What will be future of bubble?

flowforges · 2025-01-25T14:08:00.608Z

Did you tried loveable and Bolt?

lindsay_knowcode · 2025-01-25T14:19:56.585Z

flowforges:

Did you tried

Do you know of any other AI tools where I can upload screenshots and have created a Web UI? I am just using it for images in a design doc, so it doesn’t need to be proper code - just saves me creating wireframes (which I am average at anyway more average than an AI tool )

flowforges · 2025-01-25T15:01:49.477Z

To be honest they are still far behind from bubble and it can take years to build a visual editor like bubble but still they are powerful for prototypes and basic MVPs.

So bubble will need to keep adapting otherwise within next 4-5 years no one will even use bubble.

design.agx · 2025-01-25T16:01:56.064Z

I have also been using Lovable and it does a pretty decent job if you prepare your prompts beforehand
Set the database first and then build the UI page by page adding features.

They just launched their figma to lovable and it’s pretty good, and soon you will be able to sync your project with the changes you do in Figma

Also prompting with images of the UI you want, helps a lot to get it right

Delightful · 2025-01-25T18:41:30.221Z

flowforges:

What do you think will be the future?

Bubble will burst as current WF and limitations are non scalable for sure
There. Are plenty of better alternative already and time will make everything clearer.

psycholabdesign · 2025-01-25T20:20:33.901Z

Having looked through lovable doc and tutorials, definitely, you’ll need good prompt input and still will make changes and also connect APIs, DB, and stuff to deploy a live app.
But things are moving quite fast if you look at other builders.

Bubble may be in trouble in some cases but will still be used long into the future. And if they step up the game with reliable performance maybe longer, and with the introduction of mobile and better prompt engineering that can create db setup, design, and logic even more people will be able to build things with natural language.

Again, this month was a bad month for the platform’s reliability. Lost some work due to main cluster issues.

Screenshot 2025-01-25 at 22.13.242170×1540 306 KB

Screenshot 2025-01-25 at 22.13.361198×1494 105 KB

ihsanzainal84 · 2025-01-26T00:36:49.133Z

Tried Bolt and a bunch of AI builders. Great stuff if you’re an experienced developer.

AI deploying stuff for you is great and all but troubleshooting and scaling is going to be a bitch if you don’t understand what the AI has deployed.

Until the day I can use AI to deploy, manage, troubleshoot and scale an app, I won’t be trusting AI to build my serious apps by itself.

flowforges · 2025-01-26T03:33:39.303Z

Let’s see

flowforges · 2025-01-26T03:36:11.443Z

The biggest reason founders are moving from bubble is the bubble pricing

mitchbaylis · 2025-01-26T04:46:07.407Z

I’ve found claude to be the best for prototyping

Vo is also very good

chatgpt seems to be pretty terrible

although the 1o is very good at coding (just doesn’t seem to be that good at canvas - claude still wins out there)

applepatient · 2025-01-30T22:00:41.594Z

hey that is amazing. congrats. I am a beginner bubble dev with zero background in coding. But now at this point, i have grasped a lot of fundamentals about programming, though no hands on experience with coding. I bought a course which teaches you all the fundamentals you need to know to understand and prompt ai to code. IS this a good start? i was gonna go traditional route of html/css, computer science basics, then javascript etc…

But i really want to start making apps using the exact method like you. I do not mind learning or figuring out stuff. So i am curious to know about your journey, your background a little bit if you can share something? Thank you thank you in advance. You will literally save hundreds of hours of time and energy with your response i feel.

fede.bubble · 2025-01-30T22:40:05.035Z

Good to see you decided to continue your discussion from this topic a couple of months back:

Is bubble pricing mode making bubble unsustainable? Questions

Is bubble pricing mode making bubble unsustainable?

Bubble has been working on really cool stuff (like an AI builder and our Mobile apps builder) but I think we could’ve done a better job at advertising and talking about it more. Essentially, “building in public” more. We shared some updates and demos during BubbleCon but we’ve been quiet since. That doesn’t mean we haven’t been coming to work every day.
In the next few weeks you should expect to start hearing more about what we are building as we correct this and start sharing more and also, more often.

My personal take? LLMs are getting cheaper and smarter, and that’s great news for Bubble as we introduce those AI capabilities into our editor. And there’s a lot of truth when app builders when they complain that AI is great for Day 1 building, but not for Day 2 maintenance.

emir.ozgun · 2025-01-30T22:58:58.300Z

AI and mobile was part 2 of new year road map, what about part 1 ? About optimization of the infrastructure for faster reliable db operations ?

bcart0v · 2025-01-31T00:54:19.252Z

Bubble will do fine.

Creating an app with a prompt is great, but if there’s an issue can it fix it? Can you fix it, as you don’t know the code that was generated?

We’re entering a new era of programming, where some Developers know to create code for any framework via any LLM, but couldn’t troubleshoot to save their lives. AI still can’t understand complex apps that work with many different moving parts, and probably won’t for a bit longer and that’s the main job of a Developer, fixing bugs.

Learning the Fundamental’s of programing is key to becoming a good Developer, the more you know about how something works, the easier it is for you to troubleshoot issues or even create new features / apps, regardless if you decided to build it via AI or code it yourself.

I’ve been coding for almost 15 years now, I use LLM’s to do mundane tasks and quick fixes, but I know wrong code when I see it, and most AI’s love to generate code that would never work and confidently tell me it will.

georgecollier · 2025-01-31T07:03:05.567Z

People also miss a very obvious fact.

Wordpress made it easy for anyone to build and own a site a long time ago. Yet, companies still pay agencies for its development.

In fact, people still pay a lot of money for traditional development, despite it not being necessary and still being more expensive.

Point being, the total addressable market for something that can build itself is much smaller than you think

sem · 2025-01-31T07:20:41.421Z

I think you miss an obvious point

Only a fraction of all the websites out there are build by agencies. It will probably be close to 1% and 99% is build by mums and dads.

When you talk about software it is the other way around. Still there are mums and dads that can write code but in that space it will probably 90% written by agencies/developers.

The markt for software is growing. More and more is getting digitalized.

In 5 to 10 years, software will be just like Wordpress or wix. You want something automated, you tell it to a Wix/wordpress like system and you get your software.

Most NoCode and even early stage AI builders like Replit will not make it. Most of them are building better tools. Eventually companies will dominate that build better solutions to the real problems at hand.

olatoye · 2025-01-31T08:30:30.345Z

My 10 cents…

Bubble will actually emerge stronger once it incorporates more AI capabilities into how we interact with it.

As a true non-technical person (cannot code at all), I’ve looked at these AI builders and what’s very obvious is they are built for for already-technical people. There’s too much jargon, the viewable code is distracting, there’s too many things to learn which the creators are trying to breeze over, unsuccessfully so.

I doubt AI coders can transform their user experiences overnight. This is unlike Bubble - if their strategy is right, they can adapt their platform to include the best that AI has to offer within 12-24 months.

senecadatabase · 2025-01-31T11:27:12.184Z

I agree most websites are built by mom and pops.

However, what a lot of people don’t grasp is that the real magic isn’t the website, it’s the content…and that takes skill.

Two websites can be for the same product, one can have almost zero results and the other one can bring in unlimited business. It’s all about the content, how it’s presented etc.

Most mom and pops have no idea how to create content that sells out warehouses…

they can sit down and put a website together on Wix or whatever…but we all know most websites are really useless. A lot more could be said about that.

A talented copywriter can still outperform any AI out there. Our agency has a couple of them that can beat any split test with AI.

Also, when it comes to AI for apps…

using AI as a personal assistant will be the future and will negate the need for most apps.

I will be able to get up in the morning and tell my AI assistant: ‘Make me an appointment to get the oil changed in my Jaguar’

It will schedule an appointment and also keep me updated with reminders etc.

Also, complex apps used for CRM etc. will be handled by AI assistants either for personal use or business use.

Lot more to say on that, but if you’ve read some of the opinions coming from the AI world by scientists who have worked on AI you’ll get a better understanding of what they predict for the future of AI.

What is harder to predict is what will replace things. Usually the replacements are things that nobody really saw coming. I.E., what will replace websites? That’s unsure at the moment, but with changing technology, a lot of dynamics are created and reversed.

Being able to pivot and change is always important.

flowforges · 2025-02-04T07:34:34.168Z

Learn from bubble academy

design.agx · 2025-02-04T08:46:42.722Z

I think it might take some years still to see the Ai agents being the ones who use more the Ii weren’t than humans

There are some early interesting projects and research going on.

llms-txt

The /llms.txt file – llms-txt

A proposal to standardise on using an /llms.txt file to provide information to help LLMs use a website at inference time.

github.com

GitHub - wild-card-ai/agents-json

Contribute to wild-card-ai/agents-json development by creating an account on GitHub.

They propose a different way that LLM’s can interact with a website or API.
Some people have even go so far to say API’s will stop to exist as they are made for developers and there is other ways to give the AI agents more capabilities.

I’m more hesitant about that, but exciting to see the future!