¶ Building and Scaling Startups
Try to build relentlessly for the same people trying to solve the same problem , and I think that was like a core change , like it was everything that we were building was actually built to solve real problems , which was like 10 times more sustainable .
The thing that would blow the user's mind was actually to have the DM trying to fix it in like five minutes , pushing it And because we were using some like very flexible version next year stuff , it could go very quickly in tradition And like 10 minutes after the send the DM , i was saying , hey , yeah , it's now fixed .
You can see it in the app right now . So I would say , like you have two kind of competitors , like a few of them , that we just help each other , just sharing the contacts and how to just move forward together .
And there was these other categories of competitors like just trying to try to steal your customers , as you are almost out of business for the two days .
All right , man , let's kick off . Well , people are really , really excited for this I've been looking forward to for quite some time and you have quite a story , which is what I really want to get into today and just kind of go through that story and try to unpack it .
But my first point on this is like how does someone go from zero dollars to a multimillion dollar exit in less than two years , 18 months from correct ?
Yeah , exactly Like we did things so , so differently than when we failed our few startups before I'm . So I'm so glad we took this path . I would love forward , I would love to describe how much we learned about it and go for it , Right 100% man .
So you had 11 different ideas that you created over the course of three months , and then four of them had some cash flow , six of them had active users , four cash flow and one took off . Why did you do 11 ? What was that idea ?
So the main thing that we did differently this time is like we went from spending two years on an ID that could be like invalidated , leading to a huge loss of time , to this idea of running of building one new product a week and to just try to validate it as fast as possible And if it doesn't generate revenue , just go to another product or keep it in
groits . And and I think that's what we did for like four months and it led to like 11 products . Most of them didn't regenerate , didn't regenerate any revenue but , yeah , four generated revenue on like three . three of them . it just wasn't enough . It wasn't enough to really prove validation .
and the last one of them was tweet hunter , and from the very day , the very first day , when we got user on board , we just understood that there was something strong here , like the growth was really happening , even when we were not pushing the products on Twitter . So we just knew that we got into something interesting and we just continued .
What were some of the other ideas you had ?
So , since , since the product iteration happened in just one week , we had to jump into very easy to build ID . So one of them was Twitter analytics delivered as an email . Another one was like a community market community database and the search engine that could let you promote your products on so many communities like Slack , reddit , twitter accounts , etc .
Another one was about generating tweets with GPT-3 and , surprisingly at the time , like AI was not good enough to do that .
So like we started generating tweets with this ID , but we actually transitioned to tweet hunter , which was not about AI generating tweets at the very beginning , and we came back to the idea of AI generating tweets much later in the in the life of of Twitter .
So when you were developing that AI model for tweet , under what was it built off ?
The very first idea of Twitter was a huge database of tweets that was categorized , indexed and , and with AI , we were analyzing your Twitter accounts and we were trying to find viral tweets talking about the same topic in pretty much the same way as you do , to just inspire you .
So the basically the output was like a wall of viral tweets and just with this wall , just by looking at it , inspiration were coming and you are writing like 10 times faster . So that's that's how the MVP got super , super small .
That was basically just one input field where you were writing like things like marketing and , and the the engine was retrieving viral tweets talking about marketing , but with like an AI layer that was picking just the right tweets talking about marketing that actually match what you are usually talking about .
And why ? what about chat ? You be the ? was it tree at the time , or was it two that you were building ? Why ? why was not working for tweets at the time ?
That was GPT three . At the time , it was like it was early 2021 and we started doing all that in the in the winter 2021 .
The crazy thing is like at the time , gpt three was like six months old And and I remember that with Tom , we were thinking that we were so late in the GPT adventure , like in the GPT journey So many products were out there , like there was this product headline that got sold even before we began .
We got sold for one million dollar , so for us it was obvious that we missed the train and it was too late . All the AI product have been built and it's so crazy to look at it right now and and realize that it's so not true , and like you have so many opportunities coming right now and and you will have so many more in the next years to come .
Thank you .
What ?
what do you think about that net right now , because I know you tweet about a lot and you had a lot of references recently in terms of like Different applications , like do you think it's gonna start being more integrated into like social platforms that are Twitter , linkedin , like that will be people's writing , or do you think just still gonna be that personal
flavor ?
That's . That's a super good question . I really don't know . It's because you have you have so many Excitements about AI in social and so many people are jumping on AI tools to just Craft better things can be made just videos or or tweets .
At the same time , you have a lot of people and including me That's I'm not really liking , or the AI generated reply that you get on your tweets . So it's it's really mixed feelings and I have no idea how the , the platforms themselves are gonna react to the AI .
¶ Twitter Blue Subscription and AI Improvements
I think I actually think that Twitter did a very nice thing with the Twitter blue subscription , in a way that With eight dollar per month subscription , you cannot just fluid the entire market with Twitter bots Tweeting some generic AI stuff . It's it's just not possible .
So they just by , by making it a little bit premium and with a subscription you they just block an entire AI space , which I think is great .
Yeah , that's that actually is . I didn't even think about that in terms of like Moves , separating out users , just being like . These are the ones that are bullshit , made up stuff . And then you obviously have to premium users .
I know the algorithm was different , waiting on that too , which quite interesting , but some of the stuff that I think about like the AI world is that some of it It isn't like good high quality , like obviously I thought you my backgrounds and my podcasting , so even the video content you can create from it .
Like right now , just just as of right now , in 2023 , it's just not very bespoke , like I wouldn't use it for a client , for instance , and it's slightly off and whatnot . Now , of course , this can massively increase and improve , but as of right now , it doesn't look like it's something that can be used for everyone at every use case .
You know that's when you need to not be an actual robot and be able to actually write your posts or edit software , creat graphics and incorporate your own branding into that .
Yeah , i agree with that . Like it's , it still needs supervision , but I really think that it's because we We just we suck at prompting , not in a way that you will . You will find these magic prompts on on Twitter on one of the hundred threads talking about the The bed prompts .
That's that we changed the game , but more in a way that's To make the AI very good . You need , you need a very unique , huge prompt Telling the AI exactly what you are looking for , and and nobody is doing the effort Needed for you to do that , and that's that's why Twitter and tap you our products are approving .
Some value is that we are doing this , this work for you of Trying to understand what what your tweets are usually about , trying to grab the tone , trying to grab the topics , and and we , we build , like this , very huge prompts that we send to 2dpd and With that you get much , much better , better than what you would have .
So so is top you and tweet under now , based on GPT .
Yeah , they were not at the very beginning They like . Surprisingly , when they rolled out GPT 3 , open AI prevented all the use case related to social media and My guess was that they were afraid that it would just flutes Or social media and that people would just I use them to to do harms on social media .
The way they function their models , i think it leads to the model being like much , much safer , so they now feel confident to use . That's interesting , man .
That's interesting . Yeah , it's been a little while story . That's . That's reason I want to get into some of us . So I want to get into some of like the actual funding behind it . So correct from wrong , but our all your ideas self-funded and you didn't . You never actually used VC money .
Actually , like we did for our first startups , like the one that didn't work , yeah , first , like you have , the thing is in France I don't know if it's the case everywhere , but I would assume it may be the case in the US Like at school , at university , people are telling you that this is the way , basically , like you have to hire victims .
You have to raise money and that's and that's how you would make your parents happy , basically , like by by being important , by having a big team And so , and so that's exactly what we did , like a right out of of university , we , we raise money , we try to hire the biggest team ever , and and we failed miserably because it's like , when you have a big team
and you don't have product market fits , it's , it's a nightmare to pivot your company and to onboard your entire team in these people because , like , basically , everyone in the team will think that you have a no vision , you don't know what you what , what you want to do , what you are building , and and you will lose the motivation of everyone in the team .
So , and and I think I need this too big failures to learn that . And in this new adventure , when we started about two years ago , right from the beginning , we , with Tom , my co-founder , we decided to start with a new adventure , and we started with a new adventure .
We decided , like , not to raise money , not to hire anyone which we had to change after that but but that and and also to like , try to build relentlessly for the same people , trying to solve the same problem , and I think that was like a core change , like it was everything that we were building was actually built to solve real problems , which was like 10
times more sustainable .
I guess , yeah , and on that basically you're looking for like real pain , and I often see you write about that like actual pain problems . How important do you think that is right now ? Because , of course , like people are not stricter with their money , they're not tighter versus building , just like , versus , like you know once , versus needs essentially .
¶ Working on Your Own Problems
Yeah , the thing is like when you are not working on something and when you , when you are like thinking about problems that you haven't really worked on , you will come up with some very interesting ideas that sound very interesting but actually are not good business ideas .
And that's because , like , you don't have industry knowledge , you don't have the core skills to understand the dynamic of the idea . And what I realized is that you need to start working on something .
You need to , like , get your hands dirty to really understand the core problem of the users , to then work on that and provide a solution that's actually worth paying for .
How do you go through that process ? Yeah , is it because , like you're the target market , you're the audience ? are you trying to solve your own problems ?
Yeah , and that's , i think that's it's like it's a huge competitive advantage to work on your own problems . Like the first startup that I built was about motivating your kids to do their daily chores Because I , like , i love the , i love the education and working with kids .
And the thing is , at the time , i was not a parent , so I was really not understanding the core issues of the parents , and I mean by that , like the struggle of the guilt and how you want to spend time with the kids but at the same time , you want to have your own time . So I wasn't understanding that .
And when we started working on Twitter with Toma , i was actually trying to build my own Twitter audience because I realized that Twitter is one of the best drivers for growth for all our products .
So I , as I was trying to build my Twitter , i understood the problem that I had growing this audience And that's how Twitter is born trying to solve my own problem in the first place .
Man . That's . That's that's so interesting because , like so many people are hopping on like different trends and they're moving between niches just because they hear that's what meant to be doing , like the web 3.0 , metaverse guys and now the AI guys but they don't really understand the core fundamentals of what they're doing .
You know , i think that's why I've kind of had moderate like success in the podcast world is because I've just done this thing for so many years that now at this point , i kind of know what the ins and outs of like how I should be operating on a daily basis And I'm trying to , i suppose , serve in a service business other people that do the same stuff as
me . You know .
Yeah , yeah , i understand that , Like the thing is , whatever you do , even in podcasting , you will need to do it for a long time . So if you , if you are not working on your own problems or something that you are really interested in , it will not last , Like you will .
You will jump on the hype working for three months and then just give up and move to another thing And you don't have this . You don't have enough time in three months to really understand the core problems on industry and come up with like original solutions .
And from your perspective , are you working on the development side as well as the marketing ?
Yeah , Yeah , that's , that's something that's quite weird , is that ? so , between my corporate and I , i am the one who has a Twitter audience , so I was the one like doing all the push for the products and until the very late in the developments , or like , the support button inside Twitter was actually directing people to my Twitter DMs .
Therapy And I think that's like that's a super powerful tool , is like , yeah , every day , all day long , i was getting DMs about people complaining about stuff And and I had like Twitter open and my coding and code development window open . At the same time , i was actually fixing and modifying things in real time . No way .
The thing that was that would blow the user's mind was actually to have the DM trying to fix it in like five minutes , pushing it And and because we were using some like very flexible , very short next year stuff , it could go very quickly in pollution and like 10 minutes after they send the DM , i was telling you hey , yeah , it's now fixed , you can see it
in the app right now . And we got a lot of tweets from user talking about how fast we were , which was like a very nice marketing boost at the very beginning , and it's definitely not the case anymore right now , because the team has expanded and now we have , like , some kind of processes in place .
But yeah , that was fun . Man , that's wild because basically , like you're also learning what you should build , because you're getting all the feedback . The feedback loop is so positive .
You know , like I just keep comparing that to larger companies that I've worked , as I mentioned , and like you'd run it like an NPS score , which is like a net performer score , like once , once every Jesus Christ man , once every six months , and then you look at a list and be like , okay , we'll do this next quarter , this next quarter , this next quarter .
Like the rate of that you're learning and that you're solving your own problems is actually like crazy at that period . So that's very interesting , though . Like because how did you have that bridge between marketing and development ?
Because usually you'll have someone like me who can't develop , like shit , and then someone who's a developer , who doesn't understand the world outside it .
Yeah , and I think , like we are experiencing right now a total change in , like , the way we can build things . Like you have a lot of indie makers right now that are just building very cool stuff And they are talking about everything they do on Twitter or LinkedIn And just by doing that , they are building this unique relationship with their audience .
The product doesn't have to be the best . The audience is having these emotional engagements toward the creators that will make them buy your products , which is why I strongly believe , like we , it's obvious to everyone that Instagram and other social media has totally changed the way people buy consumer products .
What was not obvious was like Twitter and social in general could really change the way people buy B2B products . And it's happening right now Like people are taking decisions on which tools they are going to use based on who they follow on Twitter and LinkedIn . That's crazy . It's happening .
That is wild , but because obviously I've seen it from an info product perspective , like a course . Obviously I guess I was a product too , but courses of it are quite obvious . But I bet , like Twitter and LinkedIn , were not built with those intentions .
It's like how the user interprets the technology , whereas like Instagram , for an opposite , like actually put in a boy button , they actually put in these suggestions to build brands Because , like it's kind of like Twitter and LinkedIn , of course , grew these types of models organically And it was like , oh , like users like yourself are building products that we need
to integrate with them . Question for you , off that is how do you integrate with a platform like LinkedIn , like is the API open ? Are you able to do that ? I didn't realize it before I met Taplio .
That's a quite sensitive question because , like we have a , we have , so they provide an API , but this API is really not as advanced as the Twitter one . So , like , for example , with the LinkedIn API , you cannot retrieve your posts , your LinkedIn posts , so we have no way to build an analytics system .
So we had to bypass that , and what we did is that we built this Tapio Chrome extension That's actually using the LinkedIn internal API And thanks to that , we are able to get all the data that we would not get from the official API . Interesting man , which is like a much harder technical change . But that's also why the competition is so low on LinkedIn .
So , on Twitter , before the entire Elon Musk sheet and like a premium API , you had a lot of Twitter scheduler , like a lot of small tools to compete with Twitter . Linkedin is nothing just because the technical is behind it much , much harder .
I kind of know that . I thought about that because it was quite interesting when I saw all those tools for Twitter , because , by the way , i'm not I'm like quite new to Twitter . I've only started using it like literally really shortly , but I've always used LinkedIn . But when I'm using , i was using Lemlist for sales and marketing and all that shit .
And Lemlist has ability to scrape like details of people who are sending emails to and how to do . That is true , i think it's true . A Chrome extension as well as something similar . So if you're looking for the name and the job and all that kind of stuff , it's from Lemlist . In that , in that , in that Chrome extension .
I have to tell me I don't know , like I don't know exactly how they are building this system . We are the . The nice thing is , like we are the very , very beginning of integrating Tapio and Lemlist . There's just so many things to do Like the , you can just like .
What we have built in Tapio that we really like is the ability to get to build the list of all people who comments on your LinkedIn posts And then to take this list , find their email addresses and send it to Lemlist to then outreach on those people , because because this is not cold emailing anymore , like it's , it's almost like warm emailing , because people kind
of know you .
Yeah , 100% . Yeah , that's really powerful , especially like in the service business space whereby , like , one client can be 12 months of work .
So if you were able to have like that like profile to actually be able to get that list is actually incredibly valuable , because how people derive lists are like you know , they're crazy to build these lists in the first place . You know , i want to ask you about that Elon Musk premium API stuff . How does that unfold for you , like what was the process ?
That was a nightmare Like . So they did their entire takeover .
So we know a lot about how the employees got fires , but what most people don't know about is that very recently like I think it was early late March they announced that in the next 30 days they would transition their API from free to an enterprise an enterprise API , and that was like a black box without a very clear communication enterprise And so we just
jumped on that . The Twitter API was the number one core component of Twitter , so we just had to move to this enterprise tier . The crazy thing is that we're not paying anything at the time And Twitter and we heard from other people that the first tier was 42,000 per month , so this it was just creates a cost of 42,000 per month .
But the craziest thing is that the the business transition in 30 days , as the announce just four days after the announce that they started to cut off accesses to a lot of developers , including us , so a few of our competitors not all , we don't understand why and a lot of other products and also very big one like what press lost its access .
We just we just lost total access to Twitter , so user could not post using Twitter , we could not retrieve tweets and we had many , many other blockages , and what we experienced was like a like a huge urgency to trying to bypass all these blockages And , at the same time , trying to organize them ourselves with , with other Twitter products , founders , trying to find
a solution , and 48 hours after that , we finally got a response from one of the 10 people that we emailed at Twitter . The crazy thing is that all the people we knew got fired , so we had new , new contacts at Twitter . One of them replied and we and she facilitated the transition to the enterprise theater .
So we are now paying 42 per month 42K per month , oh my God .
¶ Building Business
And was was there loads of like controversy between you and , as it was a hype fury as well , they were like trying to like take users from people They were trying to interview .
And they did So the the Twitter contacts that unlocked the entire situation . it was coming from one of our competitors . So I would say , like you have two kind of competitors , like a few of them , that we just help each other , just sharing the contacts and how to just move forward together .
And there was these other categories of competitors that just trying to steal your customers as you are almost out of business for the two days . But , yeah , i didn't really appreciate what they did , but it's it's business . Yeah , they can do what they want .
That's the reality of the world .
Yeah , exactly .
Oh , that is insane . And so how did you feel during that period ? Like , like , what were you doing , what were you working ? You said you were in Indonesia at that time . So , like , what was your day like ?
It's like extremely focused . I remember that I ended the day . I thought that my head would explode , but extremely focused , trying to find solutions , talking to everyone , building , building crappy things to bypass the brocages .
That is insane , man . Oh my God . I want to get into some of the area around you , bringing in more people in your team . So you brought on JK Malina as , like a like an influencer co . -founder . What was the ? what was the idea behind that ? I'm quite unaware of that . Like decision , like , is that common in sauce ?
So it's really not common And that's really not intentional in the way that I think it was like a one month , two months after we had the first users of Twitter but we started to DM some big Twitter guys , like Twitter influencers , just in the hope that they would use Twitter and then then talk about it . And there was this guy , jk Malina .
He was not that big at the time , like something like four , four hours on Twitter , but he had this , this extremely weird reaction where when I DM him , he said that's crazy , i want to be in And that is . I was super surprising , like we never like , and I didn't even understand what it meant at the time .
So we we had a few calls with him and he basically wanted to be in equity because Twitter and the way it worked at the time like the way like we were building and displaying this wall of viral tweets was exactly the , the copywriting method that he would promote in his Twitter course , exactly the same thing .
So he was like let's partner , let's work together , and I will . I will promote this thing every day on my Twitter And the thing is for us like we didn't want to have to give equity to an influencer that would just promote a few times the product and then nothing .
So we basically build this contract saying that JK Malina would take the CMO position and would need to behave like , like , like a CMO .
So , yes , he would use his audience and he would promote Twitter to his audience , but he would also work and activates many other marketing channels , including like Affiates , seo , ads and other things , and it worked very well .
How did you , how did you say yes , like he was probably like 19 at the time ?
That's true , i think 20 , 20 , but yeah , we said yes because , compared to now , our ambitions were much lower And the goal was to reach 10 K MRO by the end of 2021 . And I think we did like five time . That's at the end of the first year . But , yeah , we thought that we would partner up , reach some decent MRO and maybe move on .
We didn't really expect the huge growth of Twitter That's Mar Very rapidly , we realized that we were pretty much right about trusting him . Like just , he organized and operated a Twitter launch on September 2021 . And just with this launch that he executed , we went from 3K to 20 K MRO in just three weeks . So , yeah , he , just he , he worked very well .
I heard was there some kickback from that ? Was there some negative feedback ? when he launched , you were like Oh , i want , don't want a tweet , i want to write my own tweets .
Yeah , it happened . The weird thing is that it happened at a very low volume and low frequency . Like I , there was a bunch of guys that we like . One of them was Peter Levers I don't know if you've heard the theme . Like Levers you on Twitter .
He told me at the very beginning that we would get canceled by by Twitter people because AI tweet was not something that people wanted , and the thing is it never happened . Like there was tweets about blaming us for AI tweets on Twitter , but it always stayed quite low And I really believe that's because people really don't care .
In the end , if the tweet is good , you don't care if it's AI , it's generated or not , and in most cases , very good tweets are sometimes generated by AI but fine tuned by humans .
That you like 100% touch , and you take it from other people as an inspiration , even if it's subconscious , not even not even consciously . Writing structure , tonality , like sentence structure , is literally inherited by other people And if you don't read that frequently , you're going to inherit it from people that are on your timeline .
Yeah , that's . that's what we do for like thousands of years , like we just take inspiration from others , spin it a little bit differently , add a new thoughts on the content and release it by yourself . Like true original content is insanely rare .
Yeah , and there's only like a few people in that . I even know who write truly or like originally . Like you know , yeah , it's just it's just , it's just so like rare at this stage , you know , because of course , you need to have your own abstract thinking for it . I want to ask you about that sales process . So how does that operate ?
Like you know , you were mentioned that how you built it up and you wanted to sell it at like 10 K M , or , or is that always your goal of sauce is just to build and sell , Or do people want to build and grow and keep on growing it ?
That was the initial intention because , like we really enjoyed these early stages of building , the idea was like trying to reach 10 K M or look for a 10 times multiple And so sell for a million and just sort over .
But the thing is , what we didn't realize at the time is that when you reach 10 K M or it's , it's much , much easier to go from 10 K M or to 20 K M or 20 K M more , then start over and go from 0 to 10 .
And and valuation are also increasing , like it's it's hard to get the 10 times multiple on 10 KM , or it's easier when you have a 100 KM or I was going to say the .
the multiplier was was it by two or by one of a ? or Yeah for tweet under just tweet ? under what was the ? how did that function ?
So it's hard to answer because , like in it's , in my head it's more like 10 than one .
But when M or of AR .
¶ Building and Selling Startups
Okay , okay , sorry , sorry , because we , when we saws we were a little bit below two millions in annual revenue And and we saws for an amount that could go above 10 million , if we were correctly , in the next few months and year .
But the thing is like the business is so risky and it's so platform dependent , that's , the acquisition has been built around a very big earn outs , meaning that the initial amounts that we received was quite small . And then there are like six different milestones that are revenue milestones That will make us unlock a new , a new part of the acquisition sale .
How does that work in terms of like a structuring , like how come you pick that model ? Is that just common in the industry ?
It's pretty common to have a large earn outs . It's it's common to have like 50% of the of the amounts to be unlocked in the next two years . It's it's not that common to have this amount . That is is correlated to the growth of the products and in our case , to huge growth .
Like we were below 2 million in annual revenue And and our goal to get the total amounts of the sale is to go to 10 million in annual revenue . So from below 2 to 10 , that's that's huge .
But right now .
I'm super happy that we are on track to reach the 10 and hope that it continues .
Yeah , because I was so intrigued to understand that , because I know you like sold on paper . But then I thought to myself well , with these payouts and when all the changes in Twitter , that's probably heavily influenced , like the outcome of what's actually going to happen . So it's like there's lots of different checkpoints , i guess at that point .
Yeah , yeah , and setting was not an easy decision . Like it was , the growth was so strong that you can think that you get . you can get more money but just staying in business . But it's very hard to to like to accept this very high level of risk when you are just two individuals running running this company .
That is wild , that is unbelievable . Talk me through , like the lesson in there . Like , so , like you know for people . like when do you come to these decisions to sell ? Like when do you look to sell a company ? Like what have you learned over the years ? or not ?
In the setting or in the building .
I would say like a combination , like when you're building and you build it and then like what's ? how should you frame it mentally to build something and then sell it ?
The thing is like I think right now that's picking the correct industry and picking the right space is going to be is going to be a contributor for most of your success .
Right now , like we picked social media and AI and social media is not it's actually quite hard Like social media and the creator industry was super hot when we started working and AI had become insanely more hot right now . So we just we just picked the perfect growing industry leading to we don't have to do crazy , crazy tech to have big growth .
So I really think that be careful and be like just just right now . I would advise anyone doing AI , for example , because AI is growing so fast that you have so many things to build just pick the right industry . The next thing is I would really advise to not hire to get your hands dirty and to build yourself and try to build for yourself too , in a way .
That's , if you are the first customer and if you are building something for you , it's very likely that tons of other people are having the same problem than the one you have And if you are an expert of this problem , you deeply understand how to solve it And what is value and what is just nice to have a quite useless feature .
That's very interesting , man , and how did you get that across , that message across with a tweet under ? Because , like , of course it was like a low subscription , but people might think , oh , like , i don't need this .
I could write my own tweet , like how did you get that message across that , oh , like , you need this and you can use it for your business or your brand ?
Something that really helped with the growth was like at the very beginning , when we started working on Twitter , on every tweet , you had the tool that was used to tweet the tweets .
So , because we had a very premium product , because it was expensive and very useful and deep in AI , we managed to get a lot of big tweet players And so that's a lot of people doing very viral tweets , and so we got our names , that Twitter , written under the tweets , and so people just thought like , yes , so if he's using Twitter , i should use it too .
100% . That's that influence ? Yeah , but Twitter removed that , So right now we cannot use that anymore .
That's mad , man . That's mad . But at least you got the growth when you needed it , because you didn't have the brand at the time . Yeah , that's it .
And there was something that just building in public , like sharing all the steps , sharing all the learnings , is really something that helped us grow . So we also try to . I don't know if you saw that , but we build tons and tons of free mini tools that have the purpose of serving one specific need .
So , for example , we build one that is about analyzing your Twitter audience and trying to define the best time for you to tweet when your audience is the most awake . That's something that we just shipped , it's free , it's out there .
It's something that people are writing on Google , like what's the best time to tweet , and so , since it's not static websites , it's like a nap and providing a personalized answer . It's ranking very well . So we just all the I think the 20 or 30 free tools that we created are ranking better and better every day and are bringing a huge traffic right now .
That's awesome , man . I want to ask you on this about like . so of course , that was a great success story like start to finish , But when do you cut ideas ? How do you identify that ?
You mean when I come up with ideas .
You know in general , like if you're building when you you know you've built so many startups over the years like when do you decide like enough is enough , like this isn't taking traction , this isn't growing and when you should cut something down and give up ?
So I would , i would like spin it in the other way , the way that when it's not working , you never know if it might work one day , but when it's working , you know . You know it's so much that it becomes obvious and you start understanding that all the other ideas that you had before was actually garbage .
What I mean by that is that there was this YC guy talking about product market fits .
In a way , that's when you , when you get product market fits and like true product market fits , it's , it's painful , like it's super painful because people , because people desperately want what you build in a way that will make your servers crash , will make you die from the loss of your , from the load of work , et cetera .
So , basically , when you know , you know , like when it happens , you know , when it doesn't happen , you never know .
That's super interesting , man . That's so like different compared to like why you like normally here . You know what I mean .
¶ AI and Serverless Impact on Development
On your own personal note , like what is it about ? like building that you love .
I have to like . I tend to think that , like , to be successful , you need to do things for a long time , and so you need to , you need to love it . It's like I really I guess you like podcasting And and yeah , i'm like I just I just read what I love . Was that like ? was that your question ?
No , no , you're completely , you're completely right . But like was it always kind of like instilling when you were young , like where you always kind of like into like hardware , software , quite young , and then you kind of built up from there .
Yeah , but I actually got into programming and when I got out of college I was not really good at programming And so I just like I think , released start-up . Being creating startups is one of the best way to learn hard skills Programming . You just got get into it And then you have to .
You have to solve your problems and there is just no one around it around you To help you . So you just need to learn and right now it's just crazy how , how many capabilities you can have . We just charge it to you . I use GPT for to write . Most of the codes are right right now and it's it's crazy good , it's crazy fast .
You can write your phone , yeah , and you can actually ask it to explain you how it's working And why the code is working or why it's not working .
It's like It's for me , it's like it's the end of the old generation of coders is , if you master that , and because the old generation will not Do the effort of understanding it , you can just like How to run them in a few years .
Holy shit . So so taught me true , like that process . So would you just have like a like a bug or something you can't fix , and then you would run the code to GPT for and then like , what would a point out solutions ?
yeah , and then What I do right now is like when I start a project is like I build this prompts about explaining , explain , gpt . What's the project about ? What are the main technologies , what's what is the main ?
like a file , file hierarchy in the project , and Then I reuse these parts every time and when I have a specific problem , i ask my specific question and it generally comes out a very neat solution That's perfectly on point . It's . It's crazy and You can just ask it to like a right entire scripts and tie your file . It's . It's insane .
I really think that the only thing preventing GPT for to writing full application , full , fully working application , just based on a on a prompts , is The prompt length . Like it's . It cannot do that right now because the prompt length is too small , so it cannot assimilate an entire project in the prompts once it it goes above a certain size .
But it's really something that I think it's gonna be Doable in the next month , maybe yours .
Next month , man , that's so . It's so funny how fast it moves , you know , you know . So obviously there was like a shit on a tech layoffs , like loads people , a cluster job , it's like pretty much across like the board , and it kind of opened up the world to maybe like Companies were over , you know , overused on resources to put too much money into it .
Do you think from an engineering perspective , this will have a big impact in terms of we only need one engineer that can produce for engineers work because you can utilize AI .
Yeah but even more than you think , in the way that's like all the lay of that daddy , like a Google , meta and you're the guy . They have like 15 to 20 percent of the stuff . Just just look at what Twitter did . It's , it's like it's it's 80% and it's running and it's actually and they are actually shipping more features And they ever did so .
Like , what does it mean ? like , does it mean that those big companies can be run with With 70 plus 75% less people ? Does it mean that's if , if that's the case , and if AI is coming ? because when Twitter did that chat , gpt was not here And GPT for was not here .
So , with GPT for and with the example of Twitter , i'm pretty sure it's gonna be Harder and harder and it's gonna happen more and I'm pretty sure , like meta and Google are Doing this like this 15 70% layoff , because they are just afraid of doing more .
But they would , once that you want to do more , yeah , if they had their way , they would rebuild the company again .
They'd have like less middle management , less less managers here , like less people and like even like HR and stuff , because you could . Yeah , all of it No you know , so if they had they they're pretty much overburdened as a result at the moment .
Operationally , yeah , and , and it's not only about AI , like what's Twitter and tapio are builds on the technology that is called serverless , which which basically is Is we are Running all our back end architectures without servers .
It's it's it's clouds , but it's it's like clouds 2.0 , in a way that we don't manage servers , so we don't manage loads , we don't manage complicated things , and so what we did This year Just me at the very beginning would have required like 10 to 20 people 10 years ago . It's , it's insane , oh my god . And I'm not talking about AI here .
So , yeah , i truly think that we are living in the best time ever for developers and makers to build very complicated things , very complicated things , alone or in with the very , very small teams , that is is that .
Would that be the same for mobile apps that your serverless ? Is that possible ?
like both mobile apps and SAS in in . In most cases , they require servers and codes running on servers . So , yes , all this code running on the server can be Can run on serverless technologies . That's We not require any servers and saving so much time for your developers .
Man , that is literally below on my mind . It's so . It's so crazy how , like there's this opportunity But some people would overlook it , and especially , especially companies .
You know , like one of my clients , if I'm just quite interested , when my podcast clients met me , i really go point saying that like It's , we have all these resources out here now summer , and not even AI , just like automation , ai or like any other way to streamline operations , but most people have their head in the sand looking at it and it will take them
24 months , 36 months to catch up to Any sort of speed , whereas if people like you , who are literally in the Alps right now , that's just churning out these features . You know , because you know what's available to you . I .
I truly think that we are going to leave the biggest rotation of of Software's and companies in the world like meaning , meaning .
Tons of companies are going to die And tons of new companies are going to be boring and right 100% , and I think also people can actually utilize it in a different way . Um , i want to finish off with like your next , your next steps , like what's your , what's your next goal Now ? what are you looking to do next ?
I have no idea . So , like right now in the in the short term , uh , in the short term , we work with the empire to grow twitons and tapio to 10 million in annual revenue . Um , and I think it's . We are totally on track for that . Uh , especially on , tapio is growing so fast Tons of people are jumping on it .
¶ Exploring AI and Building Businesses
After that , i don't know like I really feel this ai uh formal , like I want to build so many ai products right now . Uh , i have . No , i have no idea if , if , uh , i will , i will do it With the empire as an employer or if I will just continue to do it myself . I don't know .
Yeah , man , you have plenty of time to discuss those things , of course , you know , but it's just interesting to see how your , like , your brain works .
Like that's why I enjoy these conversations so much , because , like You have so many ideas , you can patch them together and then you can build them , which is such a dangerous combination To be able to go after those things , which is awesome . That's , that's the gap for a lot of people that you've already filled for yourself . Yep , what about you ?
Well , the idea , man , is that I'm I'm Heavy building this agency full time , but I'm utilizing everything around me , you know . So I have a very small team . I only have two people with me . Um , what's on video production and ones on graphics , and we're just looking for , you know , bigger clients , bigger company clients and Really into kind of b2b space .
So that's , that's the goal , you know , just head down on that for the next next couple years . You know , i don't have any . Uh , i don't try to set my focus , i just try to stay Within my lane , to try to keep building . You know , nice , good , nice man , do you want to say a massive thank you ? I appreciate this .
Um , of course , like there's so much different things you get , get , get , go back and forward on . But uh , i want to say a massive thank you , man .
I think you very much for having me . Was was very pleasure man .