Great news y'all. Support Dept has just reached Gold 🏆 Agency tier status with @bubble! We're just three humble guys turning up every day and working hard to deliver great experiences for our clients. Thanks for all that you do @heyluizandrade & @hectordmontoya1 And for a bit of fun, as the other gold agencies are quite concentrated in the US and Europe we get to make all sorts of bold claims now 😅like being the only gold agency headquartered in the APAC region, or even bolder - the Southern Hemisphere🌏

I agree - although I think it depends on your audience. Most people have no interest in building their own software so they will always look to something or someone else to do it for them - be it SaaS or an LLM. At least until they too are replaced 😂 Having said that, the market for micro-SaaS will no doubt explode as the barriers continue to fall, so with greater competition we’ll see prices fall too. With lower prices, the incentive to bother building also decreases.

Thanks @fede_bubble 🙏 appreciate you. I know I’ve become more and more critical of late, but do know that my criticism comes from a good place. I want nothing more than for Bubble to succeed, and I have no doubt there are so many competing priorities currently and only so many hours in the day. It’s tough. Just to build slightly further on this: I think Bubble was quite strong early on at doing an awesome job of abstracting away the complexities of building software, and then it kind of just stopped… yes the team has been shipping loads of updates and things get incrementally better month by month. But the core functions a bubbler touch everyday only ever get window dressings and fundamentally haven’t changed all so much in the last few years, other than the responsive engine 🙌. Meanwhile expectations from devs, clients, users never stop moving 🚀 Marketing may disagree, but ultimately what people want today whether it’s day 1 or day 1000 is for things to be quick, easy, and robust. AI is hot because it at least feels to be the case. If you had tailwind UI or the full shadcn components library on Bubble for people to drop into the editor, people would be falling over themselves to use it. AI prompting is great and doesn’t need to be an or, but this would be way quicker and less ambiguous. The vast majority of my clients would be over the moon with Shadcn UI updated with their brand colour palette. On basically everything else, what I see the AI tools do is basically build “recipes” that the AI can then implement. And while every app is different, the core fundamentals don’t change all that much. If there were recipes in the workflows tab, database, option sets, API connector, so much complexity gets abstracted away. Simple example: if I build a google login button today 1. Firstly, I see nothing about it anywhere, so I have to hit the forum, read the manual, watch a YouTube video, channel devine thoughts. Basically go somewhere else to try to figure it out. It’s a whole thing. 🫠 2. Now I need to build a button and spend some time styling it. If it’s day 2 and not day 200, I’ll probably make it look terrible. 🤷‍♂️ 3. I need to know to and then install the google plugin and enter my API key. Shit, how do I do this? What’s google cloud console?🤔 4. I need to create a workflow and while there’s a nice easy step, I still need to add extra steps about what happens next for it to actually work and for me to know where the data goes 🥱 If I use an AI coding assistant it asks me for my API key and tells me how to get it and then handles the rest. Theres no reason Bubble couldn’t have a “sign in with google” element that I could drop into my UI (or prompt) and this is done for me. Still no AI required, but achieves the exact same outcome. Why was it not always like this? I think the disappointing part is that it feels like we have to beg all the time for things to progress on real day 2 challenges that aren’t sexy, but do matter and impact so many every day, and ignoring them only helps the competition. @ClausISaidSo is just so awesome for all that she shares, as are many others. If only you had a dev assigned to fix Claudia’s many grievances 😂 all along.

@BubbleBackups @mneary0 @bubble @nocodepete Exactly right!

Looks awesome, but the one thing I don’t like is the need to work from their base template and then export code. Otherwise 🔥 There are other similar tools coming to market that sit over your existing code and allow you to easily switch between their app and say cursor on the fly, meaning you can always use the best tool for the job in that moment without friction.

@kennandavison Icon

@emilahlback @lovable_dev 👋

Exactly. Many moving pieces and it’s hard to know if there will be a clear winner in the long run. I suspect there may not be, but the one limitation no-code tools need to address is how closed they are. If everything reaches feature parity this is the one big detractor for them. In the short-medium term it’ll probably bounce around all over the place. I think the best advice is to do what makes sense today based on what’s available and your skills.

Depends on your timeline. If you want to launch an app from a standing start today, I’d still recommend no-code tools. We’re just not there yet to be able to build, deploy and manage an app with AI without any coding skills or experience. It’s getting way easier week by week but there’s a lot of pieces of the puzzle still. Having said that the space is moving super fast and it’s always hard to know if that pace will continue. It’s possible that in 6 months time there will be better AI and better tools and many of those challenges will disappear, or we could be 20% better than today 😅🤷‍♂️

@Elil_hrair_rah_ @mneary0 @bubble And I think that makes a lot of sense. I’d much prefer 3-4 tools that work seamlessly together but are each focused on delivering specific value and obsess over making their piece of the puzzle awesome. A Swiss Army knife will be good for all situations, perfect for none.

@GavelCoding You need to reframe your thinking and see where you can add value regardless. Often what my clients ask for and what they need are two totally different things. Helpful AI isn’t so helpful here…

@mneary0 @bubble 😅 I really should compile a list so I can track them, but here’s a couple that come to mind

I’ve also seen 3-4 visual code editors (albeit very early) that looks a bit more like webflow but sit above your code base and have some inbuilt AI generation. No code import/export - you can use it a little or a lot. You can also edit the code in cursor / vs code at the same time, so no lock in, just use the best tool for the job at that very moment. All sounds compelling and they are moving fast…

@mneary0 @bubble @nocodepete I’m surprised NoCodeFusion and @nocodepete haven’t been acquired yet! 😅 Definitely changes the game! Awesome work.

Can you dynamically set the timezone of your picker, just set that to the dropdown value and save the date. Then you can just use that date to schedule the backend workflow. Remember date/time is absolute so there’s no reason to worry about timezones in the backend in this instance. You only need timezone overrides in the backend if you’re setting an arbitrary date/time, and then timezone overrides are important

I don't think it's that, I think perhaps they don't keep up to speed on the latest tooling and approach to development so their ideas about what good looks like and what tooling they need to provide is outdated. I think what we have made a lot of sense 10 years ago. For example, I'd say a very high % of traditional devs who design their own UI would use some derivative of tailwind and a UI kit that sits over it for the sake of consistency and to get something that's clean and functional and to avoid recreating the wheel on each project. A very high % of Bubble devs design their own UI yet are stuck with basic elements from 10 years ago, very primitive base styles, and a blank page. Regardless of experience level, you burn a serious amount of time trying to overcome this every day. Even if you do work with a designer, there's still so so much basic nuts and bolts work needed that simply doesn't exist on other platforms or with code. When you go to the backend, I just don't think they understand the complexities of what clients expect logic-wise, or what's needed for certain integrations, so there's lots of unnecessary complexity and workarounds. But like, if I used my backend to trigger a welcome email, it's perfectly capable.

I think @JJEnglert made a lot of good points about this the other day. It's difficult to compete as a lot of these new tools have significantly more funding, don't have to deal with legacy issues, and will also just get better as the underlying LLMs improve. But separately from that I think the biggest reason is that they're open. If these tools operated in a vacuum and didn't allow you to fully leverage them for a part of a bigger puzzle, they would never have taken off.

I agree, I think there are a lot of small quality of life improvements they can make first before layering on AI, and will actually make the AI implementation much faster also. But the top of that list is proper logging. It's a real big issue. But for the UI part, if they just implemented even just one extensive UI library with proper modern elements and components it would save a considerable amount of time. Again, then use the AI to stitch those together - it's a lot easier than trying to generate everything from very basic foundations.