These type of posts seem to be a weekly thing now on the forum.
AI isn’t a threat to Bubble. It’s an asset.
IF we get to the point where you can enter a prompt and develop an Uber, we won’t need apps at all. AI will just handle everything we want it to without needing a third source…
we have a long ways to go before we get there.
The difference in activity on the Bubble Forum may be somewhat related to AI, but I do not see it as an indication that AI threatens Bubble.
The Bubble Forum was the main learning resource early adopters had. It was a place where you could go to for help when stuck on something as in the earlier days, there was only one youtube channel for Bubble content, one set of how to build apps video series and the bubble documentation (which wasn’t great but nothing like it is today). Because of this, there was lots of activity where people were sharing the things they came up with that likely had not been done in Bubble before.
I think that Bubble has done a good job of flattening the learning curve, which was a top priority when they got $100M in 2021, which helps to reduce the need of new bubble users to frequent the forum. The community has done a lot to expand the number of developers using social media to boost their own presence, so there are tons of free resources available for people to learn from in a format (video, long form articles) that is a bit different from the forum.
Additionally, Bubble launched their AI to new users months ago, which has apparently helped to increase new user retention, likely because it is making it easier for new users to get a sense of how to build in Bubble a bit quicker, which again, reduces the need for new users to use the forum.
Bubble has added in components, which makes it easier for a new user to get a grasp on how to setup the UI for the the features they might be trying to put together, which may lead to less users needing the forum.
Also, as the forum was the backbone of what made bubble great (look back at any listicle on no-code tools and the Bubble community was the number one asset listed under Bubble), as long time users began to prioritize their own social media, agency blogs or newsletters, their became less engagement from them. Most of them are still building on Bubble and have thriving Bubble agencies, but some have left due to personal issues related to changes in Bubble.
And then of course there is AI. Now there are tons of new Bubble users opting to use ChatGPT or something else, to just ask the question. Of course, the answers are not always optimal, but still people are able to piece together some feature sets on their weekend evenings, giving them that motivation to keep building as they see success in some features implemented. So, since AI is able to help people building in Bubble, there are less new users in need of the forum as a learning resource.
So, from my experienced perspective, AI tools that help build with code are not the reason the Bubble forum has less activity than two years ago. Instead it is more to do with the maturation of Bubble ecosystem, which is a great sign that Bubble is alive and well, and not going anywhere.
Just look at the flurry of activity around Bubble mobile on the Forum right now…
When AI reaches that point then it’s not only Bubble that is irrelevant. It will be a monopoly by whichever company leads the sustainability of your ideal product.
Funnily enough people will then complain about lock ins and then another Bubble pops up.
Perhaps because you don’t understand how the code that would achieve this looks - I’m not trying to be snarky, just illustrating that the dunning kruger effect with AI code is huge and you cannot build something you could not explain yourself.
I know you typed this yourself but I have to laugh at it inadvertently being one of the biggest GPT syntax patterns.
@Bennn it would be easier to try and answer some of your questions if you shared some of the apps you’ve built with AI and how the development process compared from your Bubble-built apps.
That way we can all talk about pros/cons or do feature-comparisons. Otherwise it becomes a more generic “this thing happened, what’s your take?” and, in my opinion, makes it less productive as a whole.
@Bennn why do you care?
If AI gets that good then you’d just leave bubble. And it would be easy.
I used to feel the same way. I stopped caring because, if this happens, I have bigger things to worry about than bubble…like why would any business need software.
I feel your pain though.
Yes, I actually did write it myself 
In fairness though, I believe it was also in a post somewhere on here that Bubble had put out.
I do use AI a lot to help me when I’m building in Bubble. It’s a great help. Although. I’ve learned to not always take the first answer it gives because sometimes it wants to overcomplicate things and also sometimes gives results that are not able to be done in Bubble. So I keep prompting it until it comes up with a reasonable solution.
I think Perplexity is the best I’ve found when working with Bubble.
I think that instead of worrying so much about Vibe Code, Bubble should really pay close attention to what’s happening with other low-code tools. AI is accelerating their evolution, pushing them closer to becoming true no-code platforms, with the added benefit of simpler and more affordable pricing structures—and in some cases, access to the underlying code. In the end, the free market will decide… and I hope to benefit from it. 
Great post and I agree. Have been on Bubble since 2020 and built my first production app for in-house use over that time for my business. Learnt a lot about data structure, logic, timings, efficiency, Ui and UX with feedback from my users. Was happy and tolerant of Bubble’s flaws. ad it allowed me to perform what felt like magic. I decided to rebuild the best parts of my learning app as a Saas rather than just an internal tool late last year. I was planning to use Bubble again, due to familiarity and experience , but realised I would need to rebuild it in code later in any case. Started buidling in Bubble again in late Feb 2025 and then realized that Ai is advancing fast, I will go through learning and iteration as with Bubble. The challenge seemed harder with code, but I decided to see how far I could get. 4 months later I have a full Saas with 43 annually paying users ($1k each. The app is much more capable and efficient than what I had previously built over 4 years with Bubble. Each time I got stuck in code, there seemed to be a new model that broke through. I started with Lovable for 2 days, but it was too limited. Then Windsurf l for a month, Cursor after than and lately Claude Code. I learnt a lot again and can now iterate my app at lightning speed using Ai agents. Compared to the editor pain with Bubble, Ai is a breeze. It wad a tough journey and I had no guarantees it would work as I am not a coder. Surprisingly, pushing through the frustration and learning ad I went (just like with Bubble), I am now miles ahead compared to where I would have been if I used Bubble to rebuild. Security is a concern, but as long as you ask your Ai agents and learn a bit, that is solvable as well. My app is MUCH faster, more responsive and better featured than my Bubble app. It has been much more stable for my users as well. Fewer bugs, happier clients and happier me. I love Bubble and what I learnt while here, but I am all in on code now and still don’t know how to code. I understand logic and systems though. That is very important with Ai code. I wish Bubble all the best. Thank you for being part of my life and up skilling me to the next level.
Sounds like cope from my perspective. Sorry
Someone please create another category here in the forum titled “AI Threat” 
Agreed! That will really help with the wait for next month’s “AI will kill Bubble because xxx” thread
Sounds to me like you don’t know that you don’t know what you don’t know.
Ditto 
No, because I mentioned it first.
Interesting.
Bubble teaches you it all back to front.
What stack did you go with?
I used ai “vibe coding” recently to build a plugin for bubble. While I got the result I wanted in the end, the journey was extremely labourous - over 80 hours of prompting and testing, over 150 versions. My takeaway is that while possible, there would have been NO WAY I could have just given it one prompt and let it build. The whole process revolved around me knowing exactly what I needed, and the right questions to ask. For example questions regarding memory leaks and clean up, edge case uses etc etc. Many times did it get stuck in an endless loop - creating errors, trying 15 ways to solve it, then going back round in circles offering up the same fixes. I ended up using three different Ai models to finally get it finished (Calude4 proving the most capable for this task) At many points I had to ask it to consider new strategies and make my own suggestions about how best approach a problem.
Without my experience gained from 5 years+ with bubble, and being able to think 360 degree about the software, I simply would not have known HOW to ask the RIGHT questions.
My point I guess, is that AI coding is incredible, but as George says, you need to have developed the brain patterns to think like a developer, in order to get the most from it. It’s a co-pilot - and if an inexperienced developer thinks they can easily create anything complex with a few basic prompts, they’re likely in for a shock. There are no doubt instances where it’s a perfect fit , and that goes for bubble too.
Will be interesting to see how this landscape develops over time for sure.
Side note - if I were one of the ai coding startups I’d probably hire people to write posts like these in the bubble forum every week
Bubble certainly did along with logical and structured thinking. Glad I was on that journey. Now using thr stack which Ai does best. Nextjs, React, Supabase hosted on Vercel.
You make solid points given where AI stands today. However, consider the rapid pace at which these agentic systems are improving, with 2x, 5x, or even 10x gains in capabilities, something we’re already starting to see empirically. Today’s challenges, like secure hosting and API management, are temporary hurdles.
When models scale significantly in intelligence, combined with ongoing reductions in energy costs, many current issues you’re highlighting will likely vanish quickly.
Also, regarding your point about AI currently only producing basic web apps: the vast majority of Bubble’s current revenue likely comes from users building exactly these “basic web and mobile apps,” evident from their marketing and strategic priorities. If AI reaches a point within a few years where it effortlessly handles even slightly more complex tasks, Bubble’s core market faces a genuine existential risk.
