Loading summary
Transcript1 lines
- [00:00]
A
Today on the podcast, Lovable has hit a $500 million annual recurring revenue and their users are building 1 million projects a week. Apple has announced a ton of new features and finally rolled out the AI Apple Redone Siri we've been waiting for. In addition, Apple is turning shortcuts and Safari extensions into natural language Vibe Coding, which is really interesting seeing Apple getting into the space. Perplexity is looking to do a IPO in 2028. That's their target and, and they say it doesn't matter on, you know, regardless of whatever's happening with Anthropic or OpenAI and all the other listings, 2028 is when perplexity is going to IPO. Sam Altman's tools for Humanity, his company has cut staff at the same time as OpenAI is planning to IPO. So OpenAI is growing and that one is shrinking. Google's DeepMind is picking 15 European robotic startups for AI, their AI accelerator program. I also want to tell you about my recent experience. Vibe coding an iOS app getting approved on the App Store and kind of the way the whole market has shifted as this is something I've been focus focused on and working on for seven years. So the place we're at today I think has a lot of implications for what we'll see in the future. Out of the gate, we have to talk about Lovable, who is on an absolute tear right now. This is a Vibe coding platform. If you haven't used it, you go there, you explain a tool you want to build and it can create it for you. I've built dozens of projects on Lovable, some serious ones, some less serious ones, and ultimately, at the end of the day, I've actually had most of the big ones I use ported off of Lovable, and I'm hosting them myself and, and editing them myself now just because Lovable got too expensive. I was paying, you know, 600, $800 a month for the Lovable credits, where something like Claude Cowork is subsidized. But Lovable is on an absolute tear right now. And if you're not a developer, this is the number one way I recommend people go and actually build apps. Anything base 44replit, these are all other great competitors, but Lovable is the one I like the best because they do a really good job with the database integrations and a lot of the AI integrations and security audits and they and SEO audits and they do a lot of really awesome stuff built in. And because all of that is happening, they just crossed $500 million in annualized revenue. That is a hundred million dollar jump. Since February, they're now seeing, like I mentioned, a million new projects launched every single week. I think there's a lot of people that are using this. Founders, designers, salespeople. You just build, you know, internal CRMs or HR platforms, basically stuff you would have paid for traditional SaaS, but maybe it didn't have all the features you want. I'm gonna be honest. As far as Vibe coding goes, I have a bunch of tools that I use. Including, by the way, as I'm. As I'm doing this entire podcast episode, I Vibe coded a platform that helps me come up with all of the scripts and all of the stories. It does all of the research and it puts it all together for me and I can kind of go through and talk about and look at all of the points about what's happening in the news. So the thing that I love about that though is every time I see like, I'm like, oh shoot, it's annoying that this happens or that happens. You, you don't submit a customer support request, you don't go watch a YouTube video to figure out how to do something or if it's possible. I just go to Claude on the side and I'm like, hey, add a button on the left side of my page here that when I click it, it refreshes all of the, you know, like all of the article descriptions and does XYZ thing and gives a couple more in detailed pieces of information. Cause I need them for, you know, the podcast. I'm doing like whatever I'm doing, I just tell Claude and it instantly adds it and then I'm off to the races and I move on with my life. It is so, so cool. And I think people are really discovering that power because you can do all of that with Lovable. So users have created over 50 million projects on Lovable since it was. It started actually in 2023 and it's not even three years old, but this thing is on an absolute tear. In August of 2024, they said that they were going to hit $1 billion in ARR within 12 months. So they didn't actually do that because they're currently at $500 million run rate. But I mean, their growth curve is incredible. And you know, their, their timeline was pretty aggressive back then. They didn't actually hit that, but I think we got to give them some flowers for just making such an insane amount of money so quickly, growing so fast. Survey data shows that non Technical users are building production grade CRMs, inventory systems and HR platforms. Categories that were much, I think commonly kind of locked behind multi tier SAS contracts. This is what I'm seeing a lot of people talking about. Basically. Okay, Apple is turning shortcuts and Safari extensions into natural language vibe coding. This is really interesting. They had their big wwdc, this is their big developer conference. Their new Apple iPad OS 26 is going to let users create shortcuts so you can type into plain English, you know, send a text to Anna with these emojis and all of a sudden it's going to become a working automation. I'm just kind of laughing because, you know, you could imagine as you're leaving your house every day, your iPhone could automatically be sending a message to your wife saying that you love her and you just have an automation doing that. I don't know if that's, I don't know how ethical that is, but you can think of a lot of very useful use cases, right? Maybe you're leaving work at certain time and it can go and send like a clock out request or something right as you're leaving or you're, you're going to an event or you're leaving an event and it's something that you do on a, on a regular recurring basis and you kind of have these texts go out. Maybe you have like a meeting and you send a text like, hey, great meeting, send me the follow up details. Just a reminder, right? So I think this could be pretty great. It's definitely a very genuinely novel AI feature that they were showing because it's not really just a chatbot or kind of like a catch up feature. It's a way to actually program your phone. And you're doing all of this by the way, without writing any code. So you're not like, you're not having to like make an app, you're just saying like, hey, build this kind of automation and your phone knows and does it. I think that's cool and I would like to see a lot more of that coming out of Apple. We also have a bunch of new information on the new Siri that Apple has finally released. I think they had to pay $250 million because they had all these flashy demos two years ago and nothing that they showed off actually came to be. So now if you go look at the wwdc, you'll notice there's not a lot of flashy demo videos. They actually are just literally showing you using it. Screen recording. I think they need to do that. To build people's trust back up. But you know, they had to pay $250 million in legal fees for false advertising or whatever. So I think they're trying to be more careful. But with the new Siri, there's a bunch of features I'm excited about. Number one, Siri is getting its very own standalone app, which is awesome. Super stoked about that. It's going to have chat history on it and it's going to let you choose what AI model you want to run. So it's kind of cool because you could switch, you know, Gemini and you're talking to it and it would have all your chat history. Then you go and switch it to, you know, chatgpt maybe to power your Siri. What I love about that is just the multi chat company kind of ability. This is something that people love my company AI box AI for because we let you chat with 80 different AI models in one place. And I think people want the flexibility even. Even myself, right. Like I use ChatGPT for image generation. It's awes Claude doesn't have that. So then I got to go have a Claude subscription, ChatGPT subscription and then, you know, chat and then like OpenAI now they're deleting Sora. They're not gonna have video anymore. So you gotta have Google VO3 for the video generation. And 11 Labs is the best for audio. So you need a whole bunch of different audio models. And of course, you know, Siri is only gonna let you pick, I think like three or four different text based AI models that you can use to power Siri. And so if you want to get all of the, you know, image, audio, text, video, all that kind of stuff, you'd need a different platform. Siri probably won't be able to do all that. So you'd probably still need to get keep using AI Box. But I do like that they're giving you the choice now. In the future, I could definitely see Apple kind of pulling the same trick they played with Google Maps, where that was kind of the default on the iPhone and they switched it to Apple Maps once they came out with Apple Maps. I could see Apple coming up with their own LLM and adding it to the mix and then maybe even if they got really cutthroat, cutting off all of the other options into Siri. So it's just Apple Intelligence powering Siri at some point. So I think this is the direction they're going. But right now they're definitely not on ahead of the curve on this. If You've seen the Apple Image Playground thing, they kind of revamped it, but it's pretty horrible in my opinion. So if their chat was anything like their Image Playground, I would definitely not want to use that to power any sort of AI feature. So I'm happy that they're giving you the option to use Claude and Gemini and ChatGPT right now and hopefully that is something that they'll keep in the future. Perplexity says that they are targeting a 2028 IPO. They say it doesn't really matter what's happening with Anthropic or OpenAI. That's what their goal is, which is interesting. They're going to do this even if OpenAI or Anthropic fail or succeed, they don't really care. But we also have news that OpenAI filed on Monday, yesterday that they're moving forward with their IPO. Anthropic already filed for a trillion dollar valuation IPO last week and, and SpaceX has already filed as well. So we have these three mega cap AI companies all going to be debuting here very quickly in basically the same window. The CEO of Perplexity tied Frontier Lab valuations to a hard metric. He said if either anthropic or OpenAI ships are no meaningful model capability advancements for six months, then investor confidence is going to crack. So I mean, really, he's saying this is what we should be looking at and I think this is very fair. We know that they're, they're working on training some big models and adding some new things. So the question is, you know, when will these new features come out? Are the, how, you know, what kind of advancements are we going to see? Because a lot of the stuff we've been seeing as of late is adding new features or adding computer use to the underlying models that were already there, not necessarily making the underlying line models better. So he says that's a metric we're going to want to look at. And of course Perplexi is a model, is a company that uses those other models and it kind of embeds them into their own, in their own software. So he definitely, you know, selfishly wants to see those models get better so that his product gets better as well. And according to a statement by Perplexity, they actually route different queries to open source models if those open source models have about a 90% reliability. So using something like Perplexity and you ask it a question and it's going to go send it to an open source model if it's easy Question it's. And it's going to send it to Claude. If it's very complex. They said that they can actually cut costs 10 to 20% versus just going, you know, straight with Claude. Sean's startup, Tools for Humanity, it's kind of the eye scanning identity startup. They're laying off staff as regulatory bans in Kenya and South Korea are basically cutting off their core growth markets. The cuts are coming the same day that OpenAI is filing confidentially for an IPO, which basically leaves Sam Altman. He's got these two companies and they are going in very opposite directions at the moment. World, which is the iris scanning orb, it's kind. Kind of like a subsidiary of Tools for Humanity. They secured a $2.5 billion valuation from A16Z and Bain Capital. A while back, I covered one of their integrations, but they have a few. They're working with Tinder, Zoom and DocuSign to basically verify that you're a real human and to verify your identities. They're kind of doing this identity verification, something like Tinder, right? You want to make sure that you're talking to real people. Zoom to make sure, you know, you're not talking to an AI avatar, which, by the way, I've had people apply to work at AI Box, my own startup, and they were most definitely just using, like, an AI to do the whole job interview, which was crazy. And of course, DocuSign, you know, making sure that real people are signing documents. So there's like some good integrations. Kenya, though, recently, they just banned World entirely. They, you know, they cited privacy concerns. South Korea fined them $830,000 for privacy violations, and they shut off kind of two of their big emerging markets where they. The unit economics depended on really cheap enrollment. I think we heard, you know, tons of people in Kenya were signing up. And the reason why is not a big shocker. Sam Altman was kind of just basically promising to give you free tokens of his cryptocurrency that was attached to World if you went and signed up. So, I mean, basically people in Kenya were just getting paid to sign up. And the unit economics made sense where you could do it pretty cheap. Tools for Humanity paid users in Kenya, India and Hong Kong about $50 in Worldcoin for every time that you scanned your iris. So, you know, if you want 50 bucks, you can go and scan your iris with their thing. This drew a lot of regulatory scrutiny, and there was a bunch of legal action that happened from that. I'm sure they collected a bunch of data from that I'm not sure exactly where they'll be selling all of these iris scans and all that kind of data from or how their identity verification will completely work there. But I think unless these some of the US enterprise deals like Tinder and Zoom and DocuSign are going to drive a ton of value, it's going to be hard to sustain this company with, you know, kind of where it's at today. Google DeepMind is currently, they just launched a three month accelerator in London and they're going to pick 15 European robotics startups. They're going to give them access to Gemini robotics models and some hand on mentorship from Google DeepMind engineers. The cohort is going across 10 different countries and they're going to tackle real world AI challenges. They're doing things like surgical micro robots, they're going to do ocean autonomy, they're going to do construction automation, a bunch of really cool areas. Norway's 3D Components AS claims that their AI welding platform delivers results 280 times faster than a traditional trial and error inspection method. There's a bunch of different startup clusters across healthcare. They're doing some brain tissue micro robotics, there's some construction stuff doing on site robotics factories. A couple other interesting areas are humanoid development and industrial automation. They have some like automotive automation, they have some waste sorting robots. So a bunch of really cool stuff are happening. I think there's a bunch of different companies that have gotten into this, this from the uk, France, Germany, Switzerland, Italy, Norway, Sweden, Denmark, Greece and Romania, which is basically Europe's densest robotic research and manufacturing hub. To me this whole program is very interesting but it's DeepMind really signaling where they think the next AI frontier is going to be which is these kind of embodied systems that are going to be able to look and act in physical spaces. Not just you know, chatting with this on chat GPT or, or on you know, Gemini, but really actually getting into the center of manufacturing and a lot of other things that robot only robots can do. I also think this is going to be kind of the feeder for a lot of the partnerships that Google DeepMind is going to be making in the future. Right. They're going to go help these companies and these startups that are working in these different spaces. But this is going to be basically their client list and how they're kind of developing a lot of the AI tooling that is going to go into other robotics companies. So this kind of the, the incubator, this is a lot of value for the companies joining the Incubation for sure. But also this is a lot of value for DeepMind itself. Over the last couple of weeks, my wife and I have had a fascinating experience revibe coding an app that we made years ago. And I want to tell you a little bit about what, what happened, what we saw, and some of the improvements that we're able to make and where I think this whole Vibe coding thing will go into the future. The first thing is I 100% think anyone will be able to Vibe code apps for iOS for Android. But our experience was, you know, in the past, the very first startup I ever started when I got out of college was with my wife. We started, it was called Self Paws, and it was a meditation and positive affirmation app. So you could record your own voice and you could kind of make these custom meditation sounds in the background and listen to yourself. Same positive affirmation. So it's kind of like if you're sick of, you know, just listening to something like calm or headspace, where it's someone else's voice, you could kind of do it in your own voice. That was the idea behind it. And it became very popular. We spent about $200,000 on this to get it developed and built and worked on it over a number of years. And since I've been working on AI Box for the last three years and this, this company didn't really. We haven't been able to really maintain it. It was too expensive for the revenue was able to generate. And so we weren't able to really keep maintaining and making a ton of updates to the platform. Now, two weeks ago, after, you know, I've been talking about Vibe coding and building so many different things, my wife said, hey, look, like we should probably go relook at AI Box and work on, you know, like, creating a new version of the app, get rid of all the bugs, add all the features we want. Like, is it possible? And so I started looking into it, and in the space of two weeks, my wife is pregnant, so she was able to, you know, she had a lot of time sitting on the couch. She didn't feel like moving, but this was a lot of fun for her. She was able to rebuild the entire app from scratch, iOS native. So she. She built it all on Swift. And it was really incredible, all of the different things we could do today that we weren't able to do in the start. Just so many different levels. One of them was like, all of the. All the icons and graphics that were on the app in the past, you know, like we hired people to go hand draw every single. You know, there's like 120 different sounds that you can layer in the background of meditation, like seagulls or waves or stuff. We had someone go hand draw every single one of those graphics that were added onto the app so that they all kind of looked the same. Well, today we went and took those same hand drawn graphics and we were able to feed them all into ChatGPT and have it make even more cohesive. Even more, honestly, is like, it improved the quality a little bit. It kind of started with what we had before, but it smoothed out some of the rough edges or some of the graphics that were a little bit too wide or too tall, that didn't quite match perfectly, or that just kind of looked a little bit funny. We put them all in and were able to get ChatGPT to redo all those. That was like a huge project. That was really fascinating to see how that had changed. We also had some bigger kind of animations and graphics that were like much more elaborate pictures. And we were able to upload those into ChatGPT and say, hey, because we, we decided to like, reskin the whole app. It was kind of like more blues and gradients in the past, and now we're gonna make it more like browns and earthy tones. So we actually got to take all of our graphics and regenerate them as earthy tones. So it's very similar to the old graphics. Not exactly the same, but now it's like earthy tones. And anyways, it's so fascinating how we're able to do that. We also, probably at some point in the future, although we haven't done this right now, we contemplated, you know, we had all these people that we paid in the past to generate or to go and record themselves doing a lot of the calming meditations. I would be curious to experiment, you know, going and adding like 10 or 15 more voices, so you could pick more voices to read your meditations. But we'd use something like 11 labs to, to actually do the voiceovers for a lot of our meditation tracks. So I know, like, that's probably like controversial to some people, but if the quality was good and people, you know, really reson with a different voice, I think I'd like to give them the ability to do that. But in any case, it's been so fascinating and we were able to get the entire app rebuilt. All of the payment integrations, all of the features that we, you know, were always buggy in the past. Like, you know, the Thing that people still use it to this day. I think there's like 300,000 or maybe like 150,000 streams a month on people listening to their own self recorded affirmations on the app. And so it's really cool because one of the big bugs we had in the past always was people weren't able to like they wanted to listen to it all night long in like a loop basically. So they fall asleep to like their meditation. And a lot of times in the old app it would just kind of glitch and it would crash somewhere after like an hour or two. Well, that's because it wasn't built native to iOS with Swift. And so when we rebuilt the whole thing like that big bug was gone. And it's just incredible. Anytime anyone has a recommendation or a suggestion or a bug and that they want to get rid of, we just tell Claude and it opens Xcode on the computer and it adds it and rebuilds it and it's like, okay, all ready to go. So this is shout out to all of you. If you've ever wanted to build a tool or build a product, you 100% can do it. If you have a vision or an idea, go and build the thing. You 100% can. My pregnant wife in the last two weeks rebuilt our entire, you know, app that we spent hundreds of thousands of dollars in the past. And also I went and built a web app which we never had before and we always wanted so you could just go to selfpaws.com and you get all the same features that the app has. But on the website, which is also nice for like account management. People can create accounts, they can pay us so we don't have to pay for like the just all of the iOS fees because they charge you like 15 or 30% of all of your revenue. You got to pay to Apple if someone buys it on the app. So we can tell people to go sign up on our website and then download the app after. Anyways, it's an absolutely, you know, just unlocks so much for this old business we had being able to completely re envision it, rebuild it and send it out to the world. So if you want to build something, my number one recommendation, if you want tips or advice on it. I have a school community which I'll link in the Description. There's like 300 members. It costs $19 a month. But I break down this week. I actually did a whole video about how you can build apps and I actually showed some of the usage metrics, the analytics how much money this app in particular is making, but I make a video like this every single week. So if you're interested in vibe coding and building things, go join the school community. There's a link in the description. And if you want to see what this app looks like that I've just been describing, go check out the Self Pause app. It's been completely. It also has like an AI chat in it where you can. It's like a life coach. An AI life coach that can help you to improve in different areas of your life. But even if you're not interested in like using the app and in affirmations, go check out this app if you're interested to see what you can actually build. It's called Self Pause and I'll leave a link to this in the description as well. Guys, that's it for the episode today. Thank you so much for tuning in. I hope you enjoyed this episode. You learned something new. It's obviously a labor of love and something I'm very passionate about doing and sharing in the future. Coming up, I am due for My wife is having a baby. Our due date is later this week and so I'm not sure how good I'm going to be at being very consistent for the next couple of weeks. So if I'm not, I apologize ahead of time. Maybe when the baby's napping, I'll find time to keep doing these podcasts, but if I disappear for two weeks, you'll know that the baby was probably born. So anyways, thanks so much for tuning in, guys. I really appreciate all of you listening. It helps me be able to do this show, support my family, and honestly, really appreciative to all of you. Hope you guys all have an incredible rest of your day.