Scott Clary (4:45)

I know that when you were raising money, you ran into friction with investors. And you exper. You've spoken about this a few times. You experienced something called the pause, which is, I'm assuming you're speaking to investors that don't Quite understand because the people that have money that are the investors in Surat, they're not the people that usually put money into a lot of tech startups. So I'm still curious what made you want to still build in Seurat knowing that some of the investors weren't as educated in AI and tech and you were running into this friction when you were raising. I, I know that, I know that the Internet doesn't have a zip code and you could build anywhere, but there was a reason in your mind why you still chose to build in this city. Like help me understand it from like a strategic perspective. Even when there's a little bit more difficulty finding the right people, talent, investors, mentors, whatever.

Vishal Virani (5:44)

So like there are like two or three things. The first thing is like kind of in DNA of building something the way I said, like I was the first engineer of my family and that's, that's the same DNA, like I want to build something first time from my city and it, it may inspire like lots of other founders. Like you can do it from anywhere. Don't, don't think about your location. So it was quite a philosophical angle. Another part is like every city has a good resources. Because of some opportunity or because of some restrictions, they were not able to move out of the city. So another part was like because I, I started my first venture in Surat which was like it consulting and service company and we like having hundreds of service and consulting company in Surat so like you can start service and consulting company from anywhere. And that's, that's how I started with Surup. And then I realized like we have some advantages. Like I cannot be Google of SF because Google is already there. But if I build something differently and if I build something exciting, I can be a mini Google of Surat and that's where I can attract a good talent. And at my peak time I may require like 100 developers or 200 developers. If I'm building a product company. That's, that's what is like team size. You require like okay, out of 20,000 engineers, do I have like those 200 sharp minds who can join me where I can help them to think differently, where they can join my vision and they can do things in an out of box way. And my answer was yes. So instead of like doing a competition in a bigger tech cities like I decided like okay, can I build something differently in my city where I get the talent and to be honest, like I got it. So that was a strategical angle that I took and definitely the third part was like okay, definitely VC generally took that micro moment of pause when they heard Surat. Like Most of the VCs in SF do not even know about where Surat is. But at that time, like okay, do I have a product which can speak on behalf of me? And again, the answer was yes, definitely. It was quite a bit of a struggle. But I know like when are you, when you are trying to do something differently? The way I, I, I attempted my engineering from the farming background. It was the first time and always like first time things challenge you. But when you, if you are able to crack that thing, then it will give you a, a different kind of a dopamine. Like okay, you are the first one who is doing it from your city. That's even a bigger moment to cherish all cherish about. And, and that's what I picked. Like, okay, it may have a little bit of a trouble, but if, if I succeed I can able to create an impact on like lot many people who are building it from different, different cities. And one of the example is like in my recent trip of sf, we hosted a dinner and there were one person who come to me and he said like I'm from Mexico. I was working with the Mexican government, Mexico government. And then I moved to SF to build my startup. But when I read your story, I thought it was a mistake. I should at least give a one attempt. Like I can also build it from Mexico and it may empower like lots of people around me in Mexico. But I give up without even without putting a single attempt. And that's what I always want to create. Like, okay, you need to just have a courage to do that attempt. If you get success, like you can impact lots of life. And that's what happened with me. So it was a kind of a struggle with me because of my location, but that, but I want to do it somehow. And I tried to navigate through all the challenges which I was facing from my city. But at the end what we celebrate is results, not the journey.

Scott Clary (9:24)

I don't think when most people are starting a company, they understand that as an entrepreneur, as somebody who's building something, you will have a lot of people that look up to you. If you're successful, you'll have a lot of people that will look at what you've built and, and you will inspire a lot of people and you will give hope to a lot of people who are ambitious, young and just starting out. I don't think a lot of people even think about that when they're first starting because they're so stressed about just making this thing work. But you actually had the foresight to say, hey, I want to give like inspiration to people who are anywhere in the world and help them understand that they don't actually have to be in, in Silicon Valley or in, pick a big city, Bangalore, wherever. They don't have to be in a major urban city that is known for tech to build something that is, that is bleeding edge, that is disruptive, that is life changing. Not just for themselves as an entrepreneur, but you know, for their customers. And ultimately like what you're doing is like really bleeding edge and disruptive and, and life changing for the world. So I think that, I think that more entrepreneurs, yeah, I mean I don't need everybody to make it harder for themselves, but they should, they should take a second and think, hey, how am I building this? Am I building it in a way that's ethical, that's responsible, that sort of, I don't know, I'm doing it in the best way possible because there's going to be a lot of people that are looking at me and are going to look at me as like the hero in their entrepreneur story. And I think that's kind of what you're doing when you're saying that doesn't matter where you are, you can build from anywhere. So I, I just wanted to point that out. It's, it's a really nice thing.

Vishal Virani (11:12)

Yeah, Just want to add one thing like don't create a wall of excuse around you if you are not from the, a good Ivy League college or if you don't do not speak good English or if you are building it from a tier two city, those are the excuses that you are giving it to yourself to become a good entrepreneur or building a successful startup. Like it's all, it's just all about your will and a courage to build something. And then once you have that goal, like how much hours of efforts that you can put to learn something, to unlearn and relearn something and like how serious you are about the goal. So that's, that's where you start making the strategies using the first principles. And once you understand the first principles, you will realize like, okay, location and everything does not matter. What matters is like, okay, how you understand your users, how you understand their problems and how you build this, build something and, and then you will find your ways.

Scott Clary (12:03)

Now this idea of the pause, like when, when you're pitching this to an investor and they're saying there's no way you can Build this in Surat. I think that I'm going to make an assumption here and you can tell me if I'm right or wrong. But I feel like if you're pitching AI to investors at all, there's already going to be a pause because it's something that's never been done before. But you're also pitching it to the employees that you're hired. Like you're constantly pitching, you're constantly getting people on board. So when you're building something that truly has never been done before, I think a lot of people are going to experience this pause, like this doubt in, in the, in the stakeholders and the people that they want to get involved in the mission. Walk me through what you experience as somebody who's building something, insurance, but also building something that again has never been done before. How do you deal with people doubting you? How do you deal with the pause from investors, from your employees? How do you overcome it? How do you keep going after. I don't even know how many meetings you went into where they said, like, no thanks, I'm not interested. Like what, what keeps you going past the point where a lot of people would throw in the towel and give up?

Vishal Virani (13:11)

So there are like two things. The first is like, okay, it's not just about the city. Sometimes it's about your educational background, like from which college you did your engineering or PhD or whatever it is and like what your ex company. Like, okay, I'm X Meta, X, Google, X Netflix and then the city. So to be honest, like I'm a founder, which is like having a fault in everything. Like I'm not from any good X companies like X Google and all. I always introduce myself as ex of my ex because right after the graduation I started my first company and then second and third venture. So I always introduce like I'm ex of my ex. I did graduation from a very ordinary college which no one knows and then then from the Surat. So what matters is like okay, there is. When you are building a startup, there should be something which can speak on behalf of you, sometimes your domain. Like, okay, I work in a meta for this particular AI project and that speaks about you and your way will be easy where people can join you or people can fund you the same way. Like maybe sometimes you're a city that you are living and from where you're building that can speak on behalf of you. If you have nothing, there is one thing that always can speak about you, which is your work, which is your product. So I knew it very well like okay, I don't have any of these three things, so let me build the product right, Let me try to get the kind of end users that can prove my capabilities. And when someone will use my product either as a VC or as a user, they will start believing in like these guys are capable enough to build something. So definitely we got multiple rejections like whoever didn't try our product, they may reject us. They raised their eyebrow when they heard about the Surat and then they have the questions how you can build it from Surat. But when the current partners, they tried my product first and that's where they get impressed, like okay, this is a world class product. And then it was an inbound interest from them, like okay, we want to talk to you. And when they understand the vision, when they understand the underlying engineering behind the product and everything, they got that conviction and we got the term sheet. So a simple thing is like okay, you need something which you can use it as a strength. Either it can be your city, it can be your background and if you don't have anything, just use it your product as advantage. Show the product, try to build the conviction you may get couple of rejection, it's fine. But once someone will understand your product and your vision they will join you. So don't worry about that. Just build right product for the right user set.

Scott Clary (15:50)

Where did the idea for Rocket come from? What was the problem that you were trying to solve?

Vishal Virani (15:56)

So I would say like before the Rocket we were building a startup named Device which was like a figma to code startup where you just need to submit your designs and we will convert into code. And it was a pre LLM age when it was a very hard time to convince someone like machine can generate the code that can be accepted by human beings. So we started in 2021 with the vision of like we want to automate the entire software development life cycle. So it was at that time it was a very hard engineering problem. People were not ready to accept the code that is generated by the machines. So we started building and then we realized like okay, LLM make it the entire process super powerful and we understand the domain very well. Like what kind of encode can be accepted by users, how they think, what they think and all. So it was like three and a half years of like entire engineering and the knowledge of this particular space. And we see like when we see like okay, now people are habituated, now people are changing their entire a kind of a behavior with the AI. Like they want to create the applications at their Own. They don't want to reach out to any software developers. Like product managers wants to build something that at their own. Solopreneurs wants to build something at their own. We see like, okay, with the AI, people now start changing the behavior. We realize, okay, this is the right time. And that's where we pivot from device to rocket. We shut down that entire startup because it was like a pre LLM era and like our entire positioning was attached to figma to code. Because we started with that, we were kind of in category creator in that particular space. So we decided like, okay, this is something which we should live now. And we want to go for a bigger vision where we can impact the world with our web solutioning approach, where we can provide a solution of every problem for our users. It doesn't matter what they are doing. So the long vision is like, okay, if you have the problem, just come on our platform and we'll give you a very deep solution about it. And the first thing, we started with coding because we already had that background. We understand that particular space very well and that's how we pivot from device to rocket. So rocket was not just an idea which we get it overnight. It was like three and a half years of learning understanding of the user, the understanding of the technology and everything which push us to get out of the comfort zone of device, which was growing steadily. We had like couple of million dollars in our bank at that point of time. And we decided with the, the way LLM is getting powerful, this will be the only small feature in future. So like, do I need to just sit on that particular product that we are building it from last year and a half year, or can we pivot and can we go for a bigger thing? And that's where my co founder and I agreed on something which can impact the world better than what we are building at device. And that's how, that's how we started Rocket.

Scott Clary (18:49)

You did a very difficult thing which is pivot away from something that is arguably successful. I think that many entrepreneurs have a hard time pivoting even when it's not successful, to be quite honest, but pivoting when it is a degree of successful. How do you know? Just as a general sort of teaching, you know, teaching moment for entrepreneurs. How do you know when you should pivot away? Like, what was the thing that made you actually make the decision? Probably like a hard decision and say, hey, it's working, but there's like a much bigger opportunity over here. Let's go after that as a.

Vishal Virani (19:22)

As an entrepreneur. See you are a chief visionary. You, you need to be chief visionary officer who always can think beyond five years or 10 years. So when GPT3 launched, like the first town hall that I did with my team was like, okay, this is iPhone moment. And this is not understand like this is just iPhone version 3. Whatever you experience with the GPT3 is just iPhone version 3. We don't need to think about iPhone version 3. We need to think a power of an iPhone version 15, how powerful it will be in future. And we need to make all the decisions which can be aligned to like iPhone version 15 power. So just imagine how powerful GPT15 will be. And when, when we try to just even imagine what kind of powerful, what kind of power AI will have after 10 years or 15 years, every one of us was shocked. Like okay, world will be very different. And that's why we always say like there is a world before GPT3 and there is a world after GPT3 and are we ready to adopt this new world reality? And the answer was no. Like okay, right now the current product is not ready for that. So do we need to choose the comfort or do we need to choose the conviction? Because we have a conviction on the new tech. Like okay, this is going to change the world. So we chose the conviction over a comfort. And that's how like that was the moment where we start doing our research, like okay, what can we build a part of this startup? And that's where we make a decision like okay, whenever we are ready with our research, we will pivot and it will take like, it took almost like two years to figure out like okay, what is the right tech where we can create a very bigger impact. And when we were ready with that pocket, we decided to pivot. So it's always about for all founders conviction and you need to understand a change in a market and at what pace. Initially when SaaS applications came, it was like after 10 years you were able to see the impact. Now every six months you are able to see a bigger impact than the entire SaaS industry. And if you are not able to see that change and if you are not able to add up those change in your product, then you will be vanished. So ultimately that was a simple thought. And every time we need to keep ourselves awake, like okay, what if GPT or Claude will do something or what foundational model is doing right now? If they do something similar for a rocket, what will be our strategy? So being a founder, you always need to see like okay, what will be the after three months or six months or a 12 month or a two year, after two years, what will happen and are you ready for that or are you doing anything which will help you to survive in that particular market? And again, I'm a very big believer of a first principle. So when you think from that angle, ultimately you will make all the decisions which makes you future ready. So doesn't matter like what ICP you are selecting, what product roadmap you are picking for your product and do you want to continue with the current concept of your product or not? Every answer will come when you look at the future. And that's what we talk about the vision. Like okay, what is your vision, why we talk about that, what is like that vision is like how, how far you can see. It doesn't matter. It's a possible vision or a plausible vision or a completely impossible vision. But are you able to see that far or not? As a founder, that's the most critical skill that every founder needs to have. And when you are able to see that future, you will able to easily pivot. Anyone can do a pivot or anyone can change their path if they are able to visualize what is going to happen in future.

Scott Clary (22:58)

Can you explain to me what it means to think in first principles? As a founder, I, I love this idea and I think that's probably why you are going to be successful. But I think that especially for first time founders, they, they are a little bit more short sighted and they do think okay, they're chasing revenue more often than not. And I think that that's why, you know, I just mentioned if you have a successful business, it's very hard to pivot completely. It's probably not because they're thinking from first principles, but explain that concept. When did you first learn about this concept and how do you apply it to rocket and how you think?

Vishal Virani (23:36)

So first time I heard about this principle from the story of Elon Musk. Elon Musk is very big believer of a first principle and that's how he built all the startups where everyone is saying okay, you cannot run two big companies in parallel where he is running X and SpaceX in, in parallel. So he was like the one, he's the one who is challenging every norms of the world like okay, you cannot build a rocket which can come back on earth again. He built it. And when, when I was trying to understand his philosophy and thinking, I got to know about the first principle. And that's where like the first time I read the concept, I tried to understand each and every element, how I can adopt it in my life. And, and that was the moment from where I start learning about it. I start practicing it. Every time I ask questions like okay, why I'm doing this, I get into granular detail of my decision like okay, if you will not do this, is it the end or not? So ultimately in the first principle it's very simple. It talks about like core fundamentals, make a decision based on the core fundamentals. Like okay, if I just give an example how I understand the GPT3 different than anyone else is like everyone was saying, this is the prompt engineering, this is the prompt technology that you need to add up and that's how LLM functions. But for me it's like okay, why LLM is acting like that? So I will do a lots of prompting with the LLM. I will try to understand the fundamental behavior. I will read like okay, how LLM is functioning. I will get into every micro details of granular details of those techniques and all. And I will reinvent my own techniques. Like okay, I don't want to follow any prompt techniques given by the world because like I understand the granularity. So I will build my own thing. So the first principle will, will help you to build your own methods. Not like okay, if you will follow someone's method, if they stuck, you will stuck. But when you understand from the first principle, you can invent your own methods. So that's why I always say like focus on the principles. If you just focus on the methods, you will stuck somewhere. But when you are focusing on the principles, if something is not working, you know how you can control it and that is applicable in all the decisions of your life. And that's what I always believe. Like okay, do not just read blog books and all understand underlying psychology or underlying psychology and the granular detail behind those decisions from that book. And then try to think like in your perspective how you can build your own methods. So that is like first principle thinking where you will not just accept anything in that pure form. You will just divide it into like smaller, smaller chunks and you will get into granularity and then you will create your own methodology out of it.

Scott Clary (26:22)

Quick question, what's your go to when you got 10 minutes before a meeting or a workout? For me, it just used to be whatever I could grab, which usually meant skipping meals entirely or just grabbing something that left me crashing an hour later because it was just full of garbage. That's why I'm partnering with Huel, this black edition ready to drink is a complete meal so it has 35 grams of protein, 6 grams of fiber, 35 essential vitamins and minerals. It is no sugar added gluten free under five bucks. I always keep a few of these in my fridge and honestly it solved the whole back to back meetings. Go go go nonstop, no time to eat problem. Super well and this one's new for me. It's Huels Daily Greens. I had the blueberry this morning. Honestly first impression it was way better than I expected. It's developed by registered nutritionists and diet. There are 42 vitamins, minerals and superfoods. Only 25 calories, 4 grams of fiber and just 1 gram of sugar. I throw one back first thing before my morning calls. Every single morning look, if you're running a business, time is the most valuable asset. Huel makes healthy eating simple and they also just launched into Target stores nationwide so you can get it everywhere. Try both products today with 15 off your purchase for new customers with my exclusive code scotthuel.com Scott tried both products today with 15 off your purchase for new customers with my exclusive code Scott S C-O-T T@huell.com Scott use my code fill out the post checkout survey to help Support the show that is Huell.com Scott they really make healthy living taste amazing even if you're on the go. Healthy eating, healthy lifestyle doesn't have to taste bad. It doesn't have to suck. NetSuite is a success story. Partner now what does the future hold for business? If you ask nine experts, you'll get 10 answers. Bull market Bear market Rates are up, Rates are down. At the end of the day, it'd just be easier if somebody invented a crystal ball. But until then, over 43,000 businesses have future proof themselves with NetSuite by Oracle, the number one AI cloud ERP that brings accounting, financial management, inventory and HR into one unified platform. Here's what I love about it. Instead of juggling multiple systems, you get one source of truth, real time insights and forecasting that actually peer into the future with actionable data. When you're closing your books in days instead of weeks, you're spending less time looking backwards and more time focusing on what's next. Whether your company is earning millions or hundreds of millions, netsuite helps you tackle immediate challenges while seizing your biggest opportunities. If I needed this product in my business, this is what I'd use. It's a game changer for business visibility and control. If you want to see how AI can transform your financial operations, download the CFO's Guide to AI and Machine Learning for free. That's netsuite.com Scott Clary that's netsuite.com Scott clary netsuite.com Scott Clay I think we're actually in an age of, of information consumption, but not actually applying it. People are reading, you know, a book a week, but they're just skimming over it and they're not actually understanding the core principles or the actual information, how it impacts their life or their business. Like you, I think it's much smarter to go deep on a few topics and understand how they actually impact and test if they actually impact your life or your business in the way that they say they will versus just accumulating facts and accumulating ideas.

Vishal Virani (29:42)

Here I want to add one example if I have the time. So let's say you have lots of time.

Scott Clary (29:46)

Go for it. I love this, I love this idea because it's something that's so relevant to me right now. Because especially for the podcast. I mean I love when people listen to the podcast. Obviously it's what I do for a living. But I think my biggest call to action is like, yeah, you listen to a great interview, but are you actually taking these ideas and are you testing them out in your life or are you just collecting all of these ideas and assuming them to be true and then you're parroting them in your board meeting or with your employees or with your investors and you sound smart, but you actually have no idea what these ideas really mean.

Vishal Virani (30:19)

Yeah. And so when, whenever you are making decisions. So let's say the way you said, like people read lots of books and then they don't know what to do next. So let's say I have a problem and someone is coming up with the concept of an okr. You will read it, you will find like OKR can change my life and you will start implementing OKR and then you will stuck somewhere why it happened simple. When you, when you decode that, when you get into root cause analysis, the first thing is like okay, when are you reading the OKR book? You need to understand what circumstances someone introduce the OKR in Google and then you need to understand. Those were like a Larry Page Sergey who understand the okr. What kind of a talent he, he, he is having at that particular time in the past. Like all, all greatest of the greatest brain of the Silicon Valley was working with those two brilliant people who were inventing the Google. And those guys were the one who opted the okr. They adopted the okr. They, they implement in real life. And now you, you see your situation, how many people you are having around you with a very high ownership and that kind of a mindset, that kind of an intellectualness. So you need to understand the situation very well and at, in what particular situation OKR was introduced by someone and adopted by someone. If you have a very similar situation, you can definitely adopt it as it is. But if you have a very different situation and if you try to adopt OKR by just reading an OKR book, you are going to stuck. And, and that's where all the founders needs to put their own situation. They need to think deeply. Definitely you can, you can bring that structure, but you need to modify it or you need to explain it to your team in a very different way. If you just distribute OKR books to everyone and on day two, if you are thinking like, okay, everyone will be adopted, everyone will adopt the okr and your company is going to be very different, it's not going to happen. And I'm, I'm telling it from my own experience. I did it before four or five years and every year when. So at that time I even I, I didn't knew about the first principles so I, I was making such kind of indices without like thinking about anything else. But now I'm, I'm thinking like 5 times 10 even like I'm not 100 perfect in my decisions, but at least I learn it from that. And I always try to implement First Principle. I'm not expert of it, but I'm learning every day. I'm trying, I'm practicing it. So maybe after like couple of years I can be like completely first principle driven founder. Right now I'm not 100 first principle driven founder, but still I think like this.

Scott Clary (32:54)

So we're in this like awkward AI phase right now, but explain to me what Rocket is actually doing and why it's sort of solving for the function that it's actually like how it's actually building apps, but also how it's different than all these other tools that are kind of like falling flat.

Vishal Virani (33:13)

So like there are, there are a couple of things. So let me start with Rocket. Rocket is a vibe solutioning platform where we help users to solve their problem right now by generating the code. And the vision is to as I said like solve the problem with a very different spectrum. So our upcoming releases are like completely aligned to the bigger vision of solving the problem in every facet. So ultimately we'll talk about that, but right now that's what we are doing. I'm talking about that particular part. So you just need to come in and give your app idea like okay, I want to build this particular digital business just like four, five liners, nothing else. And Rocket will do the analysis of that particular space geography for which particular demographics you want to build the application everything in like 20, 20, 25 minutes. And it will come up with the complete concept like okay, if you want to solve this particular problem, this can be really good solution. And it will give you complete 8, 9, 10 screens of solutions like okay, how you can design your dashboard to solve particular problem or how you can create the form, what field should be there, what kind of invalidation you can keep. So from validations to interconnections to every component, we will generate in like 25 minutes. And that's how by just putting a one prompt, we'll help you to solve, we will help you to visualize the entire solution of your problem. This is just like first part. Then we have a command. So ultimately like okay, now if, if you don't like our design theme, you can just put one slash command like okay, slash change design theme. And we'll help you with your new design theme. You can just upload your reference image. So sometimes now I'm covering a one more answer that one more question that you, that you were asking. Like many people feel flat and they, they feel like okay, this the system with solving my time problem, actually it's increasing my time and why it happens is like okay, being a human, you have something in your mind. You think like okay, I didn't like this design. This is a very broad statement. You have some reference in back of your mind and you are trying to map that particular thing with the output generated by AI system. And that's what you are not liking all about. And now if you want to explain that entire scenario to AI, it will take like a 500 lines of prompt or like 50 lines of prompt. Depends on like how dissatisfied you are with the output. And people are not willing to put that explanation inside the AI system. And that's where they see AI system is failed. And many times we, we see a. We observe like what people put in, in. In the chat box. Even with the interaction with the chat GPT or with the Rocket system. I don't like it. Can you, can you give me something else now? I don't like it and give me something else. Both are very broad statement. And that's where AI do not understand you. And we know like okay, you are not the great prompt engineer because you didn't learn about it. And like AI systems should be capable enough to understand your context and everything. And that's where we have another feature where you can say just like slash command change design theme, upload any inspiration that you're having from a dribble or from any other system. That's it. And we will automatically identify what are you trying to say and we will change the entire design theme the same way. Layout fix or let's say now if you want to implement login people are saying like okay, you need to understand the prompt engineering to implement authentication. What we are saying just use 1/command authentication, press enter and we will implement that entire thing. If you want to implement like a chat system using the OpenAI or cloud anything, just slash OpenAI implement OpenAI with chat system. That's it. In eight minutes of time we'll implement entire chatbot system with integration of an open AI. So what we are building is a smarter system than putting a instead of inputting a cognitive load on our user. And that is like a fundamental observation between the other guys who are building this building a system in this particular space versus why we are very different. Our entire focus is on providing a solution in the most easiest and the comfortable way for our users instead of pushing them to learn about a new prompt engineering techniques and define more and more prompts to get the system. So right now what everyone is doing is like okay, you need to put a prompt, they will provide you one screen of output. Then you need to put more and more prompt like okay, I need this as well. That as well. And that's how they need to take that entire cognitive load of building the system versus here we are saying just put a five liners of problem statement. Sit and relax. After 25 minutes we will notify you your system is ready to review. Just come if you want to modify anything, just do it. And that's it. That's what our current system is capable of. Now going forward what we are saying is like why we are calling our system a wipe solution? Because building an application become a commodities. Nowadays on any platform you can build the application but still the bigger problem is like what's to build now let's say many of us is having idea I want to build something in this particular space but I don't have people who can do enough research for me or they can put the enough effort to help me like what exactly I should build to achieve something. And that's where the entire vibe solution concept come in A picture in nearby future will help our user to think from a very different perspective where we'll solve the day zero problem by telling them this is the, we will do a research on behalf of them, we'll write the PRDs, we'll do everything for them and they just need to select the PRD from the prd. We'll generate the entire application. So that's how our entire focus as a web solutioning platform is solving a day zero, day one and day two problem. And when we talk about the data problem is like okay, whenever you make your system live, that's not the one time job. You need to observe lots of data, you need to think about the scale, you need to think about how you can create a better user experience for your end users. And that's where we'll give more and more context to our system. You can plug in your third party data analytics systems and all and we will suggest like okay, here you can do this improvement. That's how you can implement your, you can make your security better and user just need to select those suggestions and we will implement it automatically. So we are building a very, very advanced agentic system to solve like a problem of a day zero to day two. So that's where they, we will empower them. So ultimately the goal is to create a impact the GDP by providing them a kind of a platform where they can go digital, where they can grow, where they can take their product to a next level. Doesn't matter either they're building a D2C system sitting somewhere in let's say Brazil or Africa. But they have a very good product and if they want to go live, if they want to digitize their entire application, they should do it without any technical knowledge. Like just put a prompt, we'll handle rest of the thing for you. And this is the future that we are able to see with the rocket. And that's what we are solving right now.

Scott Clary (40:20)

It's actually you're mentioning they should be able to do it without any technical knowledge but it's actually more than just technical knowledge. It's it's like full product knowledge. It's, it's understanding everything that like you mentioned before, like understanding data and understanding like UI UX and understanding security features. So it's, it's not just saying like I want this app to work based on what my prompt is. It's like okay, I want this app to work great, here's a first version. But then we're giving you all these recommendations that probably a non technical person wouldn't even know to Think about. And we're going to implement those because when you look at a lot of tools, so what I was just, you know, discussing about how all these companies are trying to implement AI and all the people that are using AI and the companies realize that it's taking them longer than just doing it themselves. So they're, they're solving for. I, I guess the best way to put it is like what is it they're solving for? Day one is that it? And then day two it breaks or day two it, it's. Is that a fair assumption with a lot of the AI tools that are out being used right now?

Vishal Virani (41:25)

Like, true. So everyone is just trying to generate the code and give it to the user. Like, okay, you can just host it on our platform and you can build our application, but no one is talking about like, okay, once I go live, how will you help me to increase my revenue? How you will increase to, how you will help me to handle my scale and everything. And as you mentioned, like, okay, it needs to be recommended by system. Like we know the technique. Non technical guy don't know anything about the product management. He don't know anything about the data analytics. He don't know any, anything about like how to grow his digital business because he was doing a non digital business till now. Like he was running a fitness center. But because of your platform he jumped into digitalization of his business. So System should recommend him. And that is all like day two problem. You have a product, you have application in front of you, but now how to get a business out of it is something that should recommend by the AI. And that's we call a day two problem. And, and that's where our entire focus is all about.

Scott Clary (42:26)

Why do you think founders optimize for like this, this demo or this, this first version versus day two. Like it makes sense to me. Like if you want to build again, if you want to be a product first organization, if you want to build a company that centers around a really good product that solves a really important problem, solving for day two seems to be the most logical thing. But I mean we see a lot of founders not, not building for a day two problem. They just sort of build for that flashy demo or that first experience. Why do you think that is? Like why? Like, I mean I've used Vibe coding tools myself and I've experienced literally everything that you're discussing and I'm like, this kind of seems like a little bit lazy. Like it seems like they didn't like go all the way and I don't quite understand why, but I don't know. What's your, what's your thought? Is it just because things are moving so quickly? Maybe.

Vishal Virani (43:22)

See, I respect all the founders. In fact like because like Lovable and Bold started this vibe coding space, we got the idea we can also also get into this. So here, here are the things like every, every startup having their own individuality. Like okay, maybe Lovable is focusing on creating the landing page, a good branding website where people can come and create a digital portfolio. So as a founder it's always about like who is your ICP for whom you are building, what is your particular vision is all about. And even a day one problem is having its own depth. So it depends on the vision of the founder, what they are building and where they are getting the good growth. For us, like when we were deciding our icp, like we put some of the things into non ICP brackets. Like we don't want to get into that because someone is already solving it. So getting into that zone can then create unnecessary competition for you and there is no chance of in growth. So everyone, whoever is building in any particular space as an entrepreneur, they are doing some good work and they know their ICP very well, they know the GTM strategy very well and that's why they are building in that particular strategy. And when you know like okay, you are deep into it and when you are getting a really growth by solving that particular thing, you will not able, you may not want to get into something else from the very beginning. Definitely. Once you see like okay, maybe we are launching in that particular space, we are getting a good traction, they will also jump the way we started wipe coding after, after seeing the growth of Lovable and Bold. So it's all about like market research, understanding of your icp, what they want. It's. It's all about the user, what they are looking from you. So from our GTM strategy what we got is a serious audience and they were like asking us a very hard questions, how you will solve my problem. And that's where we decided like okay, we want to build the entire spectrum from a day zero to day two instead of just focusing on the day one problem. So everyone can do with the capability of AI, anyone can do anything. It's all about what you believe, what is your philosophy and what your user wants.

Scott Clary (45:28)

I will play devil's advocate and you can be an impartial founder and you don't have to have an opinion on this. But, but in my opinion, I think that some companies have grown very quickly because of the hype around AI. And if there was not the hype around AI, they would have not raised the amount of money that they would have raised. And I think that we're going to see at some point companies with real products that solve real problems. Those will be the companies that last. And I think there's going to be some companies that have grown incredibly quickly that will have to pivot or do something because their solutions are great. But I don't think they justify the amount of money they've raised or the valuations they've achieved. So that's my personal opinion.

Vishal Virani (46:19)

It's not about the valuation, but see on what parameter you are building a business. I always believe in sustainable business. And when I talk about the sustainable business, like, okay, what kind of gross margins you are having, what kind of cake to LTV ratio you are having, like is it, is it like your growth is justifiable or not? Because like sometimes rapid growth put you in that kind of a situation. Like every user cost you a penny, which is getting into negative direction. So if you are having a negative gross margin and if you're growing very quickly, you need to raise more and more money just to survive because you know you are losing your money on every penny you are earning. So I don't call it a sustainable business. Definitely if you know your metrics very well, if you know your mathematics, like, okay, this particular time, if LLM cost goes down and I will, I will be a profitable company where I will have a millions of users go, like, you can do that, but that's the bigger assumptions, like okay, LLM cost goes down, what if it will not? So it's not a problem of money, it's all a problem. Often like mathematics, like what are you doing, what are you believing and what are your assumptions that should be aligned? That what I believe, like, okay, only crazy funding will not help you to make any decisions like okay, that will help you to survive if you are growing without unsustainable metrics. So that's what I believe. This is my opinion.

Scott Clary (47:46)

One thing that's interesting about Rocket, because most other competitors in your space, they deliver code in, you know, sub three minutes. You mentioned before that rocket takes about 25 minutes. So what was the rationale behind that? Obviously I'm assuming because it takes a little bit longer. It's, it's creating more code. It's, it's taking its time. It's, it's creating like a very high quality product. But when all the competitors are pushing Things out in a very, you know, in two, three minutes. And then you're pushing things out in 25 minutes. Was that a concern? Obviously I know that like Fortune 100 companies are piloting rocket right now. You have like over 10,000 paying subscribers, so it's working. But obviously that was a calculated risk that you took.

Vishal Virani (48:32)

It was a quite calculated risk. The way you said initially, we had a very big internal debate, like, okay, will be able to create the impact, like people will wait for 25 minutes or not? And initial thought was like, let's do it and let's see it in the data. Will people wait for 25 minutes or not? And we say like, okay, 90% of our users are ready to wait because we can see like, okay, they wait for 25 minutes and then they check the preview and everything. And that funnel was like after sign up, after putting the first prompt, like 90, 90 to 92% of people were waiting just to see a preview what we are generating. And that's where we got the confidence, like, okay, we can go with this. Now why we are doing it, like, why we are sticking to 25 minutes of philosophy versus three minutes of generation philosophy. Because what we believe is like a kind of an Apple to Apple comparison. Like right now we are generating 8 to 9, 10 screens of output from a single prompt without putting any kind of a cognitive load on our users. Whereas I said like, okay, most of the people are just generating one screen. Then you need to put lots of prompts to get your final output and all. And as a user you need to think what you want in a system. But in our case, like, what we observe is like user want AI to surprise them. Like, okay, I'm putting my problem as AI, you help me to solve. And in a 25 minutes of time, actually we are doing all that research and we are giving like end to end system as I said with the validations and all, where users see a bigger level of an impact. So if I put over 25 minutes of output with this 3 minute of output. If you want to build a very similar level of a system, it may take two to three days of time for you because you will constantly put new prompt iterate it. You will do all sort of work which we are doing in a 25 minutes of time. And this is like quite a complex system. But even if I consider the easy system in general, in an average, like we are reducing 8 hours of work to 25 minutes of 25 minutes in rocket platform. And when I say 8 hours of work I'm talking about like other wipe coding platform. If you, if you want to achieve the same quality and same level of comprehensiveness, you may spend eight hours of time to doing all the prompting work versus we are able to generate in 25 minutes. So it's in actual. It's not 25 minutes. It's our first results. We are taking. For the. Generating the first results, we are taking 25 minutes of time.

Scott Clary (50:54)

I think that have you found, have you found with a lot of companies just building in the AI space, they're optimizing for speed. Like do you see that that's sort of a common thread between many companies.

Vishal Virani (51:07)

Yes. So and many times I also got this feedback like okay, the real vibe is to generate the results in three minutes because we are building a wipe coding platform. I said real vibe is, and my philosophy is like real vibe is like delivering the solutions without putting a cognitive load on your user. And that's always a debate. And like many times I heard this feedback, but this is the feedback from a generic perspective, like general feedback and all. But when I look at my data, the funnel and all, the kind of in conversion rate that we are having, we see like, okay, this is working really well for us and this is the kind of an USP that we are building against our competitor. So we stick to that particular philosophy. Like definitely we are trying to reduce our time. When we were started it was like 25 minutes like recently in the recent update of last week, now we are taking only 12 to 15 minutes of time. So it's not like we just want to stick to our 25 minutes of time. But what we fundamentally believe is like our first result should be comprehensive enough, it should provide the solution. So we want to stick to that particular philosophy and wherever we can able to reduce the time with our engineering efforts without compromising with the output quality, we'll focus on that. So definitely it's not like okay, we are not caring about our user, but like okay, if, if, if, if I need to make a conscious call between our output quality versus speed, I will always focus on the output quality.

Scott Clary (52:28)

So funny how what Rocket allows you to do and what any technically any vibe coding platform allows a non technical person to do is to create something that they wouldn't have ever even been able to have done in like a year with their current level of technical knowledge for the majority of like non technical like vibe coding users. Yet it's so wild how everybody is so is like they're everybody's attention span now is like an all time. Right.

Vishal Virani (53:02)

So fundamental, you know, fundamental behaviors of users is getting changed, right? Change now.

Scott Clary (53:07)

Yes, it is, it is. But it's not just because of, it's not just because of LLMs or, or, or, or other AI tools. I think that everybody's attention span across the board has just been reduced to almost nothing because of how we consume content. I mean like we're in like a gig economy where we can have everything delivered to us instantaneously. We're optimizing for 30 to 45 second videos on social media. But I'm just, I'm laughing at the fact that somebody is using Rocket to code, to code something. Somebody who couldn't can't code. Like somebody who can't code is using Rocket to code something. And there's a conversation about 25 minutes versus 3 minutes when a year ago like you would have had to have spent like a hundred thousand dollars, I don't even know, plus hired a developer plus like it's so wild how even.

Vishal Virani (54:01)

Like before, like couple of years, years, if someone tell you like okay, this is a problem statement, how you will solve it, what will be your dashboard looks like or how this features looks like, you may take like seven days of time just to think about the solution of that particular problem statement. Now in 25 minutes of time, like in fact like in a 15 minutes of time, you are getting a solution, you are getting a code, you are getting everything. And still it looks like, okay, you are taking a 25 minutes of time to generate this solution. Like look at this, 15 days of effort versus 15 minutes. It's like, and that's a good time. And that's where I was talking about that iPhone version 3 versus power of iPhone version 15. This is the power of AI and this is just in beginning, we are just scratching a surface. After five years it's going to be even powerful. And that's why I advise everyone if you are still in that mode of like, okay, should I adopt the AI or not, Should I get into that or not? Doesn't matter you adopt it or not, just try it. And that's where you will realize like okay, how you can change your thinking and how you can change your entire workflows and everything. The one thing that you talk about like lots of people, I forget to answer that question. In fact, like you talk about lots of people tried to implement there and then they went back to their original workflows because they were not able to figure out the values. One thing that I observed, which is quite different before, like when we were implementing the SAS SaaS applications or SAS tool in our organization, what we are trying to do is like, okay, which SaaS application is best for my workflow? And that's how you were adopting it. Nowadays the world is very different. If you think you will, you need to, you don't want to change your workflow and some AI system will come and sit in your workflow. It's a different, it's a wrong thinking. You should think about the restructuring of your flows, processes, team and everything to add up the AI and, and that's the right way to do it. And again, think like in that iPhone version three, like before iPhone, how you were operating the mobile phone or even before mobile phones, how you were operating versus and then mobile. Can you, you didn't say like, okay, mobile need to sit in my workflows. You change your entire habits, you change your work style, you change your everything to adopt that smartphone in your life, lifestyle. So the same way, if you want to adopt AI efficiently and effectively, you should think about the restructuring of your process team and everything. And then you can find out. You will figure out any tool which will help you to boost your productivity, like 10x20x that you, you, you will never think about. But when you try to fit AI inside your workflows, you will say like this is not working because the tool is didn't meant to be like that. So that's, that's the one observation that I'm having. And maybe like for all your viewers, I can give this advice like, okay, think differently. When you are trying to adapt the AI in your life cycle.

Scott Clary (57:00)

Indeed is a success story partner. Now if you're hiring Indeed is all you need. Let me give you an example. If I needed to hire a new editor for this show, I'd go to Indeed and be super specific. Not just can you edit audio. I'd say I need someone who's edited a conversational podcast for at least three years, gets our style and knows our software. Someone who's done this before. And here's the thing, with Indeed Sponsored Jobs, I'd get people who fit that description. I'm not digging through resumes from people who've edited one YouTube video. I'm getting actual, actual podcast editors who know what they're doing. People who've worked on shows like ours and can prove it. That's what makes a difference. You get people who actually are what you're looking for. According to Indeed data, sponsored jobs posted directly on indeed are 90% more likely to report a higher than non sponsored jobs and people are finding quality hires right now. In the minute that I've been speaking to you. Companies like yours have made 27 hires on Indeed according to Indeed Data Worldwide. Spend more time interviewing candidates who check all the boxes. Less stress, less time and more results now with Indeed Sponsor jobs and listeners of this show will get a 75 sponsored job credit to help you get your job the premium status it deserves@ Indeed.com Clary just go to Indeed.com Clary right now and support our show by saying you heard about Indeed on this podcast. Indeed.com Clary terms and conditions apply. Hiring do it the Right way with indeed HubSpot is a success story Partner now look, if you're in marketing right now, or if you're an entrepreneur who's hired marketers, or if you're a founder who leads a whole marketing team, you know the direct you are creating content for 12 different channels. You're launching campaigns, you're scoring leads, you're analyzing all the data and somewhere in there you're actually supposed to do great marketing. And it's exhausting and you're spread too thin. But this is what actually works for marketers to give you your life and your time back. HubSpot's marketing hub HubSpot's content hub, combined with their new AI called Breeze. Now what this actually does is basically everything that you shouldn't have to do yourself. It remixes your content instantly so you're not starting from scratch every single time. It handles lead scoring so you know exactly who to focus on, which customers will probably convert. And it pulls all your analytics and your data and KPIs into one place instead of having them scattered across 15 different tabs and tools. Plus, the AI agents that HubSpot builds for you can automate the repetitive stuff that's eating away your day. Bottom line, you get better results faster without burning out. See, with HubSpot, everything's connected in one platform instead of duct taped together. So if you're tired of being spread too thin, check out HubSpot.com marketpl to see how this actually works. I'm curious when you think about. When you think about how AI tools will continue to get better in the future, I'm wondering why, like you look at OpenAI and ChatGPT and Claude, why aren't. Why aren't they optimizing for reducing cognitive load? Because I think if you ask Sam Altman, he'll say that he is. But I don't. I mean, I use it every single day. I don't Think that it really is done a good job of reducing cognitive load on the user.

Vishal Virani (60:05)

Well, I agree it's like yes and no. Both definitely they are improving every day. But being into foundational models, they are having a lots of challenges. So if you look at the GPT3 versus GPT5, definitely they make it lots of advancement into that model. But because they are serving a horizontal purpose, they are serving everything that user, user want from that platform. So now the expectations are that high, we are not able to see those improvements and all because like in two years our life has been completely changed. So when we look at like foundational models like Cloud Gro or OpenAI, we think like, okay, now they are not improving that fast. But our expectations were very high. Like when GPT3 launched, we were not expecting anything at all. But now we are expecting like they should give us like 100% accurate result without saying anything thing they should understand us. So let me, let me tell you a very small observation. Like, okay, after the memory feature, like initially when you were doing something with the ChatGPT, you need to explain everything in the individual threads, but now they are having the memory feature. So ultimately whatever you are explaining. So let's say if I'm explaining about myself to a chatgpt in one thread, in another thread, if I'm talking about like, okay, consider my Persona and write this thing for me so it will remember your entire thing like who you are, why and what are you doing and everything. It's a very small feature, but a very, very powerful. Because now every time I don't need to explain the entire context to a chat GPT. So a very small feature but very powerful. So you need to observe like where they are moving, how they are moving. But as I said like now our expectations are quite different. So we think like they are not moving enough fast. So it's always quite a problem.

Scott Clary (61:50)

I want to, I want to talk about AI replacing people. So you've said a few things. You've said that AI won't replace people, it replaces mundane work, which I agree with. You've also said that if someone isn't adopting AI, they will, they will definitely get replaced. So for the average person who's listening, who's a developer, they're, you know, they're listening to this conversation. They're like, what do you mean AI is not going to replace me? You just told me that in 25 minutes, not only can it do research, it can do, it can do, it can do product, it can do UI ux, it can do basically everything that I do in my job. How will this not replace me as a, as a developer? What are you saying to that person?

Vishal Virani (62:35)

So see, there are like two things. AIR will not replace you is the right and wrong. Again, if you will not adopt AI, AI will definitely replace you. And what I fundamentally believe is like AI will reform and restructure everything from now. So let's say, when I say it's a restructuring and the reform, initially we have a roles like in product designer, product manager and then maybe a front end developers. The going forward maybe the role will be AI product designer where a designer is doing everything using the AI like conceptualizations to writing the PRDs, to delivering the end code. So that's how every role will be redefined. The way AI is getting powerful and now you need to see like what role will be best with the AI. And if you will build your capabilities in that particular way, AI will not able to replace you. But you just want to stick to your particular role. And if you are not a greatest of the greatest expert of the world, then AI will replace you. So all the mundan and the mediocre work will be replaced by the AI. So you need to be expert in something with AI and that's how you can make sure like you will not be replaced anytime in the world. And there are like 1% of the people who are like super expert in whatever they are doing. Either they are doing the designing or they are writing the code. And all those were not going to get replaced by any system of the world because those were the like people who are believing in fundamentals, they understand everything and those are the resources who don't need to adapt anything. So if you, if you're 100% sure you are that kind of an expert who don't need to be changed with the AI, who understand every fundamental piece of a job that you are doing, you don't need to worry. But if you, if you know like you are, if you are not that one person and if you are coming you are a part of 99% then you need to think like okay, how my role is going to redefine in the age of AI and how I can be fit for that. And if you adapt it like no.

Scott Clary (64:37)

One can replace you, say I'm a founder, I can build an app now in 25 minutes. Who do I need on my team? Do I need any developers? Do I need a cto? Who do I need? Who am I hiring now? What is your vision? What is your true vision? For the future, is it going to be everybody's a solopreneur?

Vishal Virani (64:55)

A good question. And you put me in a controversial space now. Okay, so right now if I talk about the current version of Rocket, definitely you need a developer because you need someone to validate the code that we are generating at this moment. Like okay, this is the secure code. Definitely we are saying like this is a secure code, but at least you need couple of folks. If you are building a really serious system, then in 25 minutes is not just enough. So for an example, there are multiple things and as a founder, what role you want to own. So initially founder were handling the product management and the engineering and then other things. But in the age of AI, when you know like anyone can build product very quickly, you need to focus on the GTN so you can hire like few folks who can manage your product and who can deliver very quick product using a wipe coding platform like Rocket. And as a founder you need to focus on the gtm. So a solopreneur is definitely a good thing, but it's just an just to validate your idea. So now here I talk about the fundamental shift, okay, with the wipe coding platform, when you are creating the first version initially to build the mvp, to conceptualize your problem and the product, you were taking like lots of time and you need to put a lots of money. And that were like most riskiest phase of your life, of your startup life, where you don't know will my idea work or not. And you were trying to explain your concept verbally to your users to get a feedback. Now with the AI, without hiring a developer, as a solopreneur you can go very quick in market like in just month of time, you can get the validation, you can do everything and then you can think like, okay, now I want to scale my product, I want to launch more and more features inside this and I will focus on the GTM to build my first million dollars of ARR, let's say for an example, that's where you can hire a couple of developers. You can give a Rocket platform to them so they can expand your product very quickly as a single person, it will be tough for you to manage your gtm, your rest of the things, validation of the product and also definitely will require developer. But before AI and this kind of platform, you may require like 5 to 10 people of team. Now you can do the same thing with like just like two or three folks. So that's how things are getting different. But you will not require any person at all in Your team is a wrong statement.

Scott Clary (67:18)

I'm just wondering what, like what a software, like what a software team looks like in five years from now for somebody like and, and, and for somebody who is like again they're not the top.01%. They're not CTO. They're not somebody who is, you know, has built apps and exited. They're not somebody that has those accolades. They're somebody who is like, like learning cs. They're, they're early stage in their career developer and they want to have a future. Well, describe, I want you to describe what a software team looks like in five years from now, but I want you to describe it so that, that person who is concerned about AI taking their job because they're, you know, they are, when you're just starting out, you are mediocre. That's, that's how you are. You start out mediocre. You don't start out the best of the best. So what can that person do to be part of that software team or that development team in the future?

Vishal Virani (68:19)

The one thing like now there is a scope of an AI developers, data engineers because AI needs lots of things like fine tuning of model to having like advanced reg system, the contextual system. And also definitely the developer's job will be there. But as I said like okay the front end and back end developers job will be redefined to AI developers where they will do a lots of AI work and optimizing the data database. And right now still we are having a lots of hallucinations and also ultimately the job will shift from ordinary front end, back end developers to AI developers. The way this entire industry started like software industry initially we were doing like C&VB.net and C kind of programming. Now everyone is like having a proper front end and back end developer. So whenever there is industrial revolution come or whenever there is a big change happen in your, in your industry, always remember like you will not be replaced, your role will be redefined and are you ready for that or not? So ultimately even if you are a mediocre software developer either you can think like how you can become the AI developers and you need to start learning from now so you can become expert of that. Another part is like okay, if you don't want to change, if you want to stick to a front end and back end developer and as I said like there will be need of them as well. But then they need to think using a wipe coding platform or using any AI platform like Cursor or, or a rocket or anything how they can be like 10x powerful and productive not just in delivering the code but okay now because I have a time I can own the product, I can think from the product management lens, I can deliver the code, I can test it at my own and that's how I will. I, I can do a five people, five person's job at alone like product management designer, coder, the unit tester and the DevOps guy single handedly. And that's how that role will be very different. I don't know what we call it but the way you said now we are having a forward deploy engineer fde. That role was not there even like before two years. Now everyone is talking about the Forward Deploy Engine FDEs. Everyone is talking about a GTM Developer roles are completely new in the age of AI. And that is going to happen. So you as a, if you are a mediocre resource you just need to figure out like where you will fit in and you start need to learn those things just to make sure you will, your future will be secure.

Scott Clary (70:48)

One thing you mentioned, which I think is another thing that we're all trying to solve for is hallucination. And I, I, I'm curious as somebody who's building something right now, where does hallucination even come from? And, and then how do you, how do you actually end up solving for it and, and making it so that we hallucinate less so we can trust the results a little bit more. Because I know that, I think that now they do tests and benchmarks again like on generative chat tools and it's getting better every time there's like a new version released. But what is that? What, what is causing that hallucination?

Vishal Virani (71:23)

So the first thing is like a vague requirement and the vague prompt. So what is that hallucination? You were expecting something and you get a very different result. Definitely. Sometimes every software has error. It's not a new thing to be honest. Like even human being also hallucinate. Like okay, let's take a very wild guess. Often support system. Okay, you have like support department, you have 10 people generally what is our success rate or like error rate if there is no AI at all? All human also operated like 60 to 70% of accuracy rate generally like okay, sometimes they give a wrong response to a user or sometimes they give a response where your user get offended. They also make a 30% 40% time mistakes. But our fault tolerance rate was such a high. When it comes to human to human interaction we say okay, it's a human being, they can make mistake Even if we consider the a kind of a hallucinate rate or hallucination rate or error rate with the AI system it may be sometimes like 10% or less than 20% but we are not able to. We are not having that fault tolerance rate inside us. Like okay, if machine machine should be 100% perfect. Even after like 30, 40 years of like software development, still software are having the errors. Otherwise we should not have the maintenance team at all. Still the systems are broken, which is not built on the AI or anything which is built by the human being. Still a login page maintenance having the error. Like you don't receive the OTP if you see everywhere there is a hallucination, everywhere there are errors. But with the AI, the kind of an hype AI is getting the definitely the controversial side will also get that hype. And that's where we are making this hallucination a lot bigger than actually it is. Every day everyone is improving. But now people are expecting 99.99% or 100% accuracy, which is definitely not going to be there. And the reason is the second point. You explain something to a system and you're expecting 100% result but you are not translating what you are having in your brain. It let's say, let's talk about like again. I'm taking one example. You are my client and I'm just putting a row vague requirement in a LLM. I want to write a proposal to scot and just draft it off for me. And then I will say like okay, hallucinating is having lots of error. It's not generating the way I want first is, I'm not saying who is quote what I have in mind, what kind of terms I want to write, where he is sitting, where I am sitting, what are the jurisdictions that I want to cover LLM like you didn't give any context to LLM and you're expecting 100% accurate result when you create a boundary. So when you put a very broad word like okay, create a best proposal for me, as I said, it's a too broad requirement. There is nothing like in best. You need to define what is best for you. In your case that's where LLM can give you better results. And you will see like now, now you are getting a super quality results which is not generic, which is not hallucinated and all. And as I said still because of an error of a foundational model, you still may have like 2010 to 20% of variation. But you should adapt it by the time it will it will get fixed. But be very specific. When are you putting any instruction in, in any AI system just to get the right results?

Scott Clary (74:41)

When you look at, I would actually love to do another conversation with you about like just the future of AI in general, because that could be like a very, that could be a whole other probably two hour conversation. But when we just look at, you know, how fast we're moving, the success that Rocket has had, the success that like a lot of, a lot of AI tools and platforms have had, where are we going to be in the next year from now? Like, I mean, you are in it, you're building. We're talking about hallucinations, we're talking about, we're talking about reducive cognitive load. We're talking about the effects on the actual people that are using it and how they have to upskill to stay relevant. Like, I know this is a very broad question, but just what is your, what is your prediction for what the world looks like a year from now? I won't even say five years because that will be crazy to even try and guess. But a year from now, what do you think the world of AI looks like?

Vishal Virani (75:42)

So I know that will it happen in a year or now year or like later? But we will see a glimpse of Jarvis with everyone. Like when we were moving, when we were watching that movie, everyone was fascinating with the Jarvis and everything. Like the way it was handling every work, to be honest. We'll start getting a beta version of Jarvis with everyone affordable because when we're looking at that setup in the Iron man movie, everyone was saying this is not possible or I cannot afford it. But now everyone will have their personal Jarvis in a year in some form of an applications or a multiple form where you no need to focus on any mediocre work or abundant and repetitive work that will be replaced by the AI for sure.

Scott Clary (76:26)

You're only a few months in. You have over 10,000 paying customers. You have Fortune 100 companies using and piloting your product. So when you think about the speed and the growth of Rocket, do you think that it's luck timing you as a founder? Like what, what allowed you to be successful this quickly? Have you thought about that?

Vishal Virani (76:52)

That to be honest, not. It's like five months of success was with four years of hard work is what I always think about. Like it's not just in like overnight success. So someone reach out to me like, okay, this is crazy. Like you got $15 million of fund in a 15 weeks. I said, it's just a tip of the iceberg like four years of hard work plus a 15 weeks of rocket which leads to a $15 million of seed round. So it's always about that. Like I never think about the timeline and all what I always think about like where I want to go and how quickly I can go there. And like I try to align my entire team towards that particular vision and the next milestone. And that's how we work. Like 10 hours, 15 hours a day. It does not matter. We just want to work and we just want to show it to the world.

Scott Clary (77:42)

What does success mean for you?

Vishal Virani (77:44)

Yeah, I don't think much about it. Like I always think about the next milestone. But for me the success is like when I everyone will start using AI in their day to day life. Doesn't matter what kind of background that person is. And that system is built from my city. When they got to know like okay, Surat is not just for the diamond. Surat can make a really good AI product as well. For me that's a kind of a success statement.

Scott Clary (78:07)

And another thing that I think is admirable. I mentioned this at the beginning how I love the fact that you are putting your city on the map and you're being a good role model for other entrepreneurs but you also focus on not just making like a product first business. So again you build something that people actually want to use and works well, but you also focus on building a sustainable business. So just touch on that for a moment because I think that a lot of people who are listening, who, who are entrepreneurs, they've heard of the importance of product like you know, product first marketing, product first entrepreneurship. Like have something that actually works as opposed to just, just an idea that you're raising money on. Like have something that actually solves a problem people want to pay for. I think that's a very healthy way to look at entrepreneurship. But what does sustainable business mean and why is it so important?

Vishal Virani (78:54)

So sustainable simply means like you need to have a good gross margins where you can survive your cake to ultioration. It's to be in control without a good. So in this, during the SaaS business like 80% were like a gross margin, a standard gross margin. Below that gross margin was like bad in the age of LLM when you know like your highest cost is going to be the LLM just to serve your customer. And if you are just operating at like a single digit gross margin or a 15% 20% gross margin which is like right now everyone is doing is quite a risky business in a long term. If you, if you, if you are not able to raise a big rounds and all and when, if hype gets flattered, you are the first one who will get impacted. So you still in the edge of LLM. You need to think about your engineering. You need to think about everything to build a good gross margins. And that's what I call a sustainable business.

Scott Clary (79:46)

You've mentioned a few different mental models that you apply to your business and to Rocket. And just like yourself as an entrepreneur, what would be one more mental model or idea or principle that's really helped you out that you think somebody who's just starting building a young entrepreneur would benefit from learning?

Vishal Virani (80:05)

Yeah, the first mental model that I talk about is like first principle. Then second mental model that I always implement in my business is like it's even not a mental model, but it's like you should control your mathematics, you should control all the funnels, everything. Being a founder, like if you are not in a finance space, you cannot say like I do not understand you. You need to control at the higher level, you need to have a knowledge about everything, like how your a burn rate is, where your burn rate is moving, how much fund you are having, everything. So ultimately when you know the mathematical equations very well, you will able to control it. Otherwise you will not able to connect the dots. So always take a decisions after connecting the dots. Do not make a decision just based on a one situation. You need to think from the future perspective, you need to think from the current perspective, your competitor perspective, the capabilities your team is having. So a business is all about connect all the dots, figure out your capabilities and then build it. Like okay, my computer is doing this. So you don't need to jump into that. You need to think at your own. Why you want to do that? Is it aligned to your user? Is your team capable to deliver it or not? So don't just get into a vanity matrix. Don't just get distracted by what people are doing. Have your own thinking and do something which you believe. So just stay away from the vanity and less distraction. Just focus on something which you are building.

Scott Clary (81:34)

Do you find that, do you find that when founders are scared to go into the data or the numbers, they end up making more emotional decisions.

Vishal Virani (81:45)

That's what founder is all about. In the early days, generally everyone make that mistake. And many times like in the early days, the problem is like you don't have the enough data. And I would say like it's not all every time. It's all about like making decisions after looking at the data Sometimes you need to take a bold shot like if chatgpt look at the data before building the chat GPT then they will never able to build the chatgpt at all because no one were talking about the GPT 2 and GPT 1. It's all about like this thing will work for the world and that's how they build it. So sometimes founder needs to be that visionary where you don't have the data and still you are making a call. But it should not be like out of your, all 10 decisions you are making out of emotions. Once you have a stable business, once you have a user data you need to combine both.

Scott Clary (82:31)

I want people to connect with you and I want people to go explore and try rocket and I want them to to try it out. If you want to connect with some of the people that listen to this podcast, where would you want them to go? Any kind of socials that you want to send them to website that they.

Vishal Virani (82:47)

Should go check out x.com vishalvirani91 and on LinkedIn it's like LinkedIn.com vishalvirani so these are like two platforms that I always reachable.

Scott Clary (82:57)

Last thing I like to ask is, I mean you've given over a lot of wisdom and I mean you are a serial entrepreneur and you're building something that is. Is truly bleeding edge. But if you had to go back and tell you know, 20 year old Vishal one piece of advice, what would that piece of advice be and why?

Vishal Virani (83:15)

If. If I look back at like 20 year old Vishal at that time like I was do I. I want to do too many things. Like I wants to do, I want to run a service company, I want to do products and, and many other things. After my 10 years and 12 years of experience, what I realized like okay, just focus on one thing, make it like a perfect. Just give your 100 into one thing. So if you are, if you are at your early stage and if you are having a lots of thoughts and a distraction, just stay away from it. Just pick one thing, align it to your life goal and just put your hundred percent of efforts to that and that's how you can make it bigger. Like if you are too much distracted and if you're trying to do too many things at the in the very beginning, you, you may fail. So I did that in the early days and that's where I wasted my couple of years of life as an entrepreneur where I now I think I can, I. I would able to do a much better job at that time. If I, if I know those facts. So this is like one advice that I can give it to Vishal of like 20 year old who started the business.