Subscribe
Share
Share
Embed
Founder Reality with George Pu. Real talk from a technical founder building AI-powered businesses in the trenches. No highlight reel, no startup theater – just honest insights from someone who codes, ships, and scales.
Every week, George breaks down the messy, unfiltered decisions behind building a bootstrap software company. From saying yes to projects you don't know how to build, to navigating AI hype vs. reality, to the mental models that actually matter for technical founders.
Whether you're a developer thinking about starting a company, a founder scaling your first product, or a technical leader building AI features, this show gives you the frameworks and hard-won lessons you won't find in the startup content circus.
George Pu is a software engineer turned founder building multiple AI-powered businesses. He's bootstrapped companies, shipped products that matter, and learned the hard way what works and what's just noise.
Follow along as he builds in public and shares what's really happening behind the scenes.
New episodes every Monday, Wednesday, and Friday.
George Pu (00:00)
Happy Friday and welcome to another episode of the Founder of Alipodcast. I'm your host, George Poo. As you know, this Friday we're going with the new format that's going to make it more interesting for all listeners to not only learn from me, but learning the best in the tech and business world. So we'll be combining these different resources together and also what can we take away from each of those resources together.
So to start with one of my tweet that I didn't really expand on Twitter, but I find it really, really important is that I said AI will copy your MVP in three weeks. And this is why I never build single product companies anymore. Revenue diversification beats revenue projection. So there's actually a lot more I want to expand beyond this tweet. So we'll hear that. And then we'll also connect with something that Naval I'm sure everyone knows about.
something that he just recently said and how all these things actually connects together into one theme. So for my tweet, think, you know, the takeaway is very simple. AI will copy your MVP in three weeks. Right? So this is not really for even for builders who are trying to build something right now, right? It's for builders and founders and companies and businesses who already have a product somewhere on the marketplace. And this is something to remember AI and somebody with AI is able to copy your MVP in three weeks.
personally think three weeks has been very, very generous. If we think about like some, something very competitive and complex, like intercom, you know, platform, they're being around for, I think at least 10 years plus, or maybe it take three weeks to copy exact features or at least 80 % of features those people have. On the other hand, let's say there's like a company called Carta, which is basically cap table management, right? Which is a very niche space, but they charge an exorbitant high price, you know, for basically managing your
companies cap tables and a cap table is basically a table of your investors and know, investors, investors and different things and how much piece of your equity and, you know, each investor has. So for something like that, it'll probably take a few days at most to build, right? So using AI for a new experience, for an experienced developer with a new idea. So for my company, SimpleDirect, I've spent the past four years building SimpleDirect financing. And I personally think for something that we have built, the moat is actually not in a product itself.
The mode is more about the business model, right? So the problem is for someone who has, who understands the space, who understands the business model and who studies input or financing in and out, they can probably build the MVP in a few days, in probably two to three days. So that is the new realm of possibility we have now with AI. And recently, actually just this week on Claude, introduced Claude 3.5 Sonnet, a new model that has made
their coding efforts significantly, significantly better. I'm still waiting to see exactly how much better it is since it just dropped, but it seems like we're creating and we're opening a new Pandora's box and new developers and developers are be able to build much faster. So with my tweet, let's also hear about a few of existing examples of companies being copied. So cursor, obviously everyone know they went from a niche tool that just first, I think I remember cursor was the first that came out.
with AI enabled coding, right? And then immediately after we have a giant Microsoft, which acquired GitHub launching GitHub Copilot, right? Which is basically the same thing in the Visual Studio Code or whatever ID you're using. You can just have an extension called GitHub Copilot that will answer the questions of developers and be able to help you auto-complete and do it. And I'm pretty sure GitHub Copilot right now is the biggest competitor of Cursor. They probably also have the biggest market share just because they are already, they are GitHub.
You know, so is it monopolizing? I think so. Is it fair? I don't think so. You know, is it fair? Absolutely not. Right. But cursor got copied. And then also there's other startups that copied them. think Windsurf was the one that came out basically doing the same thing. Devin doing the same thing, similar things, know, clock code is an iteration of the same product. It's very different, but also in the same space. So it was a niche tool and then it went for.
It went to I think a hundred million, a hundred million dollars annual recurring revenue and there's dozens of competitors, right? And it hasn't even been a year or so since cursor was launched. So this is really crazy. And will cursor survive? Probably yes, because they were the first one on the market. And also most importantly, I think they not only just continue building the same product features everyone else is building. They have made a really generous free tier for people to use, maintain their first lead. And they also create it into a platform.
They created, they don't even tell you exactly what model they're using. And you're able to basically experience the difference between cursor and windsurf and now open as codecs and different things. So that's definitely something I think it's a moat that they have. So they're not being afraid of being copied, but you know, we'll see, we'll see how they do because all those companies are being copied perplexity. I think we'll talk about it in Monday's episode perplexity was basically copied by Google. Right. And then I remember when it came out.
chat, GBT, basically open AI launch, GPT search, which now just became, like a roll into the main product that basically is a copy of perplexity. And everyone else is basically doing the same thing. Claude added search as well. So right now perplexity is basically struggling as a company. They're expanding different things. think they're expanding into browsers, expanding to something else. You know, they, they expand to many things. So sometimes I see perplexity product offering. like, that's pretty cool. But they do require you to pay $200 per month on the max plan, which
I'm I'm not going to be one of them for a very long time. So that's the problem with Burp-Propaxity. I think it's still interesting because they're still very innovative, but their original product, it's basically losing the moat. And I find it really hard to see how the original product is going to maintain the moat out of the number of users. So we'll see how they do. So for us as well, like I said, SimpleDrag Financing, we have launched initially. No one I think is copying us at the moment, but you you never say never, right? I think tomorrow it's possible someone's going to come in and copy the exact same product that we have.
And basically telling us screw you guys and, know, under price us and, you know, do the same thing. So I think that's entirely possible. So anyone is possible in this space. If you're, if you're like a to do list app, if we're a calendar app, if you're an app that helps people edit videos, anything you're are potentially at risk as a startup. Right. And any single feature nowadays can get commoditized in a single week. Right. Or, or to be fair to some, you know, more complex companies, probably weeks I have personally vibe coded. You know, I understand that vibe coding is not.
the best way to go. also coded with AI. I understand the limitations of these things, but however, I think good developers are able to do it. And the end consumers while paying, using something, paying something that's say $50 per month versus using something for free, but a little bit flawed or even $5, $10 a month, a little bit flawed. think users will be able to use that. They will be able to take the $40 difference and do something else with their money. And unfortunately, that's just the truth. So right now, basically with SimpleDirect, we have actually pivoted for this exact same reason.
we're pivoting to become a multi-modal product company instead of just a single product company. So we went from just like SimpleDirect financing, which is basically a ⁓ product that helps companies to offer buy-now-pay-later company to a series of new applications that's going to help developers, product companies, software companies, and hardware companies to do their work better. So for example, like what we're going to release very shortly is called SimpleDirect ChangeLock, which is a public facing change lock that will allow any product facing company to build.
and release and automatically embed all for free. So that will be our first product on the line is to buy goodwill in the community. But we are building this product completely from scratch. We're building it end to end learning from the experience I've learned before. So after a change lock, we have some other products on the lines. You know, even though I don't know which products were specifically going to release because it would be a while till then, right? We do have some product chat, which is a, which is a private AI enterprise and small businesses.
Proactivity team collaboration app. So we might have that. We have simple direct posts on the line, which handles social media posting and AI insights. So there are different tools, you know, and right now we're only certain that simple direct change lock is going to be released. However, you know, there are many different tools that we'll be releasing and in a year or in two years from now, when you look back at a simple direct brand, you'll be like, geez, they have changed a hundred percent. They went from a single product company to multiple different products. And as a CEO, I can share with you.
This is, I personally think, the only way to survive, right? In the long term, if you stretch it out to like 5 years, 10 years, 15 years, this is probably the only thing that you do right now that's going to matter to diversify into multiple products. And some founders might be, you know, might be skeptical because this is something that completely challenges the startup canoes, you know? So for example, like I have a few I prepared rebuttals for my point. So Paul Graham and Y Combinator, they certainly disagree with this.
They say do one thing extremely well, right? That's their model. And Peter Teo has a zero to one, which is the first book I've read that says build a monopoly with defensible moats, which basically means that one product that's a monopoly monopolize in the market. And also like every accelerator, every accelerators application, if you're a startup company applying, it's going to be asking, oh, what is your competitive advantage? What is your moat? Right. I think, I think these questions are still good, but they work perfectly back in the days when building a
product takes actually 12 to 18 months. And yes, that's how long it actually takes to build a successful product back then. Right. You have time to establish market position. can build brand position. You can get users and you can be the first mover advantage to be on the market. you know fairly well, it's going to take whoever it is to take at least 12 to 16 months or 18 months to come in. Right. Amazon web services are basically one of those companies. They did a headstart and I think it took at least five years for Google to actually enter the market. Right. Let alone catching up.
So that is a very interesting question. then Microsoft was even later. So they were able to use a first move advantage because building at a time was difficult. So they were able to build this giant empire of AWS because of that. But putting that today, it's a lot harder to do that than said. So AI has basically changed the entire game. I personally think, yes, we're not looking at the best AI software right now, but who knows where we're going to be in a few weeks, in a few months. Like I said,
Claude 4.5 Sonnet just dropped. I personally tried a little bit. I really liked the model so far. The demo I watched online, basically they asked Claude to rewrite the Claude.ai interface experience and they tried all the different models previously. None of them worked, right? Only Claude Sonnet 4.5 is able to recreate exact same, well not exact same, but like, you know, very, very close, 80 % close Claude.ai dashboard.
and you are even able to upload files for it to work. So if you have tried building something similar, you know how difficult that is. And this is incredible, right? So the defensive moat theater, it's basically over. I personally don't believe in the word defensive moat anymore. There are a few exceptions which I will talk about very shortly. Also like we have proprietary data or we have network effects, we have brand loyalty.
All of those, think, tend to assume that your competitors actually need months or years to catch up. And now that's just no longer the case. They don't need to wait anymore. They just need a few weeks as soon as like one week, two weeks, three weeks and access cloud open AI to build something really good. So how do we deal with this? I also talk about this week at my Twitter, but I also want to bring to you the sovereign ecosystem model. And this model is very interesting. Basically it talks about when your product do get copied and it wealth get copied, unfortunately.
your ecosystem relationship don't get copied. And this is one of the mode that I have identified. I think will be really helpful for founders. So think about this. If your product is have, you have one product and this one product has only one revenue stream, that is a little bit fragile. One revenue stream equals for fragile. But if you have three to four different revenue streams, that is anti-fragile. That is robust, right? That is having something safe. So that even one or two things don't work out. You still have a remaining two revenue streams to make you money. And I like to use the example of Stripe.
No, I do understand people say, ⁓ Stripe is licensed company, but think about how much regulatory compliance they have to do. Right. So if you take that away early on, once they establish the first product, they can actually do the other product had it not been regulatory compliance in my humble opinion. Right. So Stripe didn't actually waste money defending payments. They understand anybody can do payments and we see the marketplace, you know, squared as payments now, you know, there's a few other European companies as payments. So payments is like something common. Everyone can do payments. But what really.
differentiated them. It's like they have a billing, a billing product, which I personally use that deals with subscriptions. So if you're using subscriptions, I think Stripe building has been pretty robust. Stripe connect is also a very robust product because if you think about Stripe connect, what it does is it basically allows a marketplace. For example, I'm Uber, right? Let's say you're the driver and it allows me to directly pay you for the ride that you completed sometimes even instantly. Right. And that is not something just a payment product can do.
And I have used Stripe Connect before. just think it's really cool. And I think many on market place are using it. Now they're launching treasury products, launching fraud management, they're launching Stripe taxes to help solve taxes for platforms. You know, that's not just like scope creep, right? Because Stripe has all the resources, have all the money. It's actually a very beautiful survival strategy to make sure that their mode is long-term. Their mode is not just in payments, but their mode is everywhere, right?
Their mode is in the ecosystem of the Stripe branded name. Everyone knows about Stripe and they know different things that the F-Sharp is doing and using the same dashboard getting upsell is actually very easy, right? I personally got upsold for like the Stripe radar for fraud management. So I had to pay like a certain basis point more per transaction, right? Same for buildings or for different things. So their revenue obviously ballooning, but their costs, I'm not sure exactly the detail of the cost, but I don't think their cost is as much as the scale of revenue they're able to produce. So I personally do think.
The mode is not a product anymore. Mode is a customer switching cost across multiple different products in your ecosystem. And why do I think it's right? Because, you know, think about it. How hard is it to stripe off stripe to like turn off stripe and switch to someone else. If you have stripe taxes, if you have stripe connect, if you have stripe payments and stripe billing, it's basically impossible and they know it. And that's the way they do it is to tie you up with ecosystem.
I know Apple has done a brilliant example, but you most of us are not Apple. We're not a hardware company. So I'll also throw Apple out there as like a, know, analogy of what founders can create. You know, not many of us can reach the level of Apple, but I think it's a really good example about what ecosystem can mean as competitive modes. So each product on the line, let's say after your first product is validated, you have customers, each product added will give you something transferable. know, customer relationships can deepen because they're using multiple different products.
Even let's say the customer only use two products out of your hundred product which I, you know, just as an analogy, I still think that's a very, very competitive defensive mode that you can be building for a product to succeed in the long term. You're not just defending your features, right? You're building for the next decades and you're not playing defensibility. You're being aggressive and you're getting out there for customers. And I personally think that's one of the only models that can survive the AI commodity and commoditization era.
And if you're listening, I hope you take that away. Let me know if you agree or not. That's something I personally believe in. I'm using it for my personal, you know, weekly daily building exercises. I asked myself, you cannot build something that has like ecosystem embedded to it. And if not, I try not to spend my time on it because I know everything eventually can be commoditized. That is honestly, there is no first mover advantage anymore. There still is. It's just like that timeframe has been so slow. Let's say you do something today, right? Tomorrow or even next week, someone can do the same thing.
And this is not the same before. If you were in the last era building, you understand what I'm saying. It did take a lot longer to build something and for a competitor to catch up. And back in the days, you know, only one person or less of all people can code. Now, I don't know specific in the number of I'm guessing it's probably 5 % eventually will be 10, 15, 20 % or even at the very least, will be more people. So that reduces the mode and ecosystem is your way taking a defensive mode to above. So yeah, so that's a really good.
tweet I thought I would share my own tweet, not to like, you know, promote myself. do think that's something I want to unpack in today's podcast. And let's also talk about someone that we all know, Naval, you're right. Naval Ravi Khan is obviously very successful on Twitter. You probably know, of course, you know about him if you're on Twitter. So he tweeted something recently I find interesting. He said, the purest reason to make something is not to make money. It is not even to make the thing, it's to have the experience of making.
So when I first read this, think it hits very differently than usually the tweets that he wrote, which is more about like mindsets while building, et cetera. In this tweet, he's basically talking about something that most founders completely miss. And that is like the experience of building something is the only infinite gain. Right. And I personally think about this, you know, like to build something is not to make money, to build something is not for the, even the product itself is for the process of building. And I personally think back about my last four years of building a product.
I think I really enjoy the process. have to say, I really, really enjoy the process. And basically the process goes to have an idea, validate talking to designers, getting the wireframes, getting the sketch sketches, and also talking to developers and, you know, making sure everything works at the same time. It's not actually chaotic. If you do it well, it's very, it's very iterable. Every week is like an iterable version. And next this week will be better than last week and next week will be better than this week. So it's something that you really look forward to because your product does add more and more features.
and you do enjoy the process of making. And after all these years, as you know from last week's episode where Sunsetting is in product financing, but I'm not incredibly sad about that as well, you know, because I do enjoy the process of making it. I do know that my team and I have learned all these years about the process of making. So Sunsetting a new product is actually not sad for me at all because I already learned how to make a product. You know, I learned how to make something and I think that is infinitely transferable, right? If I want to make something new, which I am.
making simple to our change log, we are making something new. And I think that's very incredible. And that is something I think everyone should learn about, even though if you're not a software founder, you know, the process of building a business, right? The process of incorporate incorporation, for example, it's a very stupid common thing to do, you know, but if you haven't done it before, never done it before your life, you must be like, how do I do this? Right? You have to run into different walls for something as simple as incorporation. But once you do that simple process, once you do that once you'll learn the process. And if you have to do it again in future.
or explaining it to a friend how to do it, you will be so smooth because you know exactly how it works. And that is the model, the same about how to build a product. You know, I personally think that's really, that's a really attractive way of just like the process of building, right? The moment a product goes live for me, that's the most beautiful day I've ever had. And it's not because of revenue generates because we got to iterate and watch it improve and see the feedback our customers will have on it. So that's like the most important thing I think Naval talks about, which is like joy of making.
And I think to be incredibly successful in your life, you have to be incredibly joy of making something that you're making. And that applies to like, you know, content creation. Let's say you're a podcaster, you're like a social media influencer. Let's say you're a software founder, you're building a software, you're building hardware. Let's say you're like a salon, you know, hair salon owner and you're doing a salon, right? And let's say you're an e-commerce business, whatever it is that you do, you have to love the process of making, you have to love the process of building because if you don't love it.
you're not going to succeed. And someone else who might love it, who actually love it, will do it much better than you, you know? And that is the fact that's going to have moving forward. And I think it only be more true as we have, you know, AI coming into the game as well. You know, so basically that's kind of my experience. I personally think Navol says it, you know, own the experience, do things for the experience, make things for the experience and not for the outcome. Just, I think I'm the perfect example of that because some of my financing failed, but I am still happy.
about it, right? Which is crazy to most people. And I think, you know, what Neville said actually does apply a lot to what we just talked about in the, like being copied in AI era, right? Like basically when I can copy anything and copy in a product in three weeks, your only defense probably is that I can make three more products, you know, in three months because I've done this before in this domain, or I can add more features. I can know more of my customers. I know exactly what to do because I already learned how to do it. I love what I'm doing, right? Versus a competitor.
Actually, even before the ad days Airbnb had one of their largest competitors when they were just starting out. They had a competitor in Europe that basically copies everything Airbnb does, you know, which is like the craziest thing. ⁓ but eventually we all know the story how it goes, how Airbnb won, how Airbnb won, right? So there is always a defensible moat, but that was the most stressful time, I think, in their founder's entire life, just to compete with others. And I'm sure at one point, all of us will probably be competing.
with many different products and many different founders. So I might be competing with a few other really like I admire founders who I admire, but we had to fight, we had to compete. You know, it's intense, but it's probably good as well because it makes you learn. So for me to defend that I am a lot faster than making things. I leave my team making things faster because I believe speed comes from experience. If you're just out of college, you don't have any real experience, but you're a really good computer programmer. I can beat you because I know exactly how to build products that people love, right? Not just products that work.
And that comes from the joy and experience of making, you know, each product, think in your ecosystem will teach you something new, the learning that you do compounds across all products and the brand you make obviously compounds across all products as well. Right. Instead of launching five different websites, five different domains, five different emails and five different marketing experiences, you can do all of that in the same brand in the same product and probably save about 40 % of your time. You're not defending your features in this ecosystem play. You're just expanding your capabilities, which is the most beautiful thing.
Right. And also let's talk more about your experience eventually becomes remote. So I already talked about that. think product products in itself needs a soul. Products needs missions. Products actually, you know, Peter Tios is a man like this, a company or product when they were screwed up, when they started can never be fixed in the future. And I still think about this back today about founders and replications about founders. do think product themselves need a soul and a mission and they need to have a heart. Think about that way.
If your product doesn't have a heart, they don't have a mission. It doesn't have a soul, right? It doesn't have spirit and a mission. You cannot fake that by just chasing money. If I really love my products and I pour everything I have into it, and I really communicate that love to my customers, to people who are paying us ⁓ versus someone who just copies us straight out of the blue ⁓ and basically have no emotions, their goal is just to make money, right? Customers can tell the difference in details. And I think that's why I talk about in the sovereigns angle that devout links explicitly say
Which is basically when you are making the experience, when you live for the experience and you're living through multiple products, you get both. You get the joy and the soul that Naval talked about, but you also get the financial anti-fragility you need to never happen to exit or to never give you up control. Or at the very minimum, if you do take VC money or outside money, you have a raise from moat and increases your chance of exiting eventually. Right? So even if you do exit in the future, this is something that you should take away from. So I think.
For ecosystem founder in the future, be an ecosystem founder because ecosystem founders optimize for infinite creation and ecosystem founders are the ones that you do not wish to be on the opposite of competing against. So that is called freedom for those founders because they have freedom, they have control and they have decades of compounding and it's very, very hard to beat them. And to basically conclude, I think AI cannot copy your full years of customer conversations, your customer emails or support tickets.
your regulatory navigation, your partners, your business development experiences and your distribution partners or whatever integrations, right? And domain expertise. AI could not just copy that people with that cannot just copy that. They can only see what's on the surface. They can only see the interface, but they cannot see the accumulated experience. But that accumulated experience is what to get you to build product two in three months, product three in six weeks and product four, well, hypothetically in two weeks. So, which is possible by the way.
So those speeds are differential, right? You take the most time now, but as you go, you're going to take shorter and shorter amount of time to build adjacent features, to build adjacent products, to build adjacent data and experience. And that is the most important thing. And that is why when the vault is very important, you have to make something for the pure joy, the experience of making. So that is the ecosystem play.
And I think that's important for you as a founder to know. if you, there's some, I also build something like a weekend audit since today's Friday. Let's talk a little bit more about, be honest with ourselves and use this like small framework just to see where you're at. If you're building a software company or you're building a current business that service business, product business, whatever. Here it goes. I have three questions for you and be honest about where you stand for each one of them. So question one, could someone with decent market understanding in your space rebuild?
80 % of your product functionality using AI tools in three weeks or even three months. So be honest about this first question. important. Okay. Question two. So what is your next product? Right. If your answer is blank or you don't think about it, your competitive mode is a little bit fragile. So think about it. What is your next product? Do you have that in mind or not yet? Question three, what did building your current product teach you or service teaches you that transfers to other products you intermane? Right.
And here's how to read the questions and answers. So if you answered yes to question one, you answer blank to question two. So unfortunately you do have a countdown timer. You don't have a business in my opinion. So harsh to hear, I think it's very important time to basically improve. And you know, this is like a warning bell for you. If you answer yes to one, but you also have something really specific for number two question. So you're thinking like ecosystem building, which is good.
So keep your mode in your main product, but also finding something that you can launch adjacent to build an ecosystem. Right. If you answer no to number one, so, you know, you're either lying to yourself or you found something actual magic, right. And audit that answer very, very hard and see where exactly you stand for. Right. And I think this year our question to ourselves shouldn't be like the previous years, which is like, how do I defend this product from copycats? Right. This year's question should be.
How do I become the kind of founder who can make three more products, adjacent products in this domain and build an ecosystem, right? Because we know the experience of making the domain expertise, customer understanding, speed of iteration, that is your only mode left. Everything else is theater. So, let me know how you think about this week's episode. I hope you find the joy of making. I hope you enjoy the episode. Feel free to let me know what you think. As always, if you have any questions or thoughts, you can always email me at george at founderreality.com.
I'm also on Twitter as the George Poo on Twitter. Tweet me on your thoughts and then I'll read it out loud, you know, in the next few episodes. So thank you so much for supporting the podcast and I'll see you soon.