle we're on is very much a net positive and we hope/expect to reach all-time highs in objective and subjective quality in the next few years.
Great to hear, Ben! I’m glad to play even a small part in those efforts as a Trusted Member. And to add another point: the community can nurture itself - it just has to choose to.
Yup! Thank you very much Richard.
I can only speak from my own experience, but it’s a long one.
I’m genuinely passionate about knowledge sharing — and not just for a few months. I’ve been blogging for over 15 years. I have my own blog (now running on Astro, after years on WordPress), I’ve tried Medium, and yet I’ve never felt as much at home anywhere as I do on DEV.
Sure, there’s some low-quality or “slop” content. But that’s not specific to DEV. It’s mostly tied to individuals who act without scruples — chasing a few dollars, pushing shady services, sometimes outright scams. They’re on DEV, but they’re also on every other platform with a good reputation and a welcoming community. In reality, they’re everywhere, down to the smallest corners of the internet. I even deal with them on a Google Cloud Run project where the only thing they can actually do is cost me money.
And alongside that, there are truly fantastic authors on DEV. People who genuinely contribute to the developer community, who bring real value — to the community at large and to me personally. Many of them aren’t elsewhere anymore, or not in the same way.
I’ve also noticed that comment spam is filtered quite effectively. I read your article, and independently observed the same thing myself. And then there are the content moderators — which I’m part of. If we all keep playing our role and acting responsibly, I’m convinced DEV can only keep getting better.
Finally, a sincere thank you to Ben, Jess, and everyone involved in building and maintaining this inspiring platform. What you’ve created really matters — and it shows.
It must be very difficult to balance the community and content with the very real practicality of running a business, employing staff and the like - I know you've had difficult choices along the way there. This is still the place I read most tech content, and I think the corner is turned. The feed algorithm seems much improved, too. I remain very glad Dev exists.
I think the integration of AI images really hasn't helped. I mean, I know that's what a lot of people were going to do anyway, but now it just looks like every post is slop, even if it isn't.
Strangely enough, Ben, this is exactly why I don’t generate those images for my Learning with freeCodeCamp series. Those posts are as much about keeping my own learning consistent as they are about sharing it. For #discuss-style posts, I have generated images, but I’ve been very careful and thorough with the prompts I use.
Just passing by after a while. I stopped posting when you guys decided that mixing politics with a technical platform was a good idea, is it safe to assume these are totally or partially the self-inflicted bits?
More than two years have passed since it peaked in a way that reading the feed was very uncomfortable and unhinged. How's the thing going now? Which efforts and measures have you guys taken to address this?
Dev.to forever! and Wikipedia - and then Reddit and Quora - and still SO as well ...
Medium, no not really - ever after they introduced the paywall, and most articles gained ridiculous sensationalist click-bait headers, aiming to entice readers to sign up (and pay) based on some sort of "FOMO" ;-)
Fantastic post, Ingo - a perfect complement to my piece last week on the internet’s contradictory nature.
I spent a brief stint on Hashnode a few years back, when Web3 was the tech bubble of the moment, which probably dates it nicely. I became disillusioned fairly quickly, if I’m honest.
I do like DEV’s badge system to a point - my neurodivergent brain is perhaps a little too susceptible to them. Sitting on the fence, I’d say the Discuss and Top 7 badges work well, but the Warm Welcome badge may edge a little too far into gamification.
The voting system is, of course, open to abuse - any such system would be. But as a Trusted Member, I can see the detritus beneath the surface, attempting to drown out the genuinely good articles on the platform. It’s far from perfect, but the community needed something in place.
It’s really encouraging to see posts like this that shine a light on the care and enthusiasm the community has for the platform. Moments like these remind me why being part of this space is so rewarding - it’s not just about the content, but about the people who make it meaningful.
Additionally, using Medium has never and will never cross my mind!
I think the main difference between Stackoverflow and DEV is the format. Stackoverflow is a public FAQ while DEV is a forum.
A lot of people, moved to AI for answers, and that is why the public FAQ format is dying.
For me the forum format has less of a threshold than repository issues or a chat server to communicate with other people.
I'm bad at promoting myself, so a forum that already attracted people is a better fit for my ramblings and ravings than a blog.
It isn't a DEV only problem, and i wouldn't know how to get it right, but AI authors seem to be the biggest problem I encounter.
I don't care if my English is a little bit broken, but other people are not comfortable enough to write without the help of AI correction.
So in a lot of cases is very difficult to detect posts and comments written by people. My method is feeling if behind words there is a person or not, but that is not scalable or foolproof.
The reason I think it is a problem is because the content that is produced by AI authors is not going to fixed when they give wrong information.
As people we can have different opinions, but most of the time we correct wrong information because we want to spread good information.
Maybe the one thing DEV could take away from Moltbook is to have the option to identify as a person or an agent?
I have the same inclination when it comes to older posts, keep them as a snapshot in time. Sometimes that violates the good information spreading I mentioned before. But it shows a learning curve, and when following the learning curve I think it can show perspectives that are though of and dismissed with context.
I don't mind people laughing about old posts, most of the times they make me cringe. But it is the price I pay for progress.
I completely agree - forums like DEV feel more approachable than FAQs or blogs, especially for sharing thoughts without the pressure of self-promotion. The AI authors point is interesting, as it can make distinguishing genuine human perspective tricky. I always make sure to use my “Written by a Human” badge at the bottom of my posts!
Interesting point, as AI can be just another useful or even necessary assistive techhnology. Using speech input to take notes while walking around with my old smartphone changes my writing style, and so do unsolicited AI answers when using search engines. Automated language services also may get us lost in translation or mistaken for being a bot while we're not.
Trying to distinguish humans from machines by heuristic signals might or might not work, but it might miss the point altogether. I know that some people are real and human because I have met them in real life, but that does not necessarily make their posts more valuable that those by sources unknown to me.
Indeed, Ingo - I think some people who quickly point to "AI-generated" mistakes don’t always consider that English might not be the author’s native language. That can easily lead to misunderstandings.
On a slightly related note, I once had retail training with Dragon NaturallySpeaking. The representative mentioned that one of the hardest accents for the speech-to-text software to recognise was the broad Geordie accent (from the Newcastle/Sunderland area in the UK)!
Haha, that reminds me of the short film of two Scottish comedians trying to use a speech-controlled elevator in Britain.
Indeed, exactly that!
Another reason I think a distinction is important are the comments. While AI has some reasoning skills, it fails at deep reasoning.
Exactly this, David. Very well put.
I’m currently drafting a post about the misreporting around the Moltbook AI Social Network. Even mainstream coverage has been misleading, with so-called AI experts repeating claims that don’t stand up to closer inspection.
When the post goes live, I’ll go into this in more detail, but the key point is this: if humans can freely sign up and operate accounts, then describing it as a social network for AI is misleading. What Moltbook offers is closer to a modernised form of bot scripting - conceptually not far removed from the IRC bots many of us were building back in the 90s - rather than a fundamentally new kind of AI social platform.
Communities fade when signal drops and ego wins over curiosity. DEV still feels different because thoughtful posts and practical feedback get real attention here. Keeping that culture takes active effort from contributors, not just the platform.
Exactly this, Gabriel, the community nutures the community.
Thanks for the article.
DEV community rocks, because it feels like a safe place to share our thoughts, learnings and see meaningful responses, no boasting, no flashy stuff. And even if we do not see a response to our articles or thoughts, it is acceptable (at least to me) because we got a chance to share :-)
Really healthy way to look at potential interactions here on DEV, Shitij.
With that in mind, what to thank you for all the times you've responsed with kind comments surrounding any of my content and that of others - truly appreciated by all concerned!
I like DEV, I think it is currently the best out there. I terminated Medium for constantly showing me obscenely wrong "tips" and AI slop. DEV is better in that (although not perfect).
I am just not really conviced with the "followers" system. It currenly shows me I have over 3400. Most of them I got days after joining (hundreds daily) and I presume they are just some bots tracking "say hello" thread. It is not a big deal, but it is funny to have such a number and then I write new article and it barely gets 50 views in month. I would accept I am not so good in writing, but then I see clearly AI generated "comparsion" article getting a lot of attention. It is a bit frustrating after putting a lot of effort into my own work...
I don’t think DEV is dying — I think it’s going through an era transition, and those always look messy from the inside.
From my standpoint, DEV has actually been one of the most valuable developer outlets I’ve had access to — and that context matters. Where I’m from, there aren’t many (sometimes any) formal spaces where developers can show work, exchange ideas, or even feel like part of a wider technical culture. No meetups. No strong local forums. No consistent pipelines. If you have skills, you mostly carry them quietly.
DEV filled that gap.
Not just as a place to post, but as a structured environment:
challenges that push you to finish things
badges that act less like vanity metrics and more like momentum
competitions that force you to sharpen how you think, not just what you build
For me, those mechanics don’t reward noise — they reward showing up consistently. And that matters a lot when you don’t have an offline ecosystem reinforcing your growth.
I also don’t see DEV’s growth vs StackOverflow’s decline as “AI killed communities.” That feels like the lazy explanation. What actually shifts platforms is usually a mix of friction, identity, and agency.
If you look at past era changes — MySpace → Facebook, forums → Reddit, blogs → social feeds — the pattern is consistent:
people migrate where participation feels easier
where personality is allowed inside some structure
where newcomers don’t feel punished for being early in their learning curve
StackOverflow optimized for correctness and efficiency. DEV optimized for expression + learning in public. Wikipedia survives because it chose gatekeeping with a nonprofit mission. Medium stumbled because it blurred authorship, incentives, and ownership.
DEV sits somewhere different: it’s not perfect, but it gives developers room to be human while still operating inside reasonable guidelines. That balance is rare.
Yes, there’s spam. Yes, there’s AI slop. That’s not a DEV-only problem — that’s an internet-wide entropy problem. The real question is whether a platform can still surface signal, and for me, DEV still does — especially through curated challenges, featured posts, and communities you deliberately follow.
I think what we’re seeing isn’t community death, but community redefinition. People still want a space that feels like “us” — where they can flex ideas, bring personality, and grow without being instantly optimized out of relevance by algorithms or paywalls.
From where I’m standing, DEV isn’t collapsing — it’s becoming one of the few remaining places where developers can still arrive unfinished, and that’s exactly why I keep coming back.
The real risk isn’t growth.
It’s forgetting that communities don’t survive on traffic — they survive on belonging plus momentum.
So no, I don’t think DEV is dying.
I think it’s one of the last places still fighting to evolve without erasing the people inside it.
What really resonates with me is the idea that the decline of online communities didn’t start with AI – AI just pulled the curtain back. The rot began earlier with UX designed for addiction instead of learning, and business models that treated contributors like free raw material. StackOverflow didn’t die because people got lazy; it died because the platform slowly turned from a workshop into a courtroom where asking the “wrong” question felt like a misdemeanor.
DEV feels different because it still tolerates imperfection. Real learning is messy, repetitive, sometimes naïve. The old internet understood that; the new one demands polished expertise before you’ve even begun. Gatekeeping protects quality, sure – but too much of it sterilizes curiosity. Wikipedia survives because it has a clear mission. StackOverflow forgot its soul. DEV, at least for now, still acts like a neighborhood rather than a bureaucracy.
The real battle isn’t human vs AI. It’s curiosity vs metrics. If communities reward honest experience instead of volume, SEO glitter and “10 tips to 10x your life,” they’ll live. If not, we’ll all end up talking to algorithms about algorithms, wondering where the people went.
Posts like this always amuse me. Not that it's a bad post, by any means, It's just funny to me how it entirely ignores the obvious: In 2026, it's easier than ever before to break out of "web 2.0"
Even before AI it was relatively easy to host content on github pages; now that anyone can ask a computer how to do it and even to spit out some CSS to make the content more presentable, the barrier to publishing is $0 and maybe half an hour of time.
By the way, did y'all know that RSS is still a thing? And it does a splendid job at the one thing it's needed for: checking a collection of sources for updates and notifying the user so they don't have to manually check.
Everything else the quickly sloppifying web 2.0 provides is just meaningless fluff that adds nothing to the lives of authors and readers other than quick and cheap dopamine. Its main purpose is really just getting more eyeballs on ads.
A healthier internet is out there, people just don't want to use it.
I do, but I do miss the dopamine when I blog write-only and don't even see traffic spikes in my classic blog's web stats. But I will prioritize IndieWeb this year, I promise :-)
Some comments may only be visible to logged-in visitors. Sign in to view all comments.