Is anybody else worried about Bubble?

Meta
21

This is exactly it. I’ve built like 40 Bubble apps, several of those which I’ve monetized and made quite a bit of money with. And to me, I really don’t care how things are made. Ultimately what I want is scalability to be able to maintain my apps, to be able to iterate and expand them and improve them. But at the same time, Bubble has been key to building MVPs which I now can build easily with Cursor or Manus in a matter of minutes.

For those saying that AIs are currently not smart enough to build complex apps, simply look at the improvement rate for the past few months and tell yourself that it’s only a matter of time. And I believe that by the end of this year, we will have smart enough artificial intelligence to build complex apps.

22

I’ve already built a couple of single-use, very specific apps to meet various needs with a single prompt, so yes.

23

Hi there, I’m just chiming in to point out the continued growth & investment the Bubble team is clearly putting into this area:

image
image513×301 22.2 KB

Speaking as a non-technical founder, having Bubble handle the full stack was game-changing for me. Seeing their AI capabilities grow within that infrastructure is very exciting. I don’t know what the future holds, but I’m still excited to see where Bubble goes from here.

2 Likes
24

Yes, I do. If it’s not just a series of a few prompts. what’s the point?

I mean what are you doing? Like “set up the code to make this dropdown and include these terms?” If so, what’s the point? Bubble would be easier. And you’d still have to have full understanding of how to compile and debug it as well as add all the features it didn’t include and remove extras. There’s no way it’s better

It would only work well if you really REALLY know what you’re doing and can have it generate very precise code in a language and dev environment you’re comfortable with so you can put the whole project together. By then you might as well just use whatever packages and libraries that are available.

1 Like
25

Is this something you see Bubble apps as not capable of doing on Bubble?

26

This is actually hilarious because he is describing exactly what the non-dedicated “Core” cluster is. Thousands of Bubble apps running on a shared monolithic swamp.

This is the most important point. Some people find coding fun, they like puzzles. When you’re trying to solve hard business problems that last thing you want to spend any time thinking about is technical details and arbitrary puzzles. Unless your “hard business problems” are related to software itself, obviously. Personally I find coding extremely frustrating, even for languages I understand, because computers are so [insert word here I’m not allowed to say] compared to the average human and have a meltdown if things are not specified exactly right. Like if you go deep enough the math at the very core is pretty cool, the very top-level outside abstractions are cool, but I have no patience for the middle layer where 99% of SWE happens.

27

I have no idea how Bubble servers work, but if this is the case, then I suspect even Bubble is being suffering from a suboptimal system design.

Serving millions of users with a monolithic app deployed in a single instance is crazy:

  • Scaling vertical scaling of resources is more expensive than horizontal scaling, and the latter is prevented when designing monolithic software, so at the end you are charging your users much more just because of this, reducing also your business profitability.
  • Single point of failure with all the negative consequences it implies for a business

Proper system design is not a whim, it’s business strategy thinking.

I agree that most businesses won’t be popular enough for the benefits of a proper system design to be noticeable. But if you happen to have one in the top 1% you’ll need to adapt.

28

This.

But Bubble better keep leaning into AI hard, because things do keep getting better (slowly but surely).

1 Like
29

I don’t believe vibe coding will remain limited for very long, and I have several reasons for this that go beyond the current hype. Historically, there has always been skepticism about what artificial intelligence could achieve, and time has proven many of those skeptics wrong. For example, before Deep Blue defeated the world chess champion Garry Kasparov, many experts claimed that AI would never be able to surpass a human in such a complex game. That changed dramatically and showed that AI can exceed expectations.

Today, many still say that AI cannot fix complex bugs, create robust applications, or that it will take decades for AI to build decent systems. But it’s important to remember that just a few years ago, similar doubts existed about AI-generated videos. The launch of Google’s Video 3 revolutionized the market with super realistic AI-generated videos, proving that rapid and disruptive advances are possible.

In my view, vibe coding is the future of software development because it delivers the actual code of your application, unlike lock-in tools like Bubble that limit developers to closed platforms and restrictive visual interfaces. With vibe coding, you have the freedom to generate, adjust, and understand the code, allowing for greater flexibility, customization, and scalability.

Moreover, vibe coding democratizes development, enabling people with little technical knowledge to quickly create functional systems, while experienced developers can accelerate repetitive tasks and focus on more complex problems. The trend is that as language models and AI tools evolve, current limitations will be quickly overcome, making vibe coding increasingly powerful and reliable.

Therefore, just as AI surprised us in chess and video generation, I believe vibe coding will exceed current expectations, establishing itself as the best approach to create MVPs and complex systems, offering freedom, speed, and quality that traditional tools cannot match.

2 Likes
30

A friend calls it “fancy autocomplete”.

31

@myjourney it’s not hard to feel this way. I often do too.
But then I see that theres a lot of hype.

But as of yet, I’ve been unimpressed with Bubble’s AI implementation.

I’m unimpressed with all of the vibe coding implementations seen.

I’m worried about Bubble because I feel like one of these days…

One of these days, everything changes. I get it. I’m invested too. But the way I see it, Bubble will be irrelevant once web and mobile applications are irrelevant. However in 2025, we are still building web apps. More web apps not less.

It’s only a matter of time where creating apps will be a single prompt

When humans build apps, we ask questions. A lot of questions. I can never see a single prompt producing anything meaningful - other than perhaps a todo list - of which we already have many.

My fear is that Bubble isn’t moving fast enough

Quality is more important than speed. It’s not good to be first to market with a useless product. Let’s give them a chance, and put their Series A round to good use.

1 Like
32

What you are describing here is already how coders need to work to stay relevant. Coders. Not vibe coders. Honestly all of this confusion is because of the terminology (I really don’t like it), it creates an unnecessary divide in programming. It’s all just coding but AI boosted.

The real distinction is if you know how to program and read code, or not. It’s like copy and pasting functions from Stack Overflow “back in the day”. You either understand what it does, or not. SO or AI, you will pay the price for ignorance.

No sane person will entrust their business to pure AI generated code. Not even the people who create AI. If that was the case, then all code AI generates will already be minified.

If you think about it, when the time comes when people can entrust AI to build, maintain and troubleshoot programs…AI will probably come up with its own super optimized universal programming language and then there will be no need for web developers.

1 Like
34

If there is any concern about Bubble future, it would not be due to AI and vibe coding. It would be due to Bubble not fixing broken features, updating performance and uptime and all the other things that make people dissatisfied with Bubble like forcing a new workflow editor that after two months still doesn’t have the simple CSS fixes requested applied and seems to have found it’s way to the shelf of ‘forgotten and abandoned not fully realized features’.

4 Likes
35

Yeah this. I really hope that after mobile/AI ships they just completely stop shipping “features” and work only on making the core product flawless.

1 Like
36

If that were to happen, competitors would likely not be able to catch up to Bubble and they’d have a safe and comfortable industry leader position. We could all only hope this to be the case, because after all, what else would need to be done?

1 Like
37

How Bubble.io will survive and thrive imo…

  1. Focus on performance, scalability, security and reliability. Ultimately, this is what matters most. The small bugs have a negative effect on reputation and people NEED to trust the platform & it’s staff. (On that note I also include customer support in this - once stellar, now very average).

  2. Make the visual editor best-in-class.
    Sure, integrate AI to spark ideas and assist with tasks (for example, “draw me a menu bar in the style of honeycomb”), but then make editing and tweaking that output faster and more intuitive than re-prompting.

  3. Allow code export.
    Users should stay not because they’re locked in, but because they choose to. Supporting code export would show confidence in their producd and if Bubble nails the first two points, most users wouldn’t need or want to export code anyway.

  4. Marketing.
    In time, most visual builders will converge toward similar systems anyway. Bubble has a head start… It just needs a strong marketing team to clearly explain to new users why building the Bubble way is the better choice.

In my opinion if Bubble can deliver on these areas, it will remain healthy and prosperous for years to come. But now is the time to work really hard to get the core strong & direction right.

38

Vibe coding ≠ AI assisted coding

Vibe coding will never be successful.

AI assisted coding, where you fully understand what you are instructing the AI to do, and what it is doing, will be successful.

Bad Bubble devs will be good at neither. Great Bubble devs will be good at AI assisted coding as the principles of good software development are the same in Bubble and trad code.

So, if you’re a great Bubble dev, I wouldn’t be worried.

If you’re a business built on Bubble - as of last October, Bubble isn’t burning through that much money considering they raised $100 million. I’m not going to say they wouldn’t go bankrupt because, well, that will certainly mean they go bankrupt, but I wouldn’t be too worried about that either (though do recognise the platform risk).

2 Likes
39

I’ve been working with AI assisted coding for the last 8 months or so and with the amount of progress I’ve seen in that time, it would be foolish to think this method of coding isn’t the future. I’ve done things with these systems that took 20-40x less time than in Bubble. @myjourney This includes building massively complicated apps but it does require a lot of handholding, a fair bit of money and serious technical skills.

The 1st major issue with most of these systems lies in the fact that it isn’t a fully integrated environment like Bubble but the progress I’ve seen in this regard over the last few months is nothing short of spectacular.

The 2nd issue is that once you’ve learned how to craft prompts etc. you’ll find that there are limitations that, once reached, requires serious coding skills in order to move forward.

In both cases, Bubble has an amazing potential competitive advantage if they can plug AI prompting into the regular development pages.

Right now, if I want a new report, I can budget 2 dev days in Bubble due to all the complexities we have in our organization. (It just takes time to layout pages, set up DB connections, set up conditionals, build download functionality etc.) On AI based systems I can accomplish the same in 5 minutes IF something does not go wrong.

Fingers crossed Bubble brings us AI prompting goodness sooner rather than later. (Then we can start complaining about DB performance issues again. :slight_smile: Or perhaps the AI will make it easier to manage our SQL connectors as well… :slight_smile: )

5 Likes
40

We are now 3-months into seriously working with Vibe programming platforms. This week we started our first paid app builds for customers after building 10+ internal apps including a CRM gomillioncrm.com

While I echo the challenges, at the moment you can get 80% built using prompting alone but you need a coder to get you over the line.

However, it feels a bit like Bubble did 5 years ago you have to fight and do workarounds to get something over the line but you can make it work. It’s so much faster to get something to 60%, we are talking 1 day, then the next 20% takes a few weeks and the last 20% requires a coder and takes a few weeks. So overall its not much quicker but it totally shows the direction of travel.

The question of it being less of a platform like Bubble and more of just an IDE that you have to manage all the component parts (i.e. servers, databases etc) has been mostly solved by some of them so this is less of an issue in my mind.

If Bubble doesn’t pull a rabbit out of the hat with the AI then I think it is in serious trouble, the number of founders we have coming to us with a lovable site has gone from zero 6-months ago to now about 3 in 10 and growing.

Simon

2 Likes
41

Isn’t this evidence that founders are getting stuck with AI, and AI tools are just making more people realise they can have something built for them for cheap? We’ve also seen Lovable prototypes from founders who come to us after getting stuck and finding Bubble.

The use of ‘vibe coding’ as a terminology by agencies is also so concerning to me. Vibe coding means letting the AI do its thing without understanding what it’s building. That always, without fail, ends in a disaster. If something breaks, you can’t just blame the AI. I can’t wait for all the vibe coded slop to fall right into my own agency’s hands if we ever move in that direction!

6 Likes