What Happened in the Bubble AI Vision AMA?
On April 3rd, Bubble held an AMA outlining their vision for how AI fits into Bubble’s future. Key themes included:
-
Bubble positioned itself as the only platform with a truly visual development language — something it claims gives non-technical users a real edge over tools that rely on generated code, which still requires reading and understanding logic.
-
A claim that platforms like Bolt, Lovable, and Cursor will struggle to stay relevant because iteration on complex apps will require technical skills that most users don’t have. To paraphrase: “writing and reading code isn’t easy. AI is not changing that that much.”
-
The company emphasized its core advantage: that Bubble apps can be understood by anyone, not just developers — an edge it claims is baked into its visual-first design.
-
A vague roadmap for AI tools — with no timelines, no clear definition of what v1 might look like, and no commitment to improving foundational UX or stability.
Takeaways from Bubble’s AI Vision AMA
Bubble’s founders made a bold claim during their recent AMA: they believe the real challenge isn’t how they will survive this AI shift—it’s how platforms like Lovable, Bolt, and Cursor will stay relevant when non-technical users build complex apps and then struggle to iterate on them. Their argument is that reading and writing code is still hard, and no-code tools that generate code won’t solve that.
I think there’s some truth to this… but also some smoke.
In fact, some of the very things they think give them an edge are exactly where these newer platforms may have the upper hand.
AI-First, Then Visual: The Backend-First Advantage
Bubble is taking a frontend-first approach — not by strategic choice, but because it’s effectively locked into that path after 13 years of building the current system. Meanwhile, platforms like Bolt are doing the opposite—they’re building backend AI-generation engines first and then layering UI wrappers over that.
To me, the backend logic is the harder, more critical part. Once AI-generated code is solid, creating a visual layer to interpret or edit it (e.g., “show me this algorithm as a flowchart” or “visualize these if/else paths”) becomes a much easier problem to solve.
So while Bubble talks about how hard it is to iterate without reading code, Bolt’s approach may actually make iteration easier, not harder.
Bubble’s “Visual Development Language” Isn’t That Visual
Bubble is stressing that it’s a true “visual development language”, something the others supposedly lack. But how visual is it really?
Let’s be honest: Bubble’s workflows are vertical-only, there’s no way to customize flow direction, and its UI logic often ends up buried in scattered conditionals or hard-to-read expressions. That’s not exactly “visual thinking.” The platforms that started with code generation and are now adding UI layers might end up offering better visual logic tools in the long run.
Bubble Costs 10–30x More — And Gives You Less Power
Not only are these new platforms evolving fast, they’re way cheaper. Bubble is often 10x–30x more expensive—if I’m going to spend that much, I might as well hire a junior offshore developer for $30–$50/hr who can listen to my “vibe-coded” feedback (e.g., “make this page smoother,” “fix this logic flow”) and implement it live. That might sound wild, but it’s still cheaper than Bubble for many real-world apps.
Who Are We Kidding? Bubble Isn’t Really for Non-Technical Users
Bubble often says it’s designed for non-technical creators—but look at the reality:
-
Most Bubble apps don’t use backend workflows.
-
The apps that do are usually built or supported by someone technical.
-
Native elements like radio buttons, file uploaders, or scrollbars behave inconsistently. Repeating group filters break or require hacky workarounds → Most serious Bubble apps end up with JavaScript, CSS, or HTML injected into them anyway.
That’s not “no code.” That’s code with a UI layer glued on top—and a steep learning curve for anyone who doesn’t already think like a developer.
The “No If Statements” Claim
Bubble claims that one of its strengths is that it doesn’t use “if statements”—because most of the world can’t read them. But… come on.
Bubble conditionals are if statements. They’re often deeply nested, packed into long conditional expressions, and spread across multiple UI elements. Some parts are even hidden behind “arbitrary text” clicks. If anything, they’re harder to read than traditional code, because you can’t see them all in one place.
So this “no if statements” narrative? It’s not just inaccurate—it’s irrelevant. You can’t build anything functional in Bubble without conditional logic. And if you truly can’t handle conditionals, then Bubble definitely isn’t the platform for you. Bolt and friends may actually be more accessible in the long run.
Other Takeaways from the AMA
-
API streaming is coming, which is one of the few truly exciting updates. -
It was disheartening that the only answer to the question “As a new user, is there a place where I can chat with a Bubble expert about AI to get answers about expressions and best practices?” was the forum. While Bubble’s competitors don’t offer any customer service, it’s still disappointing to hear it stated so plainly, especially considering that several “gold” agencies are creating awful builds. -
Bubble claims it can push two major initiatives at once: native mobile + AI tooling. Realistically, it can’t. Even worse, core things like stability, UX, reusable components, and native elements aren’t even current initiatives.
-
The idea that “building is free; only releasing into the marketplace costs” sounds good—but in practice, you can’t build much without hitting paywalls around workflows or database limits.
-
Bubble says there’s “no greater risk” of idea theft on their platform than anywhere else. While the risk of an idea being stolen on a Bubble app may not be significantly greater, it’s hard to say it’s not greater at all when Bubble still exposes DB schemas publicly, despite promising to fix that years ago.
Final Thoughts
I’m rooting for Bubble — I use it, recommend it, and have built many apps with it. But this AMA raised more red flags than it answered. The AI vision is vague, the timelines are nonexistent, and the core platform still lacks usability and performance.
The fact that the big teaser announcement was just a contest with $3K in giveaways — less than the cost of a dedicated instance for a single week — speaks volumes about the lack of truly exciting news on the AI front.
Meanwhile, competitors like Bolt and Cursor are charging forward. The idea that they’ll become irrelevant because iteration is “too hard” for non-coders — or that “AI isn’t changing that that much” when it comes to reading code — feels increasingly out of touch with where things are actually headed.
What Bubble should do next is probably a separate discussion — but there’s a strong case for leaning into what it already does best: being a visual front end. Partnering with AI-native backend tools, focusing on visualizing complex logic, or simply fixing core UX issues would be a more grounded path than trying to do everything at once.
; I really don’t want to have to move everything over multiple times…