How to Start a Startup Per Day (with AI)

So we have talked a lot about AI on this channel and even built some pretty cool products and services. And rightfully so, AI is fundamentally changing how people and businesses work and leverage their time and ideas. And what stands out to me there is that AI is a leverage multiplier. It just lets you do more faster. That also means that it's possible to start a business in a day. And that doesn't necessarily mean you should, but what is critical here on a more fundamental level is that you can now iterate and improve much faster than previously possible. And like we talk about all the time on this channel, being able to test, fail fast, and improve rapidly is critical to making any idea work, but also just settling upon an idea that actually could work. And that's whether you want to start a business on your own or you work in a multi-billion dollar company. It's the same idea where if you want to grow and get something new to work, you need to fundamentally improve and iterate. And AI lets you do that at a previously impossible scale. So in the next few minutes, I am going to break down a full AI cycle that lets you, in one workday, build and test in front of real customers a business idea. Again, you can do all of this in less than a workday. This is something I do all the time just to test ideas. And worst case, if it doesn't work, you just go back and improve. You learn something and you try again. And trust me, even if you're suspicious about this or you understand that it takes more time than a day, obviously, to really make something work, I think you'll watch this and you'll recognize the value of something like this, which is essentially a bundled, highly leveraged way of just testing something in front of a market and just figuring out, does it not work? Does it work? And how can I make it better? Now, this approach breaks into three parts, creation, leads, and outreach. Creation is just building the thing. Leads is figuring out who you are going to get in contact with. And outreach is actually getting in touch with them. And we're gonna do it all in a few hours. So to validate a startup or business idea, you, at the very least, need a full MVP or a working product or, at the very least, a sales or landing page. If you're selling a service instead of a product, just skip this part entirely. Or focus on just setting up a good-looking landing page so that your service, your business, looks legitimate. Now, to focus on actually building the thing as the first part of this one-day startup cycle, here is the workflow I recommend. So some combination of Repl.it, Lovebot, Cursor, and Clod to actually build it or code it, then Streamlit or GitHub for deployment, Stripe for payments, and a Calendly link for booking meetings if you need to do that. I wanna show you exactly what this materially looks like, so let's go jump into my computer. And again, remember, you're not building everything here. You are building what you need to to validate the concept and get feedback on how to build moving forward. Now, what we need next, not only to deploy the code or the program or the application that we just built, but also for the critical step that's coming up next, is a domain. And .online is the world's second largest new domain extension. And the word online is part of over 500 million monthly searches all around the world. That's why a domain like your business name, .online, is a perfect domain for you and for all types of businesses, ranging from global pop stars to kombucha brands. Also, we all know that .com domains are super saturated because they've been around for so long, so it can be really hard to find a .com extension that actually fits your business or idea. Meanwhile, the large inventory of .online domains means you are much more likely to get a domain that perfectly fits your business or idea. For the startups you create, .online is available for just 99 cents for the first year using this code, johnlaw, at checkout. Head over to get.online or the link in the description to use that code and get yourself set up with a .online domain. Okay, so here we are. I have Clod pulled up, as well as Repl.it and then Streamlit and GitHub. So generally, you are gonna be better off if you understand how to program and code when building anything with AI that's gonna write code for you, and there are limitations. But as far as, again, the whole focus here is, because really, validation and MVP and whatnot, this is a really powerful stack. So Repl.it is an integrated code environment. So I can say, build me a statistical significance calculator, I'll click start building, and then we'll let that run. I did just cover this in another video, as well. And then Clod as a chatbot, and specifically 3.7 here, is really just the best kind of text or multimodal chatbot for code. So what I generally recommend is if you wanna build something that's full stack, you just want an MVP and really to visualize it, then Repl.it is fantastic. You can see right here, it just built a little plan for us. We'll approve that, and it will start. And then if you want to add certain features, if you want to pull that into an external environment, like I have an external code editor right here, then you can start to work on that through Clod. So if you don't know how to code and you wanted to work on, for example, this script that we just used in a previous video, I can say, understand this and help me expand it to multiple user URL inputs, and then it'll go and work on that for us, right? So it instantly understands that, and then it's going to just give us a new script here. Meanwhile, over here, Repl.it is still building this for us. Now, you can deploy internally in Repl.it. So if we go over here and we go to deployments, you can use a very cheap autoscale plan and then add your domain. And as I'm saying that, we can now see what Repl.it built. So you see that that was one prompt. That was one-shotted. We didn't even ask it to improve it in any way, and this is what it built off the bat there where this actually works. We can edit the inputs and then it will give us a calculator there. And you can build anything of that sort with this, and then you just talk to it and you say, this is how I want it to get better, how to change, et cetera, and then you deploy that to the web. Now, if you don't want to deploy through Reddit, excuse me, Repl.it, because that does cost, I think, $20 or $30 a month, you can do so. Oh, and Claude is still working for us over here. Just look how many lines that is. That is really impressive. The two other platforms I mentioned as far as deployment goes, so taking that code and making it accessible as a website or application on the web are Streamlit and then GitHub. And Repl.it does use Streamlit. You may also know Repl.it actually uses Claude, the Claude model, to run its AI system, so a lot of this is just Claude as well. Or you can use GitHub, which is very, very cheap or borderline free. And for example, if you just copy the pages, build the GitHub pages repository in GitHub, it'll do it all for you. You just paste your code in, you connect your domain, and then you can just deploy that, and this is a tool we built for another video, and you can see the app we built right here. Streamlit is also a very easy way of deploying where you can connect Streamlit if you sign up or sign in to your GitHub code and then push it to the web. So overall, that's a really easy stack right there. It's a little bit messy, that UI, but you can see the code right here. So again, Claude, Repl.it, and then GitHub and Streamlit for deployment. Next, I wanna tell you about the most important part of this 24-hour startup cycle that only became possible a few months ago. So for step one, you built the MVP or the landing page where customers can go to engage with some initial version of your product or service. So you can at least get them in the funnel and see how engaged they are. Now, step two is leads, and this one is super easy. You can just go to any number of cheap lead providers like Apollo and filter or search the leads that you want, which I've talked about in other videos, and then export those into a CSV. Now, in whatever form, once you have a list of your CSV leads with emails attached, now you're ready to go to step three, and that should only take 20, 30 minutes max. Now for step three, we're going to merge those two prior steps, and those cumulatively should have only taken a few hours of effort. Now we get to actually test our idea in front of the market. We're doing this with a specific tool launched by a platform called Instantly. Now, normally to send cold emails, you have to buy the domains, then wait a few weeks to warm up those emails so they don't get into spam. But now with this tool launched by Instantly, you can just buy pre-warmed domains with generic addresses. So for step three, we will literally just sign up for Instantly, buy a pre-warmed domain or two in some email addresses, and then instantly, no pun intended, start sending emails. And you can also set those domains, of course, to forward to the actual domain of your business or website. So we'll upload that list of CSV leads we created into an Instantly campaign. And then the last step here is to write the actual cold email sequence, or the three to five emails sent over the course of a few weeks that are designed to get the recipient to either respond to the email, book in a sales meeting, or visit your website, and sign up or buy your product. There's a whole world of literature out there about how to do this well. My two cents are just keep it short, keep it clean, and really make that core value proposition clear. So building the product with AI probably only took one to three hours. Leads maybe took an hour, if that. And this final step should have only taken as well one to two hours to get an initial sequence and script up there. That means if you start at 9 a.m. with a cup of coffee and a good idea, by 3 p.m., you can get campaigns out there to actually test your product or service. And again, the point here isn't to be perfect or actually build a business in a day. The point is to take action, get something out there, and start getting real feedback so you can exponentially improve the product or service that you're offering. Compared to what most people do, which is take six, 12 months to do anything or take any action, just try it. Just get it out there, test it, fail fast, and also improve fast. And that is the magic of AI and how to start a startup in one day. Again, make sure to use code JOHNLAW, all caps, to get yourself a .online domain for 99 cents for the first year. And that said, cheers.