I keep getting this from founder friends who’ve switched to Lovable or Cursor. Their argument…why learn Bubble when AI can generate a working app in 10 minutes?
My take is that the speed argument breaks down once you get past a basic MVP. Complex workflows, database privacy rules, perms, etc. I haven’t seen AI tools handle those . But I also can’t deny that the gap is closing.
Agree or stupid?
What say you? Anyone tried Lovable or Replit? Back to bubble?
I have been using bubble since the first days but i have to admit they are in trouble , with tools like lovable you they can only get better because underneath they are using claude models , fable came out and in 3 days you could use it on lovable , right now bubble can still fight but very soon there won’t be a bubble.
I think Lovable and Replit are even going to be dead soon too. You can just build it using Claude or Codex. Personally I just go back and forth. I just rebuilt my NoCodeMinute site and working on the rest of the features as well. What took probably a couple years to build, I can now build in a couple of months. The larger the app, the more time it will take to build of course. Landing pages take no time at all. I just updated my landing page in a day.
The thing is, you still have to know what you are doing. Bubble handles a lot of things for you that you don’t even think you need. I can only build strong and well functioning apps with AI because of all the experience I have with Bubble.
Point is, you still need to know what questions to ask AI. That’s why I still offer coaching even with AI. There are still important things to know when building with AI as well. 
Hope that helps!
Welcome to the forum @gjmfredrick
have you been building with bubble for a long time? Curious to hear about your background
It depends on the product responsible … if the responsible can build properly with the help of AI, maintain it and capable of scale, secure, stable, sure, use Claude or Codex… but if this person is more capable of delivering the business needs on bubble (because she understands how it works, capabilities and can maintain it in a more efficient way) then use bubble… I think the question may be: What kind of project are you referring to and who is going to build it and maintain it?
Yep. The AI gets better every day and Bubble and others ship faster…maybe? Hard to keep that pace. But I also haven’t seen Lovable handle anything super complex like user permissions, it starts falling apart. Maybe that changes fast though.
no, just started, trying to stay relevant
claude has me perplexed. the newest models not so great, but maybe I’m behind the curve
If you build to be ai compatible then ai just has to catch up and your system will scale the issue right now with bubble is they haven’t built with ai in mind and at the pace claude is moving sadly its just a matter of time
I have not tried any of the AI app generators.
What I can say from my experience using AI, following the hype around it and being wide eyed about everything to do with AI, the answer of why learn Bubble, is so you can build a production ready app.
AI builds fast, but it’s a toy in most respects, so it’s ultimately good for a bit of dopamine hit where you think you did something, with not much of a psychological toll when it fails, because you know you did nothing. That translates into, fail fast and cheap, a motto for MVP, so AI slop apps are good for MVP or personal toys for the average business owner they are pitched to…could an experienced developer use them to build real production ready apps, of course.
However, data so far, from big time tech companies, is that AI doesn’t provide ROI, costs more to use than employing humans and overall is actually around 20% slower than a human at completing most tasks correctly.
In today’s environment, Bubble lost to AI for MVPs, and enterprise custom software. But, Bubble still dominates for SME internal tools and production ready apps looking to scale. Likely will still see what’s happened since 2016, a newly funded via VC money bootstrapped Bubble app getting rebuilt in traditional code, but that doesn’t mean Bubble can not still have their corner of the overall market.
Based on recent statements in Forum, it sounds like Bubble is sobering up and realizing this, and doing something about it, not fighting the current anymore and instead swimming sideways to get to shore safely.
I don’t believe Bubble is dead because of AI. AI is a tool that accelerates development, but building successful applications still requires problem-solving, product thinking, user experience design, and implementation expertise. In my view, AI complements Bubble rather than replaces it, and developers who leverage both will have a strong advantage.
Everyone’s impressed by how fast AI can build version 1.
I’m more interested in who still understands version 127.
yeah, that’s what i was thinking. MVP is fine, but then it actually needs to work at some point. thanks for the insight!
i’m way away from 127
That’s why I still offer coaching even with AI since there are many people that want to build in AI but really don’t know how it works or what to ask the AI. There is still so much to learn.
And this is exactly the main problem with AI users: they focus on short term and doesn’t evaluate on middle and long term.
Like any other solution, AI can be part of the discussion but may not fit all needs and it’s not different for Bubble.
The whole talk about AI is really dependent on where you are with your business.
If you’re running a business with thousands of users… you’re not going the AI route with Lovable etc.
If you have an idea for a dog walker app that gets a few users, great, do AI.
If you have a business where every minute your app is down costs you thousands, no company in their right mind is going to use Lovable or whatever for their app.
I’m guessing you’re a hobbyist with an idea that hasn’t found users yet.
which means your question is for the hobby app builders.
Bubble won’t die because of AI; it will only fail if you stop innovating and offering improvements to native features—such as better backend and database performance—and if you change that WU pricing model, which drives away applications with high data volumes.
What does this mean?
To clarify what I meant: tools like Lovable are built so the AI model is the “brain” and the platform is just the “body” around it. That means when a better model ships (like a new Claude), the tool instantly gets better with almost no work on their end, they just swap in the smarter brain.
Bubble wasn’t architected that way. AI was added on top of an existing platform rather than being the core the platform is built around. So Bubble doesn’t automatically inherit those model improvements the way AI-native tools do.
My point isn’t that Bubble is dead today, it’s that this is a structural difference. Given how fast the models are improving, tools designed to absorb those gains have a compounding advantage over time, and Bubble would need to re-architect to keep pace.
