I Asked An AI Billionaire How To Start A Business (10 Steps)

0:00On November 21st, 2024, Anton Oska

0:03launched his first company. 243 days

0:05later, Lovable became the fastest

0:07growing company in history. [music]

0:08Faster than Chat GBT, faster than

0:10Microsoft, growing from 0 to 100 million

0:13annual recurring revenue. In a world's

0:15first, I've arranged for Anton to give

0:17you his step-by-step guide on how he

0:19reached the impossible. As someone who's

0:20been in business for 35 years and built

0:2219 companies, I've never seen what

0:24you're about to see. When you finish

0:25this video, drop in the comments what

0:27you've learned and how you're going to

0:29implement it into your life to make your

0:30business successful.

0:31>> For people that don't know, what

0:33business have you built?

0:34>> I built a company called Lovable and it

0:36lets anyone who's not technical create

0:38software you by explaining this is my

0:40idea. This is what I want to create and

0:42then talking to an AI and it creates it

0:44for you.

0:45>> If you've got an idea, you can make it

0:46happen. Now, well, we're going to go

0:47through now the steps in which Anton

0:49managed to make that happen. I should

0:51probably tell my audience, I have

0:52invested in this business and that's why

0:54I asked Anton to come on and teach you

0:55what he knows. But a lot of people don't

0:57have an idea, right? So, what I'm going

0:59to do is I'm going to label up what we

1:00just talked about there. To me, you had

1:01a painoint.

1:02>> Yeah.

1:03>> And you solved it. That's a good way to

1:05think about if people are listening how

1:07they could go about generating an idea

1:08for themselves. Right. So, let's go back

1:10to those early stages. Where did you

1:12start? Initially I built a tool for

1:14myself to to prove that you can um in

1:18the first example is I could tell it

1:20create me a snake game

1:22>> and this was two and a half years ago uh

1:24where it was like wow that it created a

1:27snake game that was it started it

1:29running on my my computer locally and

1:31then I put that out on a tweet and it

1:34completely blew up from that. I was

1:35like, "Wow, should" and I had to take a

1:38few days and think about is there is

1:40this something I should be uh take

1:44advantage of and build a real very very

1:46very ambitious business around. And

1:48after uh sleeping on it, I woke up and I

1:51was like, "Yes, this is the way that I

1:54unblock 99% of the world who don't know

1:57how to write code."

1:57>> Back then, just to clarify, did you have

2:00a lot of money? Did you have a lot of

2:01money to do this? How much money? No, I

2:03I started it with my I biked to my

2:05co-founder's place and I explained the

2:06vision here. We have this unique

2:08opportunity in a lifetime to create a

2:10new way to create software.

2:12>> We going to call it lovable because we

2:14want

2:15>> the people who create software with our

2:17tool to create something that creates

2:18positive emotions in their customers in

2:21their customers. So we just started

2:22building like we we that was the most

2:24important thing building um getting it

2:27out there getting feedback. How we did

2:29it was we did um like a weight listed

2:33launch so that we could get some very

2:35passionate early users and work with

2:37them to to help understand where did

2:40they get most value of the early early

2:42versions of our product before we went

2:44out in the in the mainstream. We did

2:46raise money but it wasn't the most

2:48important thing for us to be successful

2:49at all. So there's three things right

2:51that I think if people listening you've

2:53got an idea now you you have development

2:55skills

2:56>> okay so many of the listeners they won't

2:58have that the irony is you made a

3:01product that would create the not the

3:03need for your skill in the early days

3:06and now someone else can have your skill

3:07in their pocket right and then you had

3:10it sounds like some element of marketing

3:11there so you posted about it and it blew

3:13up

3:14>> now did you have a big following at the

3:15time

3:16>> not the first you just caught a cultural

3:19thing right with the snake game.

3:20>> So you plugged into a cultural trend and

3:23that's partly why it blew up, right? And

3:24then from that you created a waiting

3:26list but to create a waiting list again

3:27there's marketing. What did you do to

3:28get the waiting list going?

3:29>> We launched the first version of the

3:33tool for nontechnical people which were

3:35lovable. You just open love.dev and you

3:37write in what you want to create. We got

3:38a lot of people that wanted to use it

3:40but we we thought we have to keep

3:42working on this product. So at that the

3:44point in time was still building the

3:46product, assembling the continue to

3:48assemble the team of people who really

3:50really care about building the the AI

3:53software engineer as we call it and to

3:55get feedback and work with getting users

3:58on trying out trying it out and

3:59explaining where they get the most

4:00value.

4:01>> So you found a niche which in this case

4:03was other developers, right? Yeah, those

4:06were some of the early users. But where

4:07we found the the most interest is for

4:11founders who don't have a technical and

4:14that's to that's always been and and

4:17still is the

4:19>> the group and even though there's many

4:21many people types of people in marketing

4:22and everything that's using lovable

4:24that's the group where we have this the

4:27most raving fans. Again, I I always

4:28think about my listener as someone that

4:30doesn't have any money, doesn't know

4:31marketing, and they but they might have

4:34an idea or they're trying to come up

4:35with one, right? So, the pain point

4:36helps you come up with an idea, I think.

4:38But then, um, you mentioned something

4:40else, money. So, development, you then

4:42got the marketing going with this weight

4:44list, which I think is brilliant.

4:45Everybody I know that's been successful

4:46started with a weight list, including my

4:48platform, we had a weight list, 10,000

4:49people, which with the weight list, you

4:52can go to an investor with an idea and a

4:54weight list. Is that what you did? So,

4:55did you when did you raise the money

4:56bit? when did it wasn't of course at the

4:58early stages when

4:59>> so it was actually in the very very

5:00early stage we raised raised the money

5:01but me and my co-founder we didn't take

5:02any salaries we had very very few people

5:04on the team

5:06>> um so I think we could have done it

5:07we're just just building just building

5:10ourselves I mean of course it helps to

5:12be able to spend money on this a AI cost

5:14money and so on so so

5:14>> but tell tell people how you raised

5:16money then because you've got this idea

5:18you've launched it you've put up this

5:20waiting list how did you raise money

5:21early on

5:22>> look so when I built this open-source

5:25tool for developers it was large large

5:27interest from investors. So, and I think

5:28that's generally the best way to raise

5:30money when investors come to you. You've

5:32shown that you can you can do

5:35>> uh you can get attention for what you're

5:36building and you can show that

5:38>> you you can build a product. You can

5:40build something that's creating a lot of

5:42value and that's what we had shown and

5:44then they came to us. Invest is going to

5:46be very busy now because as people

5:48listening can now develop a product

5:50themselves on lovable build a waiting

5:52list for that idea and then the market

5:55will see that you've got an idea and a

5:56demand.

5:57>> Yes, there is a lot of

5:58>> you made that whole process that for you

5:59was manual right

6:01>> so much easier.

6:02>> Did you have some business support or

6:04knowledge or did you just go with your

6:05gut back in those early days? It really

6:07depends on what type of business you're

6:08building. But for us um we kind of

6:11followed how a typical Y cominator

6:14company would raise funding and um there

6:17is a lot of re resources online or you

6:19can talk to an AI about how that how it

6:21looks like in a sasha scenario.

6:22>> Did you have any idea that your business

6:24was going to be worth billions? Did you

6:25did you did you feel it or that that was

6:27your focus? I mean I it wasn't my focus

6:29but I knew that there was a path to uh

6:34just build the ultimate tool to create

6:36software and it's going to be something

6:38completely different than what it's been

6:40in the past going to be this platform

6:42where you have um very very intuitive

6:44experience talking to an AI that does

6:46everything across building the products

6:48iterating on the products with your

6:50users even distributing and uh like

6:54running your business around this all

6:56the things that it takes to incorporate

6:59a company even and and that platform is

7:01what we're building.

7:01>> What was your first few hires?

7:03>> It was engineers to continue to build

7:05the product because it's a very very

7:06complex product. It's not something you

7:07just use AI to put together. Maybe these

7:10days you could you could do that but

7:12back then you really needed really

7:14passionate engineers that care about the

7:18entire product and they care about how

7:20do we get the most out of the AI

7:22capabilities today. Why don't we make a

7:23separate video with one of your

7:25engineers

7:26>> about how it works because I think yeah

7:28it's good to understand what's in the

7:29black box right and understand what your

7:32prompts do. So um again I I love to

7:34understand you know the things now that

7:37happen next. So um we've got this

7:40brilliant idea that's getting traction.

7:42You've got a co-founder. How did you

7:44pick your co-founder? What was the

7:46decision for the co-founder partnership?

7:47>> For me it's very important to have

7:49someone who you work together with as a

7:51team. It's you're going to go through a

7:54lot of hard decisions and I wanted

7:56someone who compliments me and I mean

8:00it's really smart about taking business

8:02decision on across building the product

8:04and across everything else and my my

8:06co-founder had since he was like a

8:08teenager built companies by creating the

8:10products himself

8:12>> and I respect him I trust him and that's

8:15what I focused on. Uh we talked

8:16yesterday off camera about um servicing

8:19the 1% or servicing the 99%. I think

8:22it's interesting this point for the

8:23audience. One of the things that you did

8:25is you thought right how can I help

8:27people that are not developers

8:29>> which is a much bigger market than say

8:31just helping developers although the

8:33original idea was very much you know

8:34appealed by developers. Was that the

8:36very beginning or did that get developed

8:38as you were building out? You're like

8:39when did I realize from the very

8:41>> very beginning right and so so you

8:43thought to yourself I'm going to mass

8:44market support but mass market marketing

8:46is quite a different thing right so you

8:48actually your initial success is in

8:50niche marketing so did you did you think

8:51okay I'm now I'm going to market to the

8:53mass market so we I didn't think too

8:55much about the marketing in the

8:56beginning I thought about building the

8:57product making the product good and

8:59leveraging that there's an appetite for

9:01new AI tools and people wanting to try

9:03new AI tools so so that was of course

9:05play that was playing in our favor for

9:07growing so fast as we've A lot of people

9:09won't launch until it's perfect. So,

9:11what's your take on that?

9:12>> I mean, you should you should launch and

9:13get it in front of users as early as

9:15possible, but once you know that this is

9:18this is the product that you're going to

9:20make your money on and this is like this

9:22is it, then you should I think you

9:25should strive for a perfect product and

9:26especially in this these days the bar of

9:30quality is very high for most consumers.

9:33If you go very niche and you solve like

9:35a problem that no one really no one else

9:37is solving, then you you just have to

9:38solve the problem. So, and you don't

9:40have to think too much about the quality

9:42and maybe it's the market isn't that

9:43big. So, you can build um like one

9:46product line that works and then the

9:47next product line that works in the same

9:49niche category. If you're building like

9:51for for a large market, then you should

9:53really focus on perfectly quality after

9:55you've pro proven that you have product

9:57market fit. Users want to pay for this

9:59and you it's growing.

10:00>> I think this is where a lot of people

10:01get stuck, right? They build something

10:03and they're scared to put it out because

10:05people might judge them. They might say

10:06it's rubbish. I had this with my own

10:08platform by the way when I first put

10:09Help Bank out. A lot of people like it's

10:11not that great, Simon. And you take it

10:13personally. You know, your personal

10:14brand could be damaged and like that

10:15negative feedback helps you improve the

10:17product.

10:18>> But a lot of people are scared to just

10:19put it out there because of that. Yeah.

10:22I think that's an important lesson for

10:23people to learn cuz that people want

10:25perfection and they won't put it out

10:26until they think it's 100% right. And

10:28you won't know if it's 100% right until

10:30the customer's used it. This is probably

10:31going to be out of date by the time the

10:32podcast comes out, but the last headline

10:34was your company is the fastest growing

10:36revenue ever in history in Europe,

10:39right?

10:40>> Yeah. So the we're the fastest in the

10:42world to reach $und00 million.

10:44>> That is a mental statement when you

10:46think about things like open AI

10:47whichever one

10:48>> fastest growing revenue that that you

10:51know what made you go down the revenue

10:52route so early because sometimes it's oh

10:55let's get users a lot of people

10:56listening like I get users and don't

10:58worry about revenue. So when did you

10:59insert the revenue piece in and and that

11:02must have been quite a big call really.

11:03>> I knew that if you're optimizing if

11:05you're having people pay that's a very

11:08good proof that this is actually good

11:11product good enough product and if

11:12people don't want to pay then maybe you

11:14should be able to create more value for

11:16them and so we we put it in there in the

11:18beginning. I think it's quite it's quite

11:20fortunate for us in a way that the AI is

11:23very costly. So we had to charge for a

11:25significant amount of money so we don't

11:27just lose lose a lot of money on each of

11:29our user and that was a contributing

11:30factor to that uh we made um we we had

11:34to monetize a lot.

11:35>> So how did you decide on the pricing

11:37again people listening how they decide

11:38on pricing on this?

11:38>> Yeah. So the first pricing was it was

11:40actually too cheap and we lost a lot of

11:41money. So that was when we we hadn't

11:43launched publicly yet. Um and then in

11:46the public launch we made it more

11:47expensive and then we were improving the

11:49AI so much that it got it again got too

11:52expensive.

11:52>> Right. So, so we had to uh change the

11:55pricing again and then we improved the

11:58product so that it behaved differently

12:01and we had to change the pricing so that

12:03it was not per message but it's per so

12:06lovable is now an agent that goes out

12:08and takes like decisions and reads the

12:10internet and for no it takes cost tiny

12:12amount tiny amount of credits per each

12:14of those actions we we iterated on that

12:16as well.

12:17>> You went for a monthly subscription

12:18model right? So again like I think Whoop

12:20who I've interviewed the founder of

12:21Whoop and they did a one-year fixed

12:23subscription.

12:24>> Did you did you look at those different

12:25models and and

12:26>> yeah so first of all let me say that

12:28Lovable is completely free to use even

12:32it's not even a free trial. You can just

12:34keep using it for free but it then if

12:36you want to build something serious you

12:37will need to pay to use it more than

12:40just a few iterations with AI and then

12:42the free is still big part of the

12:44>> but the free cost you money when people

12:46use the free.

12:46>> Does we lose money when Yeah. I don't

12:49know if we're allowed to say the numbers

12:50cuz I know the numbers. I'm not sure if

12:51we're allowed to say the numbers so we

12:52can bleep it out, but like

12:54>> 7 million people that will use the free

12:56product. I mean, I'm guessing that's

12:58that's probably 100 million right there

13:00in potential costs. I'm guessing, you

13:01know, probably more, but like for people

13:03to use it for free, right?

13:05>> Yeah.

13:05>> I mean, you don't have to make a lot of

13:07money in the beginning, but it's

13:08worthwhile to think about who are the

13:11ones who are going to give so much value

13:13and who are not so price sensitive. like

13:16pick someone who's not that price

13:18sensitive that are going to be the main

13:20revenue driver and then you can make be

13:23generous to everyone else. So if you

13:25have built a good product, it spreads by

13:27itself.

13:28>> Yeah. We want to help our users build

13:31successful businesses.

13:32>> That's the part of the part of the

13:34>> crucial. Yeah. Yeah. I know that's what

13:35you care about. Another reason I

13:36invested by the way it's a very

13:38competitive world. How do you view

13:40competition? How do you stay ahead? I'm

13:42making sure that Lovable is a customer

13:44focused and a customer obsessed company,

13:47not a competitor focused company. And

13:50that means that we we don't look at

13:53competitors, we look at our customers.

13:54We look at what are they doing? How do

13:56we get them more successful? If we make

13:58them more successful,

14:00>> it's naturally going to be the case that

14:02they tell others that they should be

14:03using Lovable. And that that's how I

14:05think about it.

14:06>> And there's a lot of stories about

14:07Lovable. I'm actually one of them where

14:09we built a product on lovable and we're

14:11making hundreds of thousands of dollars

14:12from it. There's a lot of case studies

14:14coming out like that right now. What

14:15does success mean for you personally and

14:17for the business and and what and people

14:19are thinking about building a business

14:20or have a business out there? How how do

14:21you give them a bit of advice on what

14:23you think success is?

14:24>> I think you should focus on what do you

14:26want to get out of life, right?

14:28>> And for me it's very important that we

14:30have a world where human humans create

14:32things and they just don't mindlessly

14:34consume things.

14:35>> All right, guys. I hope you got

14:36something from this podcast. There'll be

14:37links down below where you can of course

14:39try out Lovable for free. You almost uh

14:41don't realize we're losing money. Uh

14:43Anton's losing money when he let you use

14:45it for free. But you should learn how it

14:46works. Even if you don't right now build

14:48something, just learn how it works.

14:50That's how you stay relevant in a

14:52changing world. Thanks for listening.

14:53We're off to have some chicken.