Open Office Hours - With Ray Deck (LIVE) - Episode 2

Hello, how's it going, Ray?

Very well.

Second time. Yeah, isn't that exciting?

Using, using, using the fancy new Restream technology.

Exactly, that's how you got to do it, right? That's— I think, I think it makes it very interesting, right? And we can see comments, and right now nobody's watching yet, but I think they're going to get the notification. Oh, we got one watching! That may be me, but we got someone watching here, so we'll just take a look.

Well, truly, we make the show for listeners and you make it for yourself. And when you're the listener, it sort of all comes together.

There you go. We already got one. But I remember we ended our last conversation, what will be true in a week that is not true today?

Yes.

What was definitely not on my bingo card is that the next time we meet, there will be no toddler anymore.

Yes. Nordcraft.

Nordcraft.

Nordcraft, which sounds like the latest massively multiplayer RPG from Blizzard.

I mean, maybe that's what they're going for. I was looking up Nordcraft and I was only finding Minecraft stuff.

Yeah, right. Minecraft, like a Viking flavor to it, right?

Yes, actually now I think they call themselves now Vikings or something. So it's— if they do add that northern touch to it, it's my—

my wife is half Norwegian, so you know, it's all— there you go.

So it does have that touch to it now, and the craft. So it's now, it's now all about craftsmanship.

Yes, we're going to craft the north. You know, it is, it's really too easy to sort of, you know, make fun of people for these kinds of brand decisions. They're very, they're very personal and they are, it's as much about sort of aligning the team and saying, hey, this is what we're about, right? Let's go. And the team colors that will then allow you to do better because your customers don't really care. Like we, another example that would be WeWeb, right? Also recently changed their branding. They didn't change their name.

Oh, they changed their logo to this weird AI thing, right?

Their logo, their style book, yes. And you're not the only person to have said that's weird, but like similarly, it's like what's gonna speak to them so that they can, you know, go out and produce in a way that's gonna be, you know, really outsized. And that's, you know, that's the job of it. So like when people go create new branding, it's like sort of easy to sort of take a couple potshots and, you know, sure they can be funny too. And being funny, I always approve being funny, But they're just going to do what they're going to do. And the real question is like, how do they create value for us?

Yeah. I mean, I do think their new website looks better than the old one, to be honest.

Are you talking about now about Nordkraft?

Nordkraft. I didn't like the last total website. I don't know. It wasn't fully responsive and the buttons were a little bit too big. It was— I found it a little bit weird, the last website. I liked the one before, but the, the last one was kind of weird. So I think the website is going into a better direction now. I think it looks more like— I mean, because I've done some really nice websites in Toddle, I was like, they gotta, they gotta showcase more what you can do with it if it's made in Toddle, right? You gotta have fancy animations and all that kind of stuff, and they're getting there.

Yeah, I mean, if you're going to have a— if you are selling web development services, you should have a very beautiful website, you know, like your new website. Oh, my new website also demonstrates I'm not in the— I'm not in the game providing web development services, because if I were, you look at that and say, no, no, we're not going to do that. Yeah, for me, like, the, um, the other thing about marketing is that it's an ongoing experiment. Right. You put stuff out there, you pay attention to how people are responding to it, you change it. And that was actually one of the reasons why I made the change to the site was that it was previously, it was a little too stuck in its ways because the editor I was using wasn't making it easy for me to move in a direction that would be really helpful to me.

You see your website right now?

I do now see my website. Thank you so much for putting that up.

You even got an animation on there.

Yes. I made a little animation, little, little, little Tailwind, you know, layout animation.

I just wanted to say, where did you build it? And you did it in Next.js?

Yep.

There you go.

So I largely vibe-coded my way through a Next.js app.

Oh, oh, now there are testimonials there.

There you go.

One sec.

We're going to— and like some of the most interesting visual elements are plugins like this from Sendra.

Oh yeah. Yeah, it's like an iframe. Testimonials. Yeah, you got like your wall of love, or what did they call it? That's what they call it, Beyond Vibe Coding. Yeah, that's great. So whoever wants to go beyond vibe coding, click, go on statechange.ai and go on join now and fill in your credit card information, right?

Yes, please do. It's, you know, uh, yeah, we just had a, just a superb office hours, um, actually all 3 of the ones this week so far. We had 2 on Monday and 1 so far today, uh, and the, um Just like the quality of the questions are coming through and the progress people are making is just extraordinary. But really is like that whole beyond live coding thing is a big deal.

Yeah, I just said you're always doing very exciting stuff on StateChange. Definitely worth checking out. Thank you.

Can I ask, now that you said such kind of things about StateChange, what's going on on NoCode/ProCode, especially now that you've gone from being very title-centric to you had been chomping at the WeWeb bit and now, hey, there's this Nordcraft thing.

To be honest, I think that the moment Toddle became Nordcraft, I don't know, it just feels different using it. I don't know, it's just different. I don't know, it's like if you use Toddle, I don't know, just the sound of it. Just like if you use Nordcraft, it sounds like enterprise-ready. But jokes aside, I do like the new name. I think their website makes it more sound like I'm actually using legitimate frameworks than just hacking my way through there with a no-code tool, which I appreciate. But yeah, we are definitely— we have some people in office hours who are now using Cursor that graduated from using Toddle to using Cursor. And I've actually built a React app in Cursor and I was able to render this using React DOM in Toddle. So I was like, why should I just go full React I just turned the thing that I can't do in toddle into a React component, build a little SDK, and thanks to your help from years ago, I just remembered some things what we did back then and I'm able to get React Flow running in toddle using React DOM.

Super.

Of course, you want to minify it because it's going to be very big. I minified it and then after minifying it, I minified it again. It's mini, mini fight.

Yeah, those things. Most people have pretty fat bandwidth. The bandwidth concerns we had back in 2017 when this was the main focus of discussion, and people can still be quite proud of having a 10-kilobyte web app, but who cares? You throw a couple of megabytes at somebody, they're never going to know the difference.

There you go. Mine was like 500 megabytes initially.

That's large. That's really large.

It's React. I had to get the whole React thing in there.

Yes, and you make a really good point about minification, of course, tree shaking, both of which are going to matter for getting—

I think I got it down to like 50. I think I'm doing well with that and it works actually well. I built it in cursor, it took me like 10, 20 minutes building that as a React app. Then I just put the thing that I couldn't do in toddle as a React component, build a custom component, imported that with npm. I was even able to host it for free on GitHub. I got it to jsdeliver and all that. Cursor helped me with the commands and asked for my GitHub password. I don't know if I exposed something, but it was very easy. Got it on there and it's now on GitHub floating around.

Excellent. Can I ask you, use the GitHub MCP server in Cursor or just?

To be honest, I have no idea. I just gave it my password.

That can happen.

But it works and I just create a new GitHub only for that. It's not on the Despic GitHub because my developers told me to not do that because we want to make sure that everything is secure on there. I do have my own little GitHub for my own little crazy ideas. In case something gets exposed or whatever, it's security compliant. We also have a comment.

Yeah.

You can see it actually in the chat. I can show it here. You both share the exact times when your office hours take place.

That is a perfectly reasonable question and I have absolutely no idea how to contribute to the chat.

For now it's been Oh yeah, there you go.

If you could maybe put that up on the, um, on the, the, the Chiron down there or something, we might be able to help, uh, Bill and Frank at least know when it comes around for state change.

Yeah, uh, can I— think I can put something in the chat. Oh no, you can post to some channels actually, but you can post it directly on the screen here in front of us.

Yes.

Yeah, let me post it in the chat. I think I put it in the chat. Now let me try to put it on the screen. We need to set up those links beforehand, but let's see. Let's do this in here. Let's put your link in here, State Change Office Hours Calendar. Now I'm going to create this.

Under.

And now I can show it, and we got it as a QR code.

Oh, well, that's nice.

I can edit it also to just put the link under there. There you go, look at that.

Brad, we used to be good at it, and we usually post for the current week on there. So if you take a look at the past events, you can get, get a flavor for like what the, what, what the usual cadence is. Uh, over at StateChange, we usually have 7 office hours a week. I think, Yacine, you run daily, right? 5 weeks?

We do daily 5 times during the work week. We do Monday to Friday. I can put my thing up in there too if you want to see that.

Yeah, please do.

We have it every day of the week, 9:00 AM PST, California time. Then what's happening is for Wednesday and Friday, we have it at 3:00 PM for the Australians. Yes, because they told me we don't want to join at midnight. So, yep, so we got that Australian approved.

That seems perfectly reasonable. Uh, yeah, the, the, uh, having— it is really a fun thing to have this global, um, audience be able to build something that people can use all over the, uh, all over the world, um, where people who are coming in middle of the day can pick up people from, you know, New York to San Francisco to, you know, Sydney, Australia, to Seoul, Korea, through Africa, through Eastern Europe, and all the way back around again.

We do have the same thing in my office. I just— we got people from all over the world. It's just interesting and very exciting seeing what people all over the world are building and that what they're building is all connecting them, right? It's like they are all using the same tools or similar tools, and it's just, it's just very great having this place where people can, can just come together and solve their problems. And just, you know, there are also very valuable conversations resulting out of those things.

Absolutely.

Very cool.

So what is— so, so beyond Toddle news, uh, what else have you been seeing?

Uh, I just wanted to say Nordstrom, but Nordcraft.

Beyond the fall of Toddle and the rise of Nordcraft, um, what, what, what, uh, what else have you been seeing in the market this week?

I mean, there has been also the rise or fall of WeWeb. However, I mean, people didn't really like the rebrand.

Yeah, I'm sure people at Toddle are going— the Toddle community might start feeling the same way, right?

Actually, it has been very positive in there, actually.

Oh, they have been? Excellent. I'm sorry.

Yeah, I was like, okay, Toddle is not there anymore. It's not Norcraft. Let me rejoin the Discord, right? You know, I think I found a loophole. And then Andreas said, you know, you're more than welcome to Nordkraft now. So I was like, great, that is— sometimes you need a new start, right? Sometimes you need a new start.

So, uh, so, so Nordkraft— my understanding is that, uh, Nordkraft is also their coming out party for their open source hosting?

Yes, they just released open source today.

So you're—

Instructions on how to actually set it up, at least I couldn't find them, but it is open source. It's on GitHub. You may be able to set it up. I would probably have to ask you how to do it though.

Yep, fair enough. I was sort of wondering, you often move pretty quick. I was wondering if you were already on top of that, but I kind of figured I'll ask at the end of the week then.

No, it's like there is no documentation yet, so I may need to join StateChange and ask you how to set it up, and then we'll piece it together and make a video.

Sounds delightful. Okay, so if we move beyond the world of Toddle, what else have you been seeing in the course of last week? Toddle, Nordcraft, whatever. I'm gonna— you know what, I'm gonna— I'm gonna ride or die Toddle. I'm gonna keep on calling it Toddle.

You keep on calling it Toddle?

I, I, I, I— yeah, the, um, because, uh, it's like I knew it from back when.

That's a good point, you know, like we're the OGs.

GPT Engineer as well is like that. That's the original name for Lovable. There you go.

Speaking of Lovable, Lovable is getting some traction, but their apps have lots of security issues because people just use this integration, of course. But I did try WeWeb. I do like their approach to AI. But I was like, wow, it's expensive. I've gone through like a million tokens in a day.

Really? Yeah.

Oh yeah. I do have it build custom components for pretty much everything because I'm using upload here and Weebup doesn't have a good event handling system. Remember when we were on StateChange and someone asked, how do I emit events from a custom component? It turns out you can't do it even though the feature exists. Because of that, whenever I build something in WeWeb, I have to turn everything as a custom coded component because they can emit the features. What I feel is WeWeb has hit a ceiling in how much they want to do or can do in this visual interface. They've just been putting an eye into it to give people all the features they've been asking for for a while. Without actually having to build them right now.

I think that is a really good point. That's my theory. And I think it takes us to a broader point about the value of AI in that we can give our users power to express their intent in ways that we didn't have to anticipate. When you are building in visual development, basically any visual tool needs to be something you have designed and presented to the user ahead of time. The way they combine those tools might be very different, but the actual tools themselves are are things you have to have created. And the MO is to be a gradient, right, between things that are powerful, easy to use, reliable, right? Make sure you have all three of those things, and like, that's challenging. And that limits your capacity. But when you have AI, it's like, hey, there's basically— here's the meta language, and how can AI express itself into that language? And that can lead to many, many more combinations without having to create the UI elements and facets and widgets that would be required if you—

I would prefer it in WeWeb because in WeWeb you can now do two things. You can use AI to build your custom coded component. You can go all code or you can go no code. The low code thing is completely left out now because of their AI approach, which makes development for them cheaper and they have to spend less money on building all those bells and whistles. Where what I like about Nordcraft, aka Toddle, is I can just bring code in there without having to go to VS Code or without having to use an AI. If that makes sense. It gives me that control where there is an opening, there is a gap, and I can fill that gap with some code. Where in WeWeb, I can't fill the gap myself with some code in their editor. I have to ask AI to do it for me, and then I can edit what AI did.

You have to ask AI in order to get the code?

If I want to emit— so it's a very specific thing, but if I want to emit a component event on a component I built in WeWeb, I can't do it natively using WWLib and their custom code. What I feel is they are relaxing when it comes to the extensiveness of the JavaScript that you could do in it, and they're now focusing more on the view that it can write for you, which I appreciate for people who want Prompt to do this. But if you want control over what you're building, I'm not too much of a fan of it, if that makes sense. Because it takes some of the control away. Because it's a very specific thing, but if I want to build a custom component that emits, for example, an event when there's JavaScript in it, I can't do that. I have to turn it into a custom component and create this component that can have a window function attached to it to then have this provider context component emit the event and it is like that.

Better.

And you can trigger workflows, so you would have to have the workflow ID. So when I build a template, I will need to trigger the workflow with WWIP and then change it again. It's not execute workflow anymore, it's something else. It's like global workflow or something. And, um, I need to hardcode the ID, where in Toddle they have a nicer way where I can just create the workflow and have callback directly in it for the JavaScript. Code, which I do like, which WeWeb, you know, like workflows can only go down, but they can't get out of it, if that makes sense. Like you can only get data in, but you can't get data out of it. You would then have to have one workflow call another workflow. So I do feel there is not that much logical completeness, um, in WeWeb. If you want to stay within this low-code visual ecosystem, you can get there with AI, but I think that is exactly why they have added AI because it would be a lot of work doing, you know, what Toddle has done and adding all of that in there. And it would also, quite frankly, be lots of explanation for the people who already like and use WeWeb as it is. And you don't want to scare off your existing customers.

Well, that's also a good point in terms of where the existing customers are.

Yeah.

So something else I have been wondering a bit about is There's a question like using AI to build these apps, right? And then there's sort of take it a step further, right? And what does— what, not just how are we building, but what are we building as we look ahead into 2025?

Yeah.

And the— and most of the low-code systems I've talked to, very unimaginative about this. They are— they're solving the thing of like, how can I help my customers today? Solve the problem that they have today. And there's something reasonable about that, serving your customers. But from a strategic point of view, like, it's sort of look across the market. You and I sort of have the privilege of being able to talk to, you know, between the two of us, hundreds of people, right, who are building and some of the changes in the way they're building. What are you seeing that is, you know, where's the puck going in terms of what's going to be getting built coming up in later in 2025?

I think most applications that will that are being built are mainly something that is an AI wrapper. If I take a look at my office hours, almost everybody is building some sort of an AI wrapper or some other kind of API wrapper. I think it has been always the trend. I mean, that's why StateChange is StateChange AI. I remember when I joined your original, I remember your original promo video. I think you were like sitting in your living room or something. You were talking about StateChange being a consulting firm or something like this. And you say, you know, why is it AI? Because everybody's now building something with AI. And I was like, I like that guy. I wanna join his community. And if you're thinking about joining Ray, sign up right now, it's $39. It's the best thing, best investment you can ever do. And if you want an even better investment, it's $149 and you'll get much more of Ray.

Or Ray. I actually find that the model that we have here, the people who I've seen working with you who are building, they're just able to go so much faster because they're able to bring things to your office hours.

You do the actual work.

Well, yeah. That's the key thing. The people who are building want to build. And they just don't want to be stuck. And that's so when people come into like, you know, no-code, low-code, you know, no-code, pro-code office hours. Yeah. Wow. I just can't say that out loud anyway. So when they come into your office hours, they are— what they want to do is keep running and they come in, they get their help, they get their tune-up and they're able to move on. I have seen other office hours.

We actually have lots of people in no-code, pro-code office hours and in your office hours.

Yes. People who are just, you know, maxing out, getting, you know, whatever help they, they can from those two sides.

It's way cheaper than hiring someone, and they prob— they're probably as productive, if not more.

And like, yeah, and, and what is it you need to hire for? But I really think there's a high and a low thing here, right? Like, you have— if you've got the, the AI that's there to sort of be your lackey, it'll take care of a lot of things. But if you give it instructions you don't really understand, it become— it will really go off the rails. You know, and that's really where things go sideways is when you're trying to give instructions you don't understand. So that means you come and you get educated. Okay, now you can actually use it to then go fast. I mean, just put those two things together. There's a whole lot that you can do. It's not like you need to be doing all this work, slugging away yourself. Like get just a little bit of help and a bit of tooling and the leverage is just enormous. I think in a way that much more so than I think was true even like 12 months ago.

Yeah. With Claude 3.7, it's actually a big game changer also when it comes to UI.

Yes.

I know what you're using it for.

Claude 3.5, the October version, was a huge game changer for being able to do coding because they finally loaded up with some relatively recent code, both on the Python and the JavaScript side. Then 3.7, it behaves a little differently, but I find it to be fantastic. For getting done things I need to get done, allows me to move very quickly. In fact, I've got on my screen here—

I never use ChatGPT for coding anymore since Claude. It's just not that good.

Sorry, what?

I never use ChatGPT for coding anymore. It's just not that good like Claude.

Oh, interesting. Have you tried using, you know, 0.4 or 0.4.1 for it?

I've tried all of them. It just doesn't get npm right. Because like what I found is Claude is really good when I tell it— my prompt for using something in Toddle is I have this proprietary framework. The code needs to be written exactly like this. You can't go off the rails. Just keep using it even if it doesn't make sense and follow my instructions. This is about my prompt for AI to write some code that I can use in Toddle. And then it's like, I have to run this code all over again. So you need to make sure that first it can run on multiple instances, components, and if it runs many times after each other, it can reinitialize the component. If I do it in ChatGPT, it's like, It just goes wild.

Yeah, sure.

With Cloud, it's really good at understanding the specifics of the code that you need to be generated because you're generating React code, so you can fly with both. But I'm generating total code, so I need some guardrails around what it actually does for me.

That makes sense. Just coming back around to the question of what is being built, right? We talk about how to build, like 3.7, okay, pretty useful, you know, over Sonnet from Anthropic. But what, I mean, what are you seeing getting built and what do you see sort of around the corner for getting built? I mean, you talked a little bit about like adding AI to apps. Okay, that's interesting. Like where is that, where do you think that puck is sliding to?

I do think most people are going to build some sort of AI wrappers. I think there's one in StateChange, he's building an AI Chrome extension, right? And we have lots of people in no-code, pro-code who are also building some sort of AI application. Everybody is using AI in some part of their product, even if it's not advertised as AI. That's what I feel. But in no-code, pro-code, we see actually lots of people who are building B2B applications. I think that's also what happens in state change. I don't think there's too much B2C, more is like B2B where people have a, specific way how to approach a problem. We have one, he's building like a social media scheduling app for businesses. So he's putting his expertise as a marketing expert, as an influencer, into building such an application. We have another one, he is building an application for— he's building applications for his clients who are— who need internal applications for marketing processes. So he's a marketer and he turns those ideas into products for his clients that they can use to then automate internal flows. They don't have to use ChatGPT. They have their in-house solution trained on their data. There are two ways how you can provide lots of value with AI. You can build AI systems for other people. You can build them for consumers and just sell it like a product. We have, I think it's equally, there is the B2C side of B2B and then there is the B2B side. B2B, if that makes sense. Where you do more of the enterprise engagement where you actually build custom software for the user and keep getting paid to keep it running for them and all that stuff. What about state change? What trends are you seeing in there?

I'm definitely seeing people building things that are beyond just web apps. I'm seeing people building extensions. I'm seeing people building and there's big demand for discussion of MCP, which is the Model Context Protocol for basically being able to integrate an app directly with an AI frontend. Right? So basically I've sort of seen bringing AI inside your app and then bringing your app to AI. And the first of those is relatively straightforward. That's the kind of thing you've built, right? For no-code, pro-code, right? You've got like the bot that's sitting inside your community. And that's a perfectly fine, you know, thing to do and creates a lot of value. But I think the other one is like where things— I really like what's weird, right? And what's going to be and what can also decrease your exposure to all this frontend work because as that starts getting pulled away, like where's the value now? And I kind of wonder if there's going to be like a significant market for people to be making MCP services. MCV servers for these various apps that exist so that people, instead of having to come into your site and use your site in order to get the value of your service, can just be using the same interface that they like, the ChatGPT interface or whatever, or Cursor.

I actually just built an app for as a custom GPT. I actually connected that already and works actually very well. I'd run with it. I built an API in Zeno. The nice thing is with ChatGPT is you can connect it with Zeno without having to wait till they release their MCP feature.

Yes, and also without having to go create some other kind of MCP middleware. I think the MCP spec itself—

I also built an MCP for Zeno's new npm, so you can end on MCP like a tool in ChatGPT. You can actually, it's called Xano Lambda and you can look it up. It's on the ChatGPT store, it's free. And if you download it, you can use it to just tell the AI what you wanna do. It'll find with web search the best npm packet to do it and it's going to create code that actually can run in Xano and you can just paste it. And sometimes you'll see error on line so-and-so. That's not the issue with the code they generate, it's 'cause Xano npm doesn't support every npm npm package yet.

It does not. This is true, but supports quite a lot.

Uh, we just, uh, in case anybody from Xano is watching this—

sorry, what?

It doesn't support the profanity filter in npm. We needed a use case in our office hours. We got sometimes weird office hours.

Um, okay, the, uh, but yeah, the, the, um, uh, Yes, their Lambda functionality is pretty powerful. We were just working in office hours for submitting a video and getting basically video understanding using Gemini through a Xano Lambda. Because the new Xano Lambda, we can download files, we can process them, we can ship them up to Gemini, we can wait things.

FastGen used to be?

Well, yeah, I mean, FastGen, I always thought FastGen's big competitive advantage was that they had these code steps, and they had them for Python and they had them for JavaScript. Xano went a different way. They, they said, we don't really know how to support that right now. And so they kind of stayed back. I thought that the code step was going to be a bigger deal for FastGen than it was. As it turns out, FastGen never really made that much in the way of sales, had much bigger opportunities elsewhere, and they just pivoted to something else. The Xano now has its code step in the form of the new Lambda. It's pretty darn powerful. It is JavaScript TypeScript because it uses the Deno engine as opposed to using actual Node.js, but it, it works. It works pretty well. I do think this is going to open up a new, new UI can of worms for Xano where people are going to be looking to be using Xano as a code editor, right? And Xano's features as a code editor are weak. But they have the opportunity to become good. They're certainly using the same underlying technology like VS Code and Cursor, whatever it does use, like Monaco editor. But there's some pieces to make it a kind of first-class coding environment that you're used to working in some of these other environments too.

I mean, we both have seen some things in the Xano Insider Discord that we can't talk about because there's an NDA, but I think they're going into the right direction with what they're they're doing. They're getting close with all the things that they're doing that we can't talk about because they're toxic.

They have said publicly there's going to be XAML script. They have said publicly they're going to—

I'll just let you say all the stuff so I'm off the hook.

They announced the MCP servers, the MCP support they announced early this month and their little kickoff that led to their first release, which was the text template editor., which I think was also sort of a big deal.

It was your idea, right? Like Daniel told me, I think it was your idea, or someone told me it was your—

sweet of them to give me, uh, credit, but let's be clear, I don't work at Zano. They do, and they, they built it, and it's, uh, it's a good piece of engineering and definitely, you know, feather in their cap. Um, and, and what I think is also neat about it is sort of the, the way I've seen such fast pickup on it, because there's big demand for people to be building. Like, I sort of was thinking about, like, why is this such a big deal in Zano? And yeah, sure, I made this pitch to them before. I'm going to suggest they should go down the road of maybe thinking about a feature like this. They— was that the Xano right now historically has been all about structured data, right? Which was great for building a REST endpoint. But the way that we work with AI in particular doesn't call for structured data. It calls for text and manipulated text. The irony is the technology that Xano uses underneath, which is called PHP, its first use case and its first-class support is for manipulating templated text. Like, that's what it does all day long. Someone who's used PHP back in like 2000 is like, wait, this is what it was for? And this is— and now it's sort of like coming home to mama. The, um, and now it is— that's what you're able to do in Xano, and people are doing that. And then— and I'm doing this too— and then like submitting it to one of these APIs, right? Like, hey, I've got my prompt, it's been set up, sort of complicated text. But it also lends itself to like other systems that might just want more natural language, which I think we're gonna see a lot more of in sort of the, you know, more AI-driven world, a shift from structured data to textual data. Yeah.

I also actually had a no-code tool company reach out to me and they just wanted to get my opinion. How sweet is that, right? They just wanted to get my on their tool, and I'm curious about your opinion on it. So let me share my screen here. Yeah, you do let me know if you can see it because I'm curious what you think about it. I think it's interesting. Do you see my screen?

No.

Now I can. Now you can. You can— everybody— oh yeah, I think everybody—

now you have infinite mirror going on, but now we're good.

Okay, okay, there you go. So I actually found this tool— they actually, actually, they found me, um, they reached out to me on the email that's on my YouTube It's a no-code web app builder for full-stack from front to backend. They're going after Bubble here where you can do your data, you got your dynamic data, you got your components, you got your logic. The logic is also frontend and backend. Then you got your design. You got your data, your APIs. They actually got server-side rendering, It's pretty interesting because they also have some AI agent features in it because everything is now AI. You can actually build AI agents in there, bring your own model, bring your own API key, and they have tool support and their AI can call APIs that you can set up. It goes beyond what Bubble can do, but of course, in being a full-stack tool, it's not right now there because they're very new, very, very new. But I think they're going into the right direction. Of course, they do have those pre-made integrations with Stripe, data protection, scalable team collaboration, which is actually pretty nice. I did play around with it and I actually had a great meeting with their CEO. I think we talked for like 2 hours and he was just very curious about what I think about it, my feedback, and like to listen to some improvements, how they can make it better, what I would love to see. It was great conversation. Now you looked at it a little bit. What do you think about those kinds of tools? They're like the full-stack tools, but you can of course also bring REST APIs into it.

They've been around for a couple of years.

They've been like, I think like 1 year, but they've been way bigger in China than they are in the US right now.

Right now.

So they're actually, I think, the biggest no-code tool or one of the biggest ones in China. And now they're kind of like going into the US.

Interesting. And, and I saw the DeepSync link on the very top, so that sort of furthers the, um, the Chinese connection there. Um, okay, well, that's, um, it's neat. They've been around a little bit for that. Um, but what do you think about, um, you know, what One thing I look at with those kinds of tools is like, it's a, you know, it's integrated. That means I'm not just buying code generator, right? I'm buying hosting platform, right? So it's—

the nice thing about them is they're actually React-based. They run on React, Vite, and TypeScript. So I think it's actually the first no-code tool I found that actually uses React.

Yeah, I find most no-code tools don't use React. That's a reasonable point.

I talked with their CEO. He said, you know, like React, you can't dismiss it. Missing the platform capability of React. He wanted to build on top of React so you can bring in React Flow very easily. You can bring in all those SDKs. You can bring in everything that the React ecosystem supports, which is quite neat. They do actually have a feature where you can open it in VS Code for your custom coded components. You can edit them and in real time you can see it in the dashboard.

Well, that's pretty useful.

Connect with localhost. That's actually quite cool.

It's a little bit like Noah for Flutter. Noah has that and actually the flow for building coded components in WeWeb looks like that historically as well before they had more of the AI-first way of doing it. That's cool. They're out there. What market penetration they have? Because they really aren't showing very much. Their wall of love is a few indies, but they're very big in China. I mean, like, what's, what's really the scale of the organization?

I'm not completely sure about that, but I think they're a team— like, they told me like 15 engineers, so there is some size to them. Um, but I think it's just very, it's just very early for them. Um, but I, I think they showed me what they are planning to add, and it looks very promising. I don't think it is something that I would like build a very big application on right now. I think it's a very good Bubble competitor. I wouldn't build a very big application Bubble right now. I think right now it's a very valid solution to Bubble if you're considering Bubble. Um, but if you're considering something like WeWeb or a Toddle, I think just wait a year and I think they'll be there.

Interesting. Yeah, the, um the, the, and, and be there for what? What do you imagine is the use case?

I do think it will be very nice for applications where you want to have the backend and the frontend in one place. They are really going on accelerating the development for AI wrappers because their technology for building AI agents that connect with the frontend is novel. So if I take a look at Jerry, who's building kind of like AI agent internal interfaces, for his clients, I think he'll definitely be able to make good use of something like Momen to just build those way faster because it's most of the time it's just table stakes. They have those AI agentic automations with tools, something like MCP already set up and you can just have the front end and the back end in one place. That is quite unique and for what it does is actually quite cheap. It just lets you spin up those things quite fast. Actually, we have a comment about this. Is it possible to export the code with this tool? I talked with their CEO. Right now it's not possible, but you can bring your React code into it so you can do it the other way around. But I think it's an interesting tool. I may make some videos about it once there are a little bit more features that kind of like touch the hardest parts because what we do on this YouTube channel here, and me and Ray both, we kind of like more touch the harder parts parts of visual development.

Yeah, yeah, that's cool. Uh, and, and like the, the fact that it has that, that bridge out, would love to see, um, you know, we'd love to see a demo of that, you know, like get a, get, give yourself what's going on with it.

Yeah, I think you should definitely try it. I think it's very interesting. It's very interesting tool. Any new interesting tool that you found that someone approached you with or that you discovered somewhere on the internet? Anything cool on your side?

The thing I've looked at most recently that I found to be fun is Unicorn Studio. Have you used that one yet? No, I haven't. Unicorn Studio is a visual development tool for basically making Three.js elements and special effects and artifacts. I can bring it up here. Yeah.

You know how to share your screen? You need my help with that or? Looking okay? I think it's coming through. Yes. Right now I think at your— I just wanted to say right to Spotify, but it doesn't know. I think that's Unicorn Studio.

This is Unicorn Studio. The idea here is that you have— there are things other people have built, but the vision behind it is that it is a tool that allows you to make these special effects. Creating something that might look like this based on the—

How expensive is it running this on your website?

Then yeah, and they basically create artifacts that you can embed as code, if you go buy the thing, or as embeds to basically be able to put these special effects on your site and then be able to control the way that they look. You can decide what the core image is, what the—

[Speaker] Kill your SEO, right?

[Speaker] Well, the question is, your SEO isn't driven by this element.

Of course, it is like your load times, but—

right, so like, so there's a little bit of SEO art of making sure it's saying loads after your initial load so it doesn't count against you for having a slow site. Um, but it becomes the kind of thing that you can also, you know, tie into, you know, interactivity. Um, you know, in terms of being able to see what, um, you know, how things behave. You sort of see the change in like the background here.

If you can add that into Taro, I think it's the ultimate, it's the ultimate Webflow killer.

Well, the neat thing is, is just about doing the special effects, right?

Not, not Craft.

And here you can also see how it's not strictly done through a, it's loading in the JavaScript, right? So it's coming directly. Yeah, yeah. Then you can also export it as like an MP4 video or something.

Do you want to integrate this in something? Show us how to add it.

You want to do a demo right now? Yeah.

If you want to send me the code, I can show you how to integrate into Trello.

Yeah, you should totally spin this thing up. But the thing I find really neat about this system, is how it is using these visuals for getting outside of just—

I'm not on the waitlist.

Oh yeah, you go sign up for the waitlist, it clears pretty fast. But the creating things that aren't just standard web apps, that's what I find to be most exciting here. It's why I find MCP to be pretty exciting. Like a frontend, it's going to be exciting.

You want to send me the code for the debug real quick? I think it would be interesting if we can see if we can integrate this in toddle.

Northcraft.

Do we have something with a pine tree or something on there?

Let me just see here if I can get this component. Oh, it actually has a Framer component. You want to put it in Framer.js?

Let's do it in toddle. Let's see if it can work in toddle. I think I just need the JavaScript one. Yeah, sure.

Here you go. I'm going to put into our private chat. We're going to see if this works. Private chat.

Go. Private chat. I did not know about that. Where the heck is the private chat? There you go. I got it. Let's copy that. Let me start sharing my screen. This will be interesting. We're doing interesting stuff here. Let's share that. Do you see my screen?

Yeah, sure.

There you go.

I've got a mirror. There we go.

There. Now we are at— look at that new logo.

Nord.

It's still on top of that, but it's, uh, not Nord Stream, Nord— uh, what was it? Nordkraft. Okay, uh, let's do this animation. We'll create this as a component. And we're just going to paste it in here. We're going to import. Did it import it? Oh yeah, there you go. It did import the product ID. Let's actually try going to HTML online viewer because I think we got to structure it a little bit. Okay, I got it in here now. So let's do format and we got this script. Text JavaScript will have a function. Do we need the exclamation point in the function?

Ray? I'm trying to look a little more closely at it. Does it have an exclamation point? Bang.

Do we need that?

Not clear to me that the exclamation point really would matter in there, but I don't know.

Let's try it. Add a custom action here. Let's actually first beautify the code. Let's go to JS Formatter. There you go. I still don't know the keyboard shortcuts for the Monaco Code Editor. Let's go to Nordcraft. Getting used to it. Animation. Now I can create this. We can do a little package here. Then I have the source Unicorn Studio.

I think that's it. Clearly it's creating a script tag which is shoving at the bottom. Then I'm pretty sure that it wants to have a div somewhere on there that has this data.

We got the JavaScript, so we got to go into the component and then unload of the component. Run the JavaScript. Now we got the div in here. If I reload it, look at that. We got a fancy animation. Look at this, I can just do, isn't that cool?

You can replace that lightsaber or whatever it is with what we're looking for. That background image is just the image that gets swapped out. That can be like a screenshot or something. There's lots of fun, silly things you can do with this.

Yeah, we just do width 100%. And then we'll just reload the page. And, and then I have my— and look at this animated background and toggle with like, see how it just moves.

And Nordcraft. There you go. Animated Nordcraft website, you know, just by integrating this other other little doodad.

There you go. Now I got it as a component so I can just add it. I think I got to reload the page and then I got it here. Look at that. Isn't that cool? I can add text over it and I can add a div over here and make it absolute and make it with 100%. We're going to make a little— let's do a new landing page for StateChange. Let's do in here, we'll add an H1 and I'm going to do Join StateChange.ai. I'm going to center it. I'm going to give it white background. There you go. Now if I reload the page, we got a very fancy state change landing page. Look at that. There you go. Look at that.

But that's what I look for. I actually have you know, have this idea like weird no-code. I look for, you know, what are, what are tools that allow me to do things that were otherwise hard or strange to do? And where do the AIs, um, allow me to do things that are like more than just table stakes?

Someone just said wow. Yeah, that's cool. We just made like a little— you got a new landing page. If you want that, let me know. I'll share that project with you.

Yeah, it's, uh, and you can actually export the code, which will make it run even faster and blah, blah, blah. Yeah, but that's, that's kind of weird thing I look for. So, so, uh, if anybody on here who's watching this, uh, says, said, oh wait, I know something really weird no-code tool, uh, please put that into the comments.

Uh, we would love to see it. We, we should do like, uh, next, next time we're going to just talk about weird no-code tools.

I'm down with that.

Get a list together. I'll get a list together, and we're, we're trying to on camera set the weirdest things up and mix and mash them together.

I like that. I like that. We've got ourselves a plan for our next live stream. Same, uh, you know, on the right here on this channel, wherever it is you're seeing it. Uh, of course, I appreciate it's going to be on the No Code Proko channel. It should also right now be on the StateChange channel, even the StateChange LinkedIn right now. That's why we got 5 people watching.

Look at that. Wonderful. Also take a look at statechange.ai on YouTube and subscribe to Ray's channel. It's definitely worth checking out. Also take a look, here's his office hours schedule. We got that here. You can scan the QR code and take a look at that. You can also join his community at statechange.ai to get access to daily office hours, mental models, and tools to solve your hardest 5% on low-code, no-code/vibecoding.

Yes, vibecoding. We need to do a survey, get people to give us a better word than vibecoding. Accelerated development. Accelerated development.

And if you are more concerned about the front end and how to make your buttons look nice, you can also join no-code-proko.com. And if you want a beautiful t-shirt like this, Smile, Ship, Repeat, my, I don't know, slogan for this YouTube channel, it's somewhere down this video where you can, uh, purchase that t-shirt. I think it's only $20 and it's beautiful. I think it's cotton or something. It's a beautiful t-shirt with a beautiful message. I feel like we're doing like live shopping or something here. Oh, I think we lost Ray. Or did we lose me? No, I think we lost Ray. Let me Let me DM him. Let's see. Oh no, on YouTube we see I'm moving, Ray's not moving. Let me just text him, we lost you. There you go. Oh yeah, now we lost him completely. Oh, okay, let's wait. He comes back in 2 minutes. If not, this is going to be a very awkward ending today. Does anybody have any questions in the comments that we can help with before we bring this to an ending? Let's see if there's something in the chat. Let's just activate the chat here on the side and let's go over it. We got Ray. Ray's coming. I was just improvising because you were gone. I was starting to read some of the comments.

Well, apparently we lost our connection and I apologize for that little technical difficulty there.

Perfect. I just ended it as I always do with some advertisement. But now I can put it again away. But also, when— oh, we got one comment. Actually, that comment may turn into advertisement.

Um, yeah, let's talk about Despia before we, before we go. Tell us what was coming down the road with Despia.

We're actually right now, we got the whole version 3 dashboard done. We got the whole version 3 framework done for iOS and Android, so we're done with that. One thing that we are not done yet is the automated distributing for Android, which has been right now in Despia a painful process because you actually had to export the code, open it on Android Studio, publish it yourself, sign it, manage your keystore. So what we're done with is that we manage the keystore for you. You can export it anytime, you can bring in your own keystore. We'll build that app for you on our M2 machines for you. We'll send it to the App Store. One thing that's not done is actually, um, we need to replace the name of the app because right now if you use it with our with our automated publishing for Android. That's the only thing that's, that's stopping us from shipping it out there. Your app will be called Despia Framework, so we need to adjust our CodeMagic YAML file to actually replace that name with your app name. That's the last thing. It's always those small things that hold you back, but I'm very confident that we have good news next week. But I'm not going to promise anything.

Um, it is what—

what the only thing missing is Replacing a name of an app name shouldn't be a big thing. It's actually quite a simple thing, but I can't make any promises. I don't want to get your hopes up, but the framework is done and we are definitely going to have v3 for iOS next week with widgets and app clips and all that exciting stuff and reading user's clipboard. And I think we lost Ray again, but I hope that answers your question. And if anybody wants to try the V3 editor with a little bit of bugs. And if you're not necessarily wanting to publish a good Android app right now, because I wouldn't suggest using the V3 pipeline to publish the Android app, I would use V2 because it just works way smoother and there has been more testing to it and it's just more reliable. But if you want to still test it out, definitely, you know, put a comment in there or, you know, contact the support and we will be able to set that beta up for you because we definitely need some people who are willing to test it before we'll release it. But thank you, Johnny, for that question. And we lost Ray again. Let's see if he comes again. In the meantime, since we've been talking about Despia, you can also scan this QR code here to— I feel like we're doing something like TikTok live shopping or something. You can also scan this QR code here and take a look at Despia and start your project for free and get that all set up. There you go. I'll be waiting for Ray. Let's see if we can get him back. We got Ray back. There you go. We got you back again. We got to hurry up in ending this now.

We have to wrap this up because I live in a rural part of the world and sometimes it is the connection isn't totally perfect, but very rarely do I have this problem. I certainly apologize to you, Yacine, everyone else is watching.

No worries. We got one last question. Let's do this quick. Would you say no-code, pro-code is beginner-friendly? Ray, what's your impression on that?

I think your community is much more beginner-friendly than mine is. StateChange is deliberately focused on this question of hardest 5%. I would say people who are just getting started with their journey, there are lots of other resources, including Yacine's, that are much more about helping you climb those earlier parts of the learning curve. Where state change is at strongest is gonna be where, you know, hey, you, you, you've gotten to a certain level and now like you're at that point where like issues are really weird. Right. And that's, that, that's the kind of thing that I enjoy the most. We, we sometimes joke about this 5% versus 10%. But I would, I would say that if you are, if you are, if you are serious about building, right, but you're a relatively beginner, you would, that would still be a good fit for your scene. If you are serious and you have come along and it's like you're using other resources and those resources aren't helping you as much as like you really are in that last 5% territory, that's really when I would recommend coming to StateChange. We have lots of people who have been members of ours for a couple of years, but tends to be after they have gone through that first part of their journey. Yeah.

We actually have lots of members who started out a year ago with NoCodeProCode and who recently canceled their NoCodeProCode subscription and moved to StateChange because they're just evolving in the process. They are just looking for, okay, now I know how to make buttons work. I know how to do API calls. Now I don't want to use Xano anymore. I want to use Supabase. I want to go beyond that. I don't want to use Tarnlet anymore. I want to use React. Then Ray's your guy.

Cool. All right. Let's bring in—

those things than I am, but I think a no-code pro-code is at its strength. If you are within the TOTL ecosystem, if you're within the no-code ecosystem, you try to stay within that ecosystem and you either need help getting started in that ecosystem or you need getting help, you know, keeping up your momentum within that ecosystem. Once you're trying to look outside of that ecosystem, We got to have Mr. React here.

Mr. React. There's also a little bit of a frontend backend thing too, where you are a very talented friend. You know your way around the backend, but I think really the weird backend stuff.

I learned it all from you. I built a pipeline in Xano to put 2,000 files together to build an iOS app. I couldn't have done it without being 2 years a member in StateChange.

Very cool. All right, well, on that very kind note, let's bring this in for a landing before I lose my connection again.

Yeah, you have a motivational— let's do it like this. You'll do a motivational outro and I'm going to put up your advertisement here.

Look at that. Oh yeah, well, let me make it full screen so it's even more motivational.

Look at that.

There you go. Yeah, uh, there, there's so much new software, uh, that I, I think we're going to be building in the course of this coming, you know, year. And, you know, I'm really looking forward to next week because it's going to be, uh, weird stuff. Um, and the, um, and, and I think new weird tools that are about how we create real extraordinary value is going to be the just, just what's next. Um, and I can't wait to continue doing these Bidcasts with you, Yacine, and learning about how people who we've seen and people, what we see in the market are helping us make those, what's going to be next for software.

I think that's a wonderful way to end it. Thank you so much for your time, Ray. I really appreciate that. And we'll see each other and all of you next week.

Thank you.