Buildspace - live w/ Sharif from Lexica | from Zero to 1 million users in one month

Hey, hey, hey. If you can hear me and see me all right, press 1. Sorry, we were muted there.

Okay, nice.

Boom, boom, boom, boom, boom, boom, boom. Love it. How's it going, everybody? 300 people here today, Sharif.

Impressive. Wow.

Yeah, welcome to Nights and Weekends, uh, week 3, talk number— talk number 2, right? Talk number 2. Yeah, um, so, um, just a Sharif drip unmatched. By the way, how is it Does anybody know Sharif? From like zero, you don't know Sharif, to ten, you know Sharif. How much do you know this guy? And it's fine if you're zero. One. All right, we can just even just look at the chat here.

Nice.

One zero zero zero. Most people don't know anything about you. It's amazing. You're just a nobody.

A whole lot of blank slates for me to leave my impression on.

Dude, nobody knows you. You're you're gonna die unknown.

I see a four.

That's pretty good. Oh fuck a four. That's good. Okay, so let's just kind of get right into it. I'm gonna test this other camera really fast. Sorry for testing live. But boom.

Wow.

And yeah, let me know if we sound good on this too. Press 1, please. If you see Sharif and Sharif, you want to say something?

What's up?

What's up? What's up?

How's it going, everyone?

Amazing. Okay, cool. All right, Sharif, we're working. Everything's perfect. We tested it live and let's get into it. So kind of set the scene for you and to set the scene for everybody here. Welcome to the week 3 of Nights and Weekends. Everybody now has an idea. Everybody has like a goal they're running towards for Demo Day, which is on January 28th. And everybody also has a, at this point, a link that they can actually send to a person with their thing, no matter how simple that link may be. So that's kind of where you all are at at this point. And everybody should be at this point off localhost. Press, uh, press 1 if you're off localhost, press 2 if you're like still on the struggle bus. I forgive you. But at this point you should be off localhost. Cuz that was the, that was a requirement for this week. So, okay. Everybody's off localhost.

Nice.

Everybody has something.

No more localhost:3000.

No more localhost:3000. It's amazing.

They're on their, uh, .versel.app domains now.

Okay. Someone says I'm a Sharif fanboy. Aperture is wild.

Oh, nice.

So, so I guess we could kind of just jump into it and, um, Just gonna go a little bit further. You all now are off localhost. You have an idea and you have a goal. And we have Sharif here. And again, the point of these talks is to, I don't know, it's really at the core just to get you a little bit excited about your week. 'Cause I think it's like pretty hard to keep on finding motivation within yourself. So these stories will always help you out. So hopefully this helps out too. But Sharif, do you wanna maybe, Introduce yourself and tell the people who you are for the people who are at zero right now.

Yeah, so my name is Sharif Shameem. I'm a founder of a company called Lexica, and we're building a suite of creative tools built on top of generative models like Stable Diffusion. Before that, I worked on another company called Debuild, where we were using GPT-3 to kind of build full-stack web apps, where the models would generate the code and you would just come in with an idea. And before that, I worked on another company called VectorDash, where first we were kind of like a GPU cloud provider, but instead of running our own machines, we had distributed network of hosts. And then after that, we turned that into a cloud gaming company where you could stream video games from GPUs within your neighborhood. So yeah, I did work on a whole bunch of stuff over the years. But yeah, Lexica seems to be working out pretty well, which is nice.

Yeah, sick. You can also— if you could talk like right here.

Sure.

Just go ASMR mode.

Oh man.

Just like go like test, test into the mic.

Let's go.

There you go. You should be fine now, everybody. He's gonna ASMR you. But yeah, basically Sharif is One of my first buddies in startup land, and he went from zero in college to dropping out to going through YC to living on Furkan's couch to to just making a bunch of shit. A lot of it, a lot of it failed. A lot of it, a lot of it did well, right? So I guess we'll start with where things did well, you know. So we can talk about Lexica. So just to kind of set the scene for everybody, Lexica is like two months old or something like that, maybe three months.

About like 2 and a half, 3 months old, I think.

And Lexica started actually at a hackathon that me and you were at with 6 other people on a Saturday on the weekend. You were working on your own company, a different company. You were working on Debuild, which was like a no-code builder with GPT-3. And then you, one day on this hackathon, I think I was working on this thing where you could bring people back to life with GPT-3, and you were building this Lexica thing, and you were scraping the Stable Diffusion Discord for all, all the, all the, all the images. But yeah, maybe you could just start there. Like how this thing started from like a, a fucking hackathon on the weekend, random one with me and you with 6 people.

Yeah, that's a good, yeah, that's a good story. Uh, so I, I was one of the very first users of, uh, DALL-E 2. I was like pretty lucky to get off the waitlist pretty early on and I was pretty mind blown that we have this technology that can pretty much turn words into images, which is kind of mind blowing. Like we take it for granted now, or most people do at least. But like a year ago, this was like pretty insane. Like I still remember seeing the first DALL-E 1 demo where they had like the avocado chair. And that was just like, what? Like you woke up that morning, you saw that like tweet and you like went to the OpenAI blog post and it was like, we can do this now? That's crazy. So I used DALL-E 2 for a bit and it was pretty cool, but it being a closed source model, there's only so much you can do with it as a developer.

Yeah.

They didn't even have an API at the time. They do now though. So a few months later, Stable Diffusion was being trained and they had like a Discord bot where you could go into the Stable Diffusion Discord and reply to the bot with a prompt and it would generate like 9 images for you.

Yeah.

And I spent like, when I discovered that Discord, I spent like 4 hours a day, maybe like 10 hours a day at some point, just like every single day, just like generating images. It was like so much fun. It's like the first time humans have ever been able to create images without much effort. It's kind of mind-blowing. But the thing that was really bothering me was all these prompts were like really large and convoluted. Like, it was never just like, show me a picture of a cat. It was like, show me a picture of a cat, fluffy fur, gorgeous, Unreal Engine, trending on ArtStation.

Yeah.

And I was like, what is this? It doesn't make any sense to me. Why can't you just say a picture of a cat? Then I realized, oh, these like tags to your prompt make your images look better. It was pretty cool. But I was like, wait, does this actually make it better or are people just copying each other? So I would like look at the prompts that were the generations that were like flying through the Discord at like 100 images a second. And there's really no way to like do it unless you like generated a bunch of images and like there was no scientific approach to it. And the Discord search bar was like pretty bad. So you could search on the sidebar, but it showed you like 5 search results at a time or something like that.

I guess, why did you even wanna, why did you wanna search images?

I wanted to learn how to create really good images and I wanted to learn how to craft my prompts to create really good images.

Yeah.

So I wanted to like study other people's images.

And so I guess like, why were you even doing this when you were working on your own little company and like at the time, right?

Yeah.

Like, was this, yeah, like what made you wanna wake up that day and just do this? 'Cause you had your own company, you were YC funded like millions of dollars, but then here you were working on this like search engine. I think most people would find that crazy, right?

So yeah, it wasn't a search engine yet. It was more so just me like trying to be curious about these new models. Like I, I was just mostly curious about them. And the prompt thing also came out of curiosity. I was like, how can I make my prompts better? Yeah. I wasn't really thinking about how to start a company at the time. That didn't even come until much, much later. So it was like figuring out how to create these really, really good images. And I was like, the Discord search bar sucks. I can definitely make like a much better search engine for this. Just like literally just the images so I can use it for myself. And I was like, oh, you can like inspect the Discord search endpoint and you can actually see the query being made and you can just like call it like 50 times a second or something like that. It's like, cool, I'll just do that. create like a small collection of images and just like search it myself.

Yeah, I think you like had Discord. It was something like where you hosted Discord in like a VM, right?

Yeah.

So you had like 10 instances of Discord and you were scraping the Stable Diffusion Discord.

Yeah. So I used a project, this open source project. I think it was like called Discord Explorer or something like that. But it was really slow. So I actually had to write my own Discord scraper, which was pretty straightforward. It wasn't too bad. And then just use that for a bit. And then I think we got all the images in like a week. So that was pretty cool.

Yeah. How many images was it the first week?

First, I think it was like 5 million images for the first— yeah. So I think it was about 5 million images for the first week.

Yeah.

And so I realized that, oh, cool, I can build like a really basic UI around this now that I've scraped all these images. And I can just make it so that you can just search any word you want and see what those images look like. So I could search for like the term Unreal Engine and it would just return all the images that had that word in the prompt. And I could see what that word actually modified in terms of like what the images look like.

Okay. So you, yeah, I guess, did it feel weird for you to like just start working on this on the side?

No, it was a hackathon. It was a hackathon project. So I was like, oh, this is cool. It's like useful and I could probably use it. And I thought it was like a, It's like mostly trying to fulfill my intellectual curiosity than anything.

Yeah. So you ended up like making this crazy scraper, getting 5 million images. What happened next?

So I think this is where I got really lucky. So I built this search engine. I called it Lexan because I thought the name was really cool. And Lexicon was like definitely taken.

Yeah, I remember you were showing me the fonts. You were trying to pick a name. And you were like, what looks best in this font? Yeah, yeah, yeah. And you were like, and you showed me, you were like, Lexica. And I was like, oh, Lexica, sick name.

Oh, nice. Yeah, yeah, I remember that. That's pretty cool. So I bought like the lexica.art domain for like $10 or something. I was like, cool, this is great. There's like one other Steam game called Lexica, but it has like 20 reviews. I was like, all right, we're okay. We can definitely overtake the Lexica game. So I built this search engine and on Monday, so this was, I think in October, I think this was October, Monday, Stable Diffusion's weights were open sourced.

Yeah.

This was the very first time a large image generation model had its weights open sourced. Um, and then on Tuesday, I had been working on Lexica for like the past week at this point. And on Tuesday I launched Lexica and I timed it like almost perfectly.

Yeah.

And it kind of just like the zeitgeist of Stable Diffusion, we really just like rode the wave. Lexica was just like a surfboard on the wave of Stable Diffusion.

Yeah.

So that's where we got really lucky.

Yeah. I feel like it's like worth summarizing just that, 'cause it just, it just sounds ridiculous just for everybody. It's like, Saturday, me and you are, I'm working on something, we have Stuvan there, me, Stuvan, and you, and we're like, oh, Sharif's working on this thing, and you're just scraping the Discord, right? 'Cause Stable Diffusion had come out like a week prior. And then that's Saturday, and you're just starting. Sunday, it's like, yo, does Lexica sound good? And then I think like a couple days later, it's like live.

Yeah.

It's like out there within 4 or 5 days.

Yeah, so—

5 million images, right?

I really didn't care about making it like look good at all. I really was like—

No, but that's beautiful, right? 'Cause you didn't care.

I didn't care. And you were just like, I don't, So my one thing was like, I'm gonna ship this once I can type in a word and see images that have that word in the prompt. That was like my one thing. If I met that thing, it was gonna be live.

Yeah, why do you think you had, 'cause I think it's kind of funny, you're someone who cares a lot.

I do care a lot.

You care a fuck ton about everything, but in this situation, it took you 4 days.

So the thing is, this wasn't my company, this was like a side project. So 4 days was like way too long. I was like, fuck it, I gotta get back to work. Like, I just gotta ship this and get it over with and like get back to the real thing.

Yeah, yeah.

That's why I didn't spend too much time on it. But I think in retrospect, that was a really good thing to do because it really made me focus on the only thing that mattered, which was like creating something new and that was valuable to the people that wanted to use the tool that I was working on.

Yeah, in some ways it kind of forced you to get to the most like pure version of the thing.

Yeah, so pro tip, if you want to start a good company, if you want a good idea, work on a shitty idea and then work on a good idea on the side that forces you to ship it really fast. And just—

no, yeah, honestly, I feel like a lot of people, even the chat, feel like their idea is like Kind of shitty, or like they feel like their execution is kind of shitty. But to be honest, dude, like yours was just yours came out like four days. Like it was so basic. It was like I mean now you look at Lexica three months later and it's like you've added so much shit because you're full time, you've raised money, you have people working on it. But dude, day one, this is you and you and you on your little laptop. Yeah, I mean fucking hunched over. You got to fix your posture, by the way.

I do.

Yeah, that's fucked up.

Let's go.

All right, let's fucking go, dude. Everybody, everybody in chat, please fix your posture. Thank you. All right. Anyways, far from Foster. But yeah, I think, I don't know, any, do you feel like, so anyways, at this point, we, you kind of shipped the thing. Stable Diffusion weights are out and you just launched it on Twitter, right?

Yeah. Launched it on Twitter. I was like, hey, I realized nobody really knows how these words in the prompts affect the images. So I built a tool that could hopefully turn this lesson to, like, this random thing where we write random words and hopefully more of a science where you can collectively as a community learn how to create better images.

Yeah.

And that did pretty well. And I think that was the initial ignition moment for Lexica, where I think we got our first 100,000 users that day. And then that week we got around 500,000 users. And really it was just timing it really well with the launch of Stable Diffusion's weights. If I had done it a week later, I don't think it would've worked. And if I had done it earlier, just no one would've cared as much because The weights weren't open sourced. So yeah, that was really lucky timing. And I think timing is a huge factor in a lot of things.

Yeah. How much of it was timing? How much of it was just like something good that you made?

I feel like it depends on the thing you're working on. If it's like, if it's something that's clearly going to take many, many years to succeed, then timing is probably less important if you're working on like clearly what's a good idea. So like if you're trying to build AGI, then Yeah, timing is good, but like, I doubt if you're a month late to it, that's a bad thing. If you're gonna get it a month later, that's okay. But if it's something where it's just like an image board of images and tags and prompts, shipping it really fast is really important. And I think shipping velocity is probably one of the most underrated things to focus on when you're a very fledgling company.

Well, so once it kind of came out on Twitter and you were like, holy shit, like 100,000 people. And you know, at this point you have like your side project, which is Lexica.

Yeah.

And then you have Debuild.

Yeah.

Which is like the thing you've been working on for, you worked on for 2 years almost. Right? Like 2 years of your life spent on like perfecting this thing called dBuild.

Yeah.

But then here this thing is, the shitty thing that you made in 4 days.

I was in denial.

Yeah, like how'd you feel? Like, like how did it feel to like be 10x bigger than you were in 4 days and what took you 2 years?

Yeah, so I, I was definitely still considering Lexica a side project, but like over time the amount of like— the hours in the day are limited where I could spend like maybe I could spend like 8 hours on dBuild, like 2 hours on Lexica. And at one point like They just like shifted where like I was spending more time on Lexica than Debuild.

Yeah.

And then I was spending like 2 hours a day on Debuild and like no hours a day on Debuild like 3 or 4 weeks into it.

Yeah.

And then I was like, I was still in that, I was like, oh, it's just a side project. Like once things are stable, I'm gonna stop working on it. And I think it was when we like broke a million users that I was like, wait, I should probably just focus on this full-time. And that was kind of like the moment I realized I was just, lying to myself about this not being a company because I was like, how could it be a company? It's like just a collection of images. Like, how do you— how do you— like, why is that a thing? Like, oh man. Uh, but yeah, that's, that's, uh, that's the story of Lexica.

No, I think, I think it's just crazy because like there's so much stuff that happens in 2 weeks. But like, I mean, everybody looks at Lexica. I mean, the thing is like you can look at Lexica and be like, wow, viral overnight success in like 4 days, right? But then knowing you for like so many years, like dude, you show up every day. You've been grinding every day for years and years, right? And I think, I don't know, there's something there. So again, someone could look at Lexicon and be like, oh yeah, one-time viral hit type thing, Sharif, whatever, whoever the fuck that guy is. But then I know it's like, oh my God, dude, this guy's been grinding for like 4 years, 5 years, like even probably longer than that, right? So I don't know, there's something to that. But then also like even just watching you, do you feel like you take this, I don't know. There was so much beauty in the beginning to it, right? Where you were like— it was like a kind of pure thing that you were doing for fun. Like, it was probably the first time you were just kind of happy, and a bit about happy about just doing something different.

Yeah.

Do you feel like that feeling is maintained now? Or do you feel like it's— because now it's a company. Like, maybe you could tell people what happened once you realized, OK, I'm going to make this a company, and I'm going to pivot my actual company to my side project.

Yeah, yeah, yeah.

Like, how did that go?

So I feel like the magic in the thing you're working on is less so whether it's a company or not, but rather whether you feel like you enjoy working on it. And I think for me, a large part of that enjoyment comes from whether I'm like doing the things I find to be interesting and I'm like fulfilling my curiosities. Whether it's a company or not doesn't really matter, but more so, is this something I actually want to learn more about and work on? Then yes, that'd be great. If it's something I don't really care about, then I probably wouldn't enjoy it, whether it was my company or even a side project.

Do you feel like you were trying too hard? So, Debuild was the previous company. Were you trying Like, looking back, were you trying too hard to make it a company?

I think so. Was I trying too hard to make it a company? I don't, I don't think I was trying too hard to make it a company. I think we spent a lot of time working on things when we should have, like, shipped shittier versions earlier. Instead, we spent way too long trying to perfect, like, the perfect thing.

Yeah.

And I think we also tried to rely too much on GPT-3's capabilities when in reality we should have used it as, like, a way to augment a really great tool in the first place. So there are a few things I would have done differently if I could go back in time. But in retrospect, I think the 2 lessons I learned the most were ship fast and ship early. And 2, it's okay if it's shitty as long as it's useful.

And probably a 3rd thing would just like work on stuff that you think is actually interesting.

Yeah, yeah. I mean, I do that regardless. But yeah, for people who don't work on things that are interesting, like Definitely work on things that you personally find interesting because it doesn't really feel like work. You're okay, like working all day long if it's something you would do anyways.

How old is Sharif? How old are you now, dude?

I turned 25 like two days ago.

You're 25?

Yeah.

Oh man! Yeah, it's crazy. Yeah, maybe we could talk about like. By the way, it's crazy not because he's 25, but because of when you were when I met you, just a little boy, dude. Just a young boy.

I think you're like 19 or something.

Like yeah, yeah.

I was like 19 and you were like. I think it was like. I think it was like 18 or 19.

Yeah, I met you in like 18, 19. I was like 21, 22. And then just a quick story for people. I met Sharif on a DM. Like the whole reason I'm here, even here just working on Build Space and in San Francisco is because of Sharif and because of a random message that I sent him. So maybe we'll talk about that in a bit.

Yeah.

Happy birthday, Sharif.

Is your birthday, bro? No, it was like why the fuck are twenty the twenty second? Today's like the twenty seventh.

Oh shit. Twenty eight.

Wow.

Happy birthday, bro.

Thank you.

It's not your birthday though.

It was my birthday five, six days ago.

Six days ago?

Yeah.

Dude, what the fuck? You even tell me?

Oh yeah. Oops.

This is weird now.

Anyway, we still gotta get IHOP. We have like this yearly tradition where we get IHOP.

Yeah, we get IHOP every every time on our birthday because our birthdays are pretty close.

Yeah.

So yeah, we we go to IHOP together and we get the shitty pancakes. There's only one IHOP. It's like thirty minutes away. Anyways, we we let's talk about like uh. Like going backwards now, before Lexica, you're working on dBuild for a bit. And then before dBuild, I don't know, anything you wanna talk about? You wanna cover dBuild? 'Cause then we can talk about your experience like going through YC too.

Yeah, yeah. So dBuild was after YC. I think dBuild has a lot of pretty fun things. So dBuild was one of the first companies that was using GPT-3 as like a core part of what the company did. And this was March, April, May, June. I think it was like around June 2020. when GPT-3 was first released. I was— so funny story about how I heard about it. So it's like someone on YC's forum Bookface was like, oh my God, OpenAI has made AGI and it's called GPT-3.

So real quick, GPT-3, this is like 2 and a half years ago?

2 and a half years ago. Yeah. This was, yeah, when their API was first released.

Okay.

So someone was like, oh my God, OpenAI has made AGI. It's mind-blowing. And then it had like zero comments and like one upvote. And I was like, what? So I, I like tried to get access to it, but I didn't have access 'cause it was like a private thing. So I just like tweeted out to like my 1,000 followers at the time. I was like, hey, does anybody have access to GPT-3? I have like cool idea I wanna work on. 'Cause I wanted to make a, a coaching chatbot that would coach me on like things I wanted. So I like, I wanted to be able to say like, hey, my goal is to run 5 times a week. And I wanted the coach to just create the plan for me and then help guide me towards it. In retrospect, it was a bad idea, but for many reasons. So I wanted to use GPT-3 to make the chatbot feel more human-like, because right now it's just all rule-based. It was like, if you didn't run the past 5 days, then send this predefined bad message. Then I tweeted that out, and then Greg Brockman, the current CTO of OpenAI, was like, hey, sure, here's a link to your— or what's your email? We'll send you a link to the beta. And I used it and I was like, wow, this is pretty good. I can make a pretty good chatbot. I like distinctly remember sitting on the couch and using my laptop at the time. And I like typed in a message where it was like, you type in the name of a person and then you could say like interviewer, and then you could like have a conversation with them. And I like typed in Steve Jobs and I was like, hi, I'm Steve Jobs. And like the interviewer was like, welcome Steve. Uh, what are you up to today? And then like you could actually have a conversation with Steve Jobs. This was like May 2020, which was like pretty mind-blowing to me. Yeah.

May 2020. This is the craziest shit of all time, dude. It was 2020. I guarantee you, even like my 12-year-old cousin on TikTok knows about this already.

Yeah, it's crazy.

That's crazy, yeah.

It's pretty insane. Like, yeah, ChatGPT-3 is pretty awesome. So I was like, oh, okay.

Why was that a bad idea, by the way, what you were working on?

It was a bad idea just because like it was too general. Like the whole point was to make a chatbot that could coach you in anything. In retrospect, I should have focused on like one really specific vertical, like fitness or like education or something.

Yeah.

So I was like, oh, if you can have a conversation, maybe you could write code. So I like tried writing, uh, like the part, the first part of a JavaScript React function, the component header. And I was like, this is a component that renders a button. And I pressed enter and it generated some code and I wasn't sure if it would work. So I just like copied and pasted into like my VS Code.

Yeah.

And I refreshed the page and it worked. I was like, what? It wrote code?

So this is back when GPT-3 writing code was the most mind-blowing thing.

This was back when nobody outside of like people who worked at OpenAI had heard of GPT-3. Yeah. Uh, so that was pretty cool. And I was like, wait, how is nobody talking about this? Like, we have an AI that can write code for you. What?

So this is actually really interesting because you were one of the first to figure this out outside of OpenAI.

Yeah. I don't know if I was the first to figure it out, but I was like probably one of the first to like put it in an IDE or an environment where as soon as you typed what you saw, we would compile the code in the browser.

You like popularized it, right? Like you like made, you made it a thing. I remember you made a tweet about it, right? Yeah. And that went super viral. And yeah, I don't know, like, that to me is a— like, you were one of the first to figure— I wouldn't say figure it out, but one of the first to figure it out and then talk about it and then be loud about it.

Yeah.

Right?

I think, yeah, I think there is a lot of value in training really great and amazing models that can do incredible things, but there is also value in trying to understand how those models work from, like, the perspective of someone who likes to use really good products. So, like, GPT-3 in the playground is, like, a cool tool, but really is a lot of what matters is the interface and how people interact with these models and how you integrate good products into them. So ChatGPT is effectively like, it's kind of almost the same model as GPT-3. It's not quite because they use reinforcement learning to make it a lot better and more conversational and natural. But like, it's kind of almost the same model with some additional fine-tuning on top. But ChatGPT has blown up in a way that GPT-3, the API, hasn't, which is pretty interesting.

Yeah.

And it just goes to show that The way you interact with these models matters a lot, and it's a really hard problem in how to get it right.

Yeah. So that was like 2 years ago. Copilot wasn't out, right? Copilot wasn't a thing. People writing code with GPT-3 wasn't a thing. But you were like, holy fuck, man, you can write code with GPT-3.

Yeah.

I'm gonna make an IDE that helps people write code and make— help make, uh, make it easier to help others like write code, right? Because that's probably gonna be net positive for the world. More people can write code, more people can build the things they wanna build. Great. You don't have to learn what a null pointer is. Amazing. Let's do this. And I distinctly remember you put that tweet out and you actually raised money from one tweet in a week. And by the way, both me and Sharif, I don't think either of us are good at raising money. We're actually shit at raising money for the most part, but we're okay. In the classic sense of, I don't think me and you are DMing people, constantly cold DMing, pitching all the time. Maybe you can talk about that, how you went from kind of this random idea for Debuild like 2 and a half years ago to raising from a tweet in a week like $2 million.

My fundraising strategy is like— so I think one of the most memorable things from one of these YC presentations during the batch was the best way to fundraise is not to be good at fundraising or not to get good at fundraising, but instead it's just to make a good company. And then focus on making a good company. You will get investors that want to invest in your company. Like fundraising in and of itself is like not really a skill worth learning. But really it's more like build something worth investing in the first place and like the money will find you, as cliché as that sounds.

No. Yeah, it's— you're like living proof. Yeah. Yeah. And I don't know, I just feel like so much of— if I take out the moments in your life, they just sound ridiculous, right? It's like, I was fucking around with Steve Jobs chatbot, and then I figured I can write code. And then a week later you have millions of dollars. Yeah. And you're working on this thing. But then I guess, uh, yeah, just, just for the audience timeline-wise, we're going backwards, right?

We are.

So you did Lexica, then we did dBuild. Like, you spent 2 years on dBuild. Maybe you could talk about that for everybody, like what it was like getting that magic aha moment of this is just the most obvious. Because if you think about it, Debuild didn't work out. Like, you actually don't work on it anymore, right? But 2 years ago, you were like one of the first people to be like, holy fuck, like, I'm gonna make a product around this thing and let it help me write code.

Yeah.

And even looking back now, it's like, that's a fucking genius idea. Like, that should have worked. That— why didn't that work? Why isn't Sharif working on that now? Because I feel like a lot of people are looking for like— I don't know, dude, even me sometimes I look for the great idea. And if the idea doesn't sound good, I'm kind of like frustrated. But then in this case, you had a good idea, you had people online excited by the idea, and you had millions of dollars to build the idea. So then where did it all kind of— where did it all kind of go from there? Because I don't think anything went wrong per se, but it's more like you went down a different path now. So what actually ended up happening?

So we, we worked on dBuild for around 2 years, and I think a few things we did wrong was we were, one, focused on being a very general-purpose tool. We didn't focus on any one specific vertical. So to summarize what we were trying to build in one line is, what if it was like, what if you had like a programming environment where you could have a model do a lot of the annoying stuff you do when you're building web apps, such as writing the JavaScript functions and like writing SQL queries, and you just did the fun stuff like assembling your UI and like picking the colors for your button?

Yeah.

And you could have this tool where Humans did a lot of the stuff, but models did a lot of the stuff that you didn't want to do as a human. And it was meant to be this like human collaboration tool where you could both do things at the same time. I think, so I still think it's a pretty good idea, especially now that like GPT-3.5 has come out. And it's only going to get better. But it was just that Lexica was like, so DeepL was like growing like this. when Lexica was just like growing like that. Yeah, I was like, at one point you just had to like look at the numbers and just like, like respect reality.

Yeah.

So it was like, it was a hard decision, but at the same time, like, it's hard to argue between having like 1,000 users versus a million users.

I guess when you look back at those 2 years, I mean, dude, I've been the same thing, right? Like, I spent 2 years working on ZipSchool before BuildSpace did anything.

Yeah.

Like, and I, I don't see this, I don't see it as a waste, right? But how do you look back at those 2 years? Because I think a lot of those people In the chat are starting those two years today without even knowing it. Like they're expecting their thing to work out next week and the week after and the week after. But here we are, like years spent in the shithole, right? So like when you look back on it, were you like in the shithole or were you enjoying your time? Like how are you looking back at it now? Those two years, like how were they spent?

Yeah. So we spent almost.

I guess how do you like feel about like how they were spent?

Yeah.

Do you feel like it was a waste?

I don't feel like it was a waste mostly because I enjoyed what I was working on. If I didn't enjoy what I was working on, I was like I just need to. struggle and get through this and then we'll be good, I would've considered it to be a waste. But luckily for me, like, I don't really work on anything unless I find it like really fun and enjoying, which I think is a good life hack for anything. Just only work on things you find fun and enjoying, 'cause you'll never like waste your time on anything. 'Cause it's like stuff I would've done like anyways, if I, it was just like, if I was working at another job, I would've done this for fun anyways.

Yeah.

But in retrospect, I think having more contact with reality in terms of like your users is really important. Not isolating yourself in like this golden tower where you're like, oh, I know what they want. I'm just gonna go and build the car for them, even though they're asking for like electric horse-drawn carriages or something. Like, no, you actually need to go and talk to them. You're not like—

Yeah.

Very few of us are actually Steve Jobs and like you're just gonna craft the iPhone out of nothing.

Yeah, and also that guy's story is just so twisted at this point. Like, that guy was fucking testing the shit out of that thing like all the time. If anyone's read the book on like the iPhone, right? Yeah, yeah, yeah. They got— the guy didn't just make the iPhone with 200 people. It's like hundreds and thousands of people. But anyways, um, just by the way, we're gonna answer questions at the end for 10 minutes. Uh, but I think there's one interesting question I saw just because I thought— I know, I know what your answer is, but, uh, it's, uh, uh, did Lexica worry about things like a competitive moat?

Did Lexica worry about things like a competitive moat? So at the time, no, I wasn't even thinking about that. I was like just thinking about like Shit, our like cloud image bill is like $20,000 a month. How do I get that down to like something that's not gonna like make us broke? And then I was thinking about like, oh, that's better. Like, oh wait, now we're getting like DDoSed from like these random IPs in China and Korea. How do I stop that? Oh, you can just like block the entire country. Cool, that worked. It's like, it's hard to focus on long-term strategy when you're just kind of in the trenches.

Yeah.

I think competitive moats are important and like you do have to think about them, in the earliest of stages, I feel like even finding something that works in the first place is a much harder problem and much— it's worth spending your time on just making something people want in the first place.

Yeah.

And then once you do that, then you can figure out the competitive mode aspect.

Yeah, it's interesting now. Like, now we got Lexica and we got Lexica today, debuild, 2 years on that thing, right? I guess my last thing on this, on this thing is like, how did I think the most important part about these two years, like when I was working on Zip School, right? Like those two years, even before that Zip School, like even before I was working like other things, right? Like so many years. Like I don't also I also don't look look at it as a waste. I think it's I could have spent some time better for sure.

Yeah.

But I think that again this is also why I don't consider Lexica an overnight success because like the skills that you learned for like the last five ten years.

Oh yeah, definitely.

Like for example, even you just writing the tweet. Like how you write the tweet, how the OG image looks, how like the first, the copy headline looks. It's like, to other people, this could be like, oh, he got lucky and he just made it great the first time. But dude, you've made so many of these in a shitty way already, right?

Yeah, yeah.

Like you've done, you've put in the reps almost. Like that's how I see BuildSpace at least. Like BuildSpace may come off as something that looks pretty good today, but it's because the team and me, we've just fucked up for decades. Like collectively, like we're like 9 people at BuildSpace, right? So we've probably fucked up for like a collective 50 to 70 years already.

Wow.

Think about that. So you have 50 to 70 years of fucking up at this company only to make the next— this year, um, actually like, like not a fuck-up, right?

Yeah, I, I think, yeah, if I hadn't worked on Debuild, there's no way I would have been able to work on Lexica because like all of the skills I, I learned while working on Debuild made it possible to build Lexica in under a week. If I hadn't worked on Debuild, I would not have known how to do that. And I wouldn't even known— I probably wouldn't even known about Stable Diffusion, like focused on, uh, generative models in the first place. Like, I probably would have just like, probably would have heard about it like 2 or 3 months after the fact. So I think it's important to always be swimming is a good way to put it. Like, always be moving forward. Never just like stand still. Always try to be creating and shipping things. Like, I know it's not really good advice, but like, it's really just like make things and like put them in front of people and then see if they use it. If they use it, cool, build more of that thing. If they don't, then stop building that thing and build something else. It's like as simple as that.

Yeah, but I guess like the hardest part about this now is like how do you how do you stay how do you what do you call it stay stay motivated? Like how do you continuously wake up in the morning? And you know, I think I kind of understand like you are enjoying what you what you're doing.

Yeah.

But let's be honest, man. It's kind of hard to even like you can love the shit out of AI images, but sometimes in the morning you're like, man, fuck AI images. Like fuck all this. And I don't know, that's how I feel sometimes, right?

Yeah.

Like I just don't want to work on any of this. I'm like, I want it all get away from me for a minute. Like how do you kind of continuously keep it, keep it going, right? Because now you're like, I don't know, you're on like this, you're on these, you're on a really long-term journey at this point. Like you're already, you've been here for, for a while at this point compared to most people. Most people like drop out of this journey after like a year or after a couple months. But here me and you are for some reason. But I think we all have different philosophies. Like I have different philosophy from you, but I'm like, what's your Yeah, what keeps you going?

Yeah, so there are a few things. I mean, curiosity is one of them. If it's something you find intellectually interesting, you'll keep doing it even if you don't want to. I guess you will want to do it in the first place. That's why you'll keep working on it. The second is having a duty to your users. When you have people using your product and they're asking for things and things are broken, you feel like this moral sense of responsibility to making things better for them on their behalf. The third thing is like kind of pushing the boundaries of what's possible.

Yeah.

Because you quickly realize that the market is not efficient and that there are things that are possible today that people have not built yet that are pretty useful for the world. And you as a person who is capable of building and creating things has like this moral responsibility, this like social duty to the world to go and show it and to go and create these things and then put them in front of people and like hopefully make their lives better for it. It's like the collection of these 3 things that I think helps me stay motivated in the long run.

Yeah, I think that's good. It's like none of those things are necessarily like— I think it's really important everybody catches that. Like, nowhere did you say you want to build a billion-dollar company.

Oh, yeah, that's fair. I mean, like, money is important, but—

No, but like, that's like— I feel like 20— I feel like 18-year-old me and you would have said something different.

Probably, yeah.

I think we would have said billion-dollar company.

Probably, yeah.

It's like, what's your motivation to wake up in the morning? Well, dude, I just want to build a billion-dollar company. Yeah, like, of course, that's what I want to do. But I feel like that's still what we want to do at a level. But yeah.

Yeah, I feel like a billion-dollar company is like a proxy for whether or not you've like created something meaningful for the world. Like, oh, it's worth that much money, then surely what you've made is useful, right? But I think what really matters is like you creating useful things for the world, and especially things that have not existed yet. And I think a lot of people think just because like you've thought of something that Like, people assume the market is way more efficient than it really is. And there's this great quote by Nat Friedman where he's like, it's probably a better idea to model the world as 500 people than like 8 billion people.

Yeah.

So, if you thought of something and you think like it doesn't exist because it's not a good idea, it's probably just because like no one's actually gone and done it.

Yeah.

And I think it's like, yeah, like better to model the world as 500 people than 8 billion. Cool. Have all 500 people missed this? That's probably pretty possible.

Yeah. I feel like both me and you changed so much in that aspect where like we were really chasing like the billion-dollar company. Like that was our core motivation. Like we still want to do that, but now like you have other motivations, these 3 things you laid out, right?

Yeah.

I have other motivations. Like, and I think that's what makes it beautiful. 'Cause it turns out it's really hard to wake up in the morning for like years and years and years. And be like, I'm gonna build a billion-dollar company today. It's like, tech turns out it's not that exciting after like 365, 400 times of doing that, right?

Yeah.

What's more exciting is like, I kind of just wanna be happy. I kind of just wanna be curious and like fulfilled, right? So it's like—

Yeah, no one's gonna wanna like look back on their lives when they're 80 years old and like think about all the years they spent working to build a billion-dollar company. But if you look back on your years and you think back to like how you helped push like the technological frontier forward, you're gonna consider that probably a pretty good—

Yeah.

Probably a pretty good way to spend your time or like, I created things that didn't exist, and I made people's lives because they're using the things that I've made. That's also a pretty good thing to look back on. Like, I don't think anyone's ever gonna be like, I've made so many billions of dollars, this is great, my life is such a fulfilling— I spent such a fulfilling life making all that money.

Yeah.

No, like, no one—

like, it's also not about it getting to— honestly, it's probably not about it even getting, um— I don't know, how important is it for things to get big for you?

I think it's an it's important in the sense that it gets big if it's useful for a lot of people.

Yeah.

So, if it gets big, that means you're probably working on something important. So, I think creating things is important, but creating in isolation is not that good of an idea for me. Like, I think there is merit into, like, working on things just for, like, pure curiosity, but also I think having that, like, motivation of wanting to build something that helps a lot of people or, like, makes that, that is useful for a lot of people is also really important. And I think by proxy, like, if what you're working on is used by a lot of people, it's probably also very useful. Like, it's like the area under the curve, like how useful is it multiplied by the number of people that use it?

Yeah.

You can make something incredibly useful for like 100 people or like slightly useful for like a billion people. Like the area under the curve is like approximately the same.

Yeah. Oh my God. Like, AJ is just in the back rolling in.

Nice.

Everybody just saw that too. That's funny. That's AJ, by the way, founder of Neurocity. They make the headset. Yeah, we all work together here, by the way, for those not initiated. We all work in the same place. Oh, that was funny. Nice. Okay, cool. So that's Lexica. That's Debuild.

And that was— You want to talk about origin story of Sharif and Farzad?

Oh shit, the origin story of me and you.

Yeah.

I think what's kind of interesting. Yeah, maybe. Yeah, maybe. Maybe you can start it off because you know the whole reason we're here that this guy's rolling in behind us now talking on the phone is just it's just kind of crazy this all even happened right now. So maybe you can talk about it because it's it's a pretty wild. So I guess to preface this because I don't I don't think people give a shit about me and you, our friendship or our relationship. But I think a lot of people here like are trying to make some buddies in the on the internet. Or they're trying to find their tribe, right? They're trying to find these people nights and weekends like who are who are also building cool stuff. And I think everybody has this idea of like I want that community around me that that's building cool shit to help motivate me because honestly, dude, it's pretty hard. One of the main things that's probably helped me and you is just like like a small group of five to ten people.

Oh yeah.

Like so, I think that I think that's a big one. And also look at where we grew up. You grew up in Maryland, like in a random suburb. And then I grew up in South Florida, a random suburb. Like nobody builds companies where we are, right? But somehow we've ended up here. So yeah, maybe you could talk about it.

Yeah. So I think it was, I don't remember exactly, but you like DM'd me on Twitter and I was like, hey, this thing you're working on, VectorDash at the time is like really cool. And I was like, oh, hey, thanks. Like, I forget you were working on something. I think you were working like at League of Legends, like deep learning, like a bot.

Yeah. At the time I was working on a, I was actually working on a very— it was called BuildSpace. Oh really? It was called, uh— oh no, this was way before that, dude. This was, uh, this was like— I was working on the augmented reality menu app.

No, you were still in Florida at the time.

Yeah, this was— I remember when I met you, it was the AR menu app. Really? So I forgot— for basically, you can— it was a, it was a fucking great idea. What a fucking great idea. Should have been worth a billion dollars. I was scammed by society. The fact that it's— but basically you would take your phone. This is when ARKit had just come out. You'd point it at a table And it would show you that food in augmented reality. Um, that was the idea. So I would go to restaurants, I would 3D model their food, and I would, uh, um, I would, uh, I would put their menu in, in AR. And remember, this is like 3 months after AR came out. Very similar to your story actually with GPT-3, where I'm sitting there like, why is nobody talking about ARKit? This shit is crazy. I had like 10 ideas and this was one of the ideas. And I spent a year working on that. And you didn't know that? I spent a whole year working on it.

Really?

Yeah.

Oh my God.

Yeah, man. Again, like, year not wasted though. That $10 MRR after one year hit real hard. It was like taking just the best line of crack ever. No, I'm just kidding. But, but yeah, you met me in that scenario where I was just working on that app and you were working on VectorDash.

Yeah. So that was pretty cool. Yeah. And then we ended up being roommates for a while.

Well, what was VectorDash, by the way?

VectorDash was— so the TL;DR was I was like training neural networks and I was like, AWS is really expensive when it comes to renting GPU servers. And I had a friend who was mining Ethereum on an NVIDIA 1080 Ti and he was making like $4 a day. I was paying $60 a day on Amazon. And I was like, whoa, there's this huge arbitrage opportunity. What if we gave the people with the GPUs access to like— or gave the machine learning researchers access to the crypto GPUs?

Yeah.

They'd make way more money. The researchers would pay way less. So we built like a tool that would do exactly that. Cryptocurrency miners could list their GPUs on VectorDash. Researchers could rent them out. And we worked on that for like almost 2 years. And it was a good idea until we realized that students have no money and they're awful customers. Machine learning researchers, like graduate students at universities, are also awful customers because they also don't have money.

Yeah.

Startups get a lot of credits from Amazon and Google Cloud and Azure, like hundreds of thousands of dollars of credits.

Yeah.

And large enterprises are not comfortable having their data on GPUs or servers that no one or anyone can really get access to. So we were stuck selling to undergraduates and graduate students, and we were the cheapest cloud GPU provider for like a solid year. And then, um, we switched to cloud gaming because we realized we have like 100,000 GPUs. We have the lowest latency to any—

Yeah.

Like the lowest latency GPU cloud provider out there. So what matters for, uh, What products can you build with a low-latency network? It's cloud gaming, where you can stream games from your computer.

So basically, you started with the cloud network of GPU idea.

Yeah.

Which actually sounds like a great idea on paper, right?

It's a bad idea on paper.

No, but Sharif, you don't get it, bro. Imagine that you could rent out your GPU when it's not being used. It's just sitting there. You're not playing League of Legends. You're at work. Imagine if you could rent that GPU out—

Uber for GPUs.

What's that?

It was Uber for GPUs, literally.

It was literally Uber for GPUs.

Actually, no, I pitched it as Airbnb for GPUs. No, but it ended up not working out.

And even now, looking back, me and you can look back at that idea now and be like, wow, kind of obvious that that wouldn't work out. Or not maybe obvious, but I think me and you today could look at that idea and be like, that's not going to work.

Yeah, right. I don't know. It sounds like a good idea today, but like, now that I've worked on it, I don't think it's a good idea. But the thing is, people are still working on variants of this idea. Yeah. So it might work today. It might not. I don't know. It might work today.

It just all depends on execution. But I think it's kind of funny. We're kind of much better at picking ideas now too, like after working on so many of these ideas. But anyways, you posted that on Hacker News, like VectorDash.

Actually, I didn't post it. Someone else did.

Someone else?

Yeah.

I was like, that's crazy. So someone else ended up posting that on Hacker News. And then at the time I was in Orlando, Florida. I was like, get me the fuck out of here. And I was on Hacker News. I was like trying to make friends. I didn't know what I was doing. And I see, I see you or your company VectorDash at the top. And I look at the company, like I looked at your name and I saw that you were— there were 3 other people working on it. It was you, Arbaz, and Abhishek. Yeah. And I was like, all these fuckers are in college just like me. Like, why do they look more legit than me? Like, your landing page is better, your copy. I was so mad. But at the same time, I was just like, I— this guy's so cool because I looked here, I went to your Twitter, I scrolled it and I was like, this guy's so cool. I want to be friends with this guy. And then I add you on Discord and all of a sudden, like, we're just— we're buddies, you know? Yeah. We're like giving— we're like DMing each other. We're FaceTiming each other. Talking about stuff, and then that's how we met. Just a DM where I was like, "Hey Sharif, I think Vector Dash is really cool." Yeah, we both convinced each other to move to San Francisco.

Yeah, that was pretty crazy. That was pretty crazy.

But I think all to say, it's like I it wasn't my intention to try and like be your buddy. I was just like, "Bro, this this is sick." Let's be honest. Yeah. But and then yeah, from there we we end up moving out to San Francisco. It's always better with a buddy, right? It's hard to do things alone generally.

I remember my first impression of you was like, oh shit, you're taller than I expected. Wow.

What's the first thing we did? I like bought you something, right? Yeah, yeah, we—

I remember this. You, uh, we like got ice cream. I think you bought us ice cream. Yeah, yeah, yeah. Oh yeah, you bought me and my co-founder, uh, Arbaz ice cream.

Yeah, I remember the first time I met you guys. You guys were so awkward. Yeah, you were like so fucked. Imagine meeting your online friend for the first time. It's so awkward, dude. These guys walk up, they're barely talking. I'm like, yo, can I get you guys an ice cream? You're just getting ice cream. Nice. Yeah, I remember that. That was cool. But anyways, all to say, it's like, I don't know, any thoughts for people like watching this now or like live right now and just thinking, damn, like I don't really have a community around me. Damn, I haven't really met that many people. Yeah. I haven't, I don't really, I'm kind of just alone in this game right now, it feels like.

So I'll give like 2 varying levels of effort answers. The highest effort answer is like move to a city that There where you can surround yourself with a lot of people. So if you're in the United States, that's probably like moving to San Francisco.

Yeah.

If you're not in the United States, then probably moving to like another city within your country or somewhere nearby where there's like a tech hub or a community of people who build a lot of things. That's the high-effort answer.

That's like the highest effort.

That is— it's like uproot your entire life and move to the city so you can like build things. That's like the highest effort answer you could possibly give. The more realistic one is like just use the internet. and reach out to people who you think are doing interesting things. Ask them questions that you genuinely want answered and don't have this ulterior motive of networking. Just like, oh, this person is continuously talking about cool things. They have a great blog. I should just message them and ask them about this thing they mentioned in their blog, or I'll ask them about their company, or really just anything they've mentioned that's interesting.

Yeah.

And people are generally pretty receptive to cold DMs or cold emails if they're thoughtful and like not a pre-crafted sales pitch, because people are pretty good at filtering those out and not replying to them. But yeah, just use the internet. The internet's like the greatest resource you have.

I really do think a lot of people think that their only option is to come to a great city.

That's not their only option.

But like, think about it. Like, we started— I spent forever in a suburb, and you spent forever in a suburb. Yeah, like you didn't You only moved out recently. I think we moved out like like to get like like it was a couple years ago even, right? So it's like or whatever it may be, but it's like we spent most of our time at home at our parents' house basically making stuff. But then we just got really good at talking, finding the right people online because I think it's pretty exhausting too, right? Can because I think what most people end up doing is they'll like they'll DM someone, they won't respond.

Yeah, that happens.

Or like they'll DM someone, it's like they get on a call on it. They're like just like I'm not I'm not into this. Yeah, that's right.

I mean like. That's gonna happen if you like DM like Elon Musk. Of course he's not gonna respond to you. But like for most other people, if your DM is thoughtful, if it's relevant to the person you're DMing, and if you're not like trying to sell them on your company, then that's also just like a pretty good DM.

It's like fine to reach out to people. I don't know, because dude, back then I had no— I had no idea who was good and who was bad. Like I didn't know. I think both me and you were nobodies. We were absolute nobodies. I was a college student, you were a college student, and I was like All right, like, this guy's a nobody, I'm a nobody. And I think that's what made it beautiful, right? I wasn't trying to reach out to a YC founder for advice. I was like, yo, I'm a nobody, you're nobody.

You got to find people that are like in the same grind set as you in terms of like where they are in their life or their company. Like, if you're a college student, try to make friends with other college students working on companies. If you're like, uh, if you've been working full-time for a few years and like trying to make a leap into working on your own thing, find people who are in similar positions to you. It makes it a lot better. It's just like find people who are in similar stages and similar situations in their life and then befriend them. It just generally having people who are like going through the same struggle as you makes it way easier.

Yeah, I think I probably messaged like 10 people or 20 people like, I love your thing before even anybody responded to me.

Nice.

Like, and I think, I don't know, it's fine for people to not respond, right?

It's fine because you have to realize like most people are one, pretty busy. Two, they get like a lot of spam. So like It's fair for them to assume that your random cold DM or cold email is like another spam message where they're trying to sell you on like some enterprise software that they don't need.

Honestly, it's also easiest just to find people at the same level.

Yeah.

Like, not the same level, like in the same level in the sense of like, when I— when we met, we were both nobody college students. And like, that helped a lot. Like, I had no pressure. I, I had like terrible imposter syndrome. I was like, I don't want to reach out to anybody. And then like, uh, and then I don't know where you were at, but then I don't know, it felt like nothing. Like, whereas if I'm like the founder of YC and I reach out to you, it's like a whole different, whole different thing.

Yeah.

But yeah, I think if you're in nights and weekends now, like it's almost— what would be like the one thing you tell people if they're nights and weekends now and like they can see each other on Twitter, for example?

I was just like, reach out to people who are also doing it. Reach out to people who are working on ideas you find are interesting. Reach out to people who you would like to be friends with. And most of the time they're probably going to be pretty receptive to whatever you have to say.

Yeah, cool. All right, so that's kind of like the, the last of this, like, last arc of your life, which was like Lexica, then D-Build, then VectorDash. We didn't even talk about YC, um, and going through, uh, VectorDash, going through YC, you dropping out of college, all that shit.

We pivoted our idea the day of our interview and we got into our YC interview.

Yeah, wait, is that true?

Yeah, yeah, we, we realized we were—

Wait, we interviewed together, right?

No, no, I don't think so.

I remember, no, I remember I interviewed with the— my, my AR app idea, and then you interviewed with the, the ML idea. Yeah, we—

so we That was what we applied with. And then the day of our interview, we walk into the interview room and we're actually like, hey, we pivoted our idea because we realized it's not working. Now we're a gaming company. Yeah. Like, what the fuck? And we explained it to them and they're like, this kind of makes sense. Yeah. And then we were like, oh, we didn't get it and we're fucked. All right, it's okay though.

But then I remember you came home, you 3 were depressed. Like, I also, I also applied, but then I got rejected. I got— I immediately got rejected. Uh, but you got the interview in person.

Oh no, no, I got rejected.

They were like, fuck, this is stupid.

I was gonna say, like, if they reject you in person, I've never—

no, no, they rejected me on But we were living together and these 3 come home. They were just down. You guys were so down. And I was like, guys, like, it's gonna be okay. They're like, bro, we didn't get in. The world is over. Everything sucks. Like, we were trash. But then I remember, uh, what happens from there?

So we, we get the call and they're like, yeah, you're in. And we were like, what? What? And then we were like, okay, cool. This is great. And then, uh, we're like riding that euphoria for like a day. Then we were like, wait, we still have to like keep working and like nothing really changes.

Yeah, it's like, yeah, you got the $150K or like $100K at the time. I think that was a big one, right?

That was a big one, yeah, because I think we were running out of money at the time. So we actually like needed the money really badly. So that was a huge relief. Yeah.

That's crazy, dude. And then you pivoted the idea in YC or right in the interview?

We pivoted it about like 20 minutes. So we were in like the waiting room before our interview where like all these like nervous founders are sitting and just like nervously eating a granola bar graciously provided by YC. And we pivoted it 20 minutes before our interview was— like, literally 20 minutes before we walked into the room. And we just explained why we pivoted it. We were like, hey, we don't think this is a good idea, but here we have this other idea that uses a lot of the things that we built for this previous idea. We think it's going to be amazing.

Yeah.

And I remember the point I realized we might have had a chance was when Jeff Rawlinson, one of the YC partners, was like— he worked on Yahoo Mail, and he was like, oh, Wait, so you're saying that because you have a peer-to-peer network of GPUs, you're going to have lower latency than anyone? And because you're peer-to-peer, the streaming of these games will be like, you can play like really high frame, or you can play like high frame rate first-person shooters and stream that? I was like, oh yeah, like that's actually right. When they start pitching it back to you, you kind of realize you have a chance. So that was pretty cool. But we still think we didn't get in. And then like that afternoon around like 8 or 9 PM. Yeah, we got the the fame to call. Yeah, but at this point number.

But at this point, like you were already in. We had already moved out. Yeah, we were already out. We were already all in on it. We were. Yeah. And then we had like not that much money. It just did. I can't believe this even happened, you know? But like we both took like a massive risk, and I mean yeah, here we are. It's kind of funny.

But things generally work out for most people. I've realized. Yeah, I think what people.

I think one thing that's kind of interesting is like things usually— the worst-case scenario is not that bad. Like the worst-case scenario of us moving out to SF and like it not working out was that worst case, we're not getting a job at Starbucks. Yeah. And then we're like now like grinding for like survival.

I remember this one thing that, uh, this one guy told me. So this was when I was still in Maryland. I was considering dropping out of college. It was my junior year and I was like failing most of my classes. because I was just like working on VectorDash and not going to any of them.

Yeah.

And I wanted to like drop out, but I was like, this is a shitty idea. I'm not gonna— I'm not Mark Zuckerberg. Who am I kidding? Uh, and then I spoke to this guy named Alex Metall, who is a YC founder who turned investor. And he was like, hey, actually, um, I dropped out of college. I dropped out of college to do YC.

Yeah.

And I started this company and, uh, we did YC, but the company didn't end up working out. So I just returned to college and I ended up graduating along with my classmates. I was like, wait, what? You can just go back if it doesn't work out?

Yeah, it's like, it's like that doesn't even make sense. Like, it's like you have— it's like all—

it's like, yeah, it's like not this permanent irreversible decision. It's like something you can just try out.

Also, I feel like— sorry, I keep going.

I don't cut you off. Yeah, so that, that was one thing. And then this other thing was like, I read this book called The Case Against Education. It's a really good book by this, uh, George Mason professor of economics named Bryan Caplan. who explains the benefits and drawbacks of a university degree. And that was a really good book in helping me understand why degrees were useful and why they weren't.

Yeah.

So, if you're considering dropping out, I recommend reading that book. It's also just like a good book, even if you don't consider dropping out.

Yeah. Also, dude, we are kind of skipping over one of the biggest things and one of the most hard— the hardest things to get over, which is we both have brown parents.

Yeah, we do.

We both have brown Muslim parents.

Yeah.

Who the idea of dropping out, the idea of not getting a job— like, when I told my dad, I'm like, I'm like doing whatever the fuck I want, dude, he was so mad. So mad. I don't know, how did that— how did that go for you? Because I think a lot of— like, press 1 if you're like in a similar scenario where you're like, you kind of want to go down this path, but your, your parents aren't like all in. Yeah, to be honest, it's a never-ending problem. Like, even when you're 35, your, your parents could not be all in.

Yeah, my parents still ask if I'm gonna go back to school.

Do they do?

Yeah.

What the fuck? That's crazy.

Not as much anymore, but like—

My dad still asks me if I'm gonna do the PhD.

Really?

He's like, for them, Stanford is right there. Like, you could do the PhD.

And I'm like, Father, just take, take night classes and just like show him your, uh, show him your online diploma.

Yeah, yeah. But I guess, how did you kind of navigate that, if any, if any way?

So I think this was also partly what allowed me to like mentally prepare myself for dropping out was I didn't think it was permanent. I thought it was just like a temporary, like one semester thing. And if it worked well for that semester, I'd do another semester. And I just explained to my parents, yeah, this is not permanent. If it works out, I'll continue to work on it. If it doesn't, I'll go back to school. And they're like, that sucks, but it's also kind of reasonable. And sure, why not? Like, it definitely still took a lot of convincing, but the fact that dropping out isn't a permanent irreversible decision is like a huge factor. And I think you can like explain that to people and they're pretty reasonable about it. Yeah.

Yeah, I think that's like the most important part about this that we've probably gotten a little bit better at, which is that to me, the way I vocalize it in my head is like, the worst-case scenario is not that bad.

Yeah, exactly.

Like, the worst-case scenario of you going full-time, like any of you all going full-time on your idea, no matter how shitty you think it is for 6 months, is that at the end of it, you're at $0. At the end of it, yeah. The worst-case scenario for me would probably be $0 and no job. But then if I think about it, it's like, okay, what do I do? Well, okay, I'd probably get some random job. I'd probably start walking dogs. I'd figure it out. You know, like the worst-case scenario isn't that bad.

The worst-case scenario is you have this thing which you think is a really great idea and you don't work on it because you're afraid you'll— you're not gonna be able to make ends meet for those at the end of those 6 months. Yeah. And you're living with the regret of not having tried this thing that you really wanted to try. I think arguably that's worse than trying the thing, failing, and then having to start over and get a regular job again because you're not living with the doubt of what could have been.

Also, dude, how many job— do you know how many job offers I got after working on that AR menu app that failed? I— everybody wanted to hire me, from like Google to Amazon. They were all just pitching me like, we love your AR app idea, we love that you're even working on your own idea. And it's— I'm like, you guys know there's zero people using this, right? But it's like, turns out working on an idea and taking a risk and failing Yeah, it's actually like, it's a huge life hack, the best thing on your resume.

People, people love it. Like, recruiters love seeing like you were a founder of a company, even if it had like zero customers.

It's because it's so rare for anybody to do anything like different. So it's like when you— when someone does, even if they fail, people, people often think that like, oh, if it fails, then I'm a nobody, I'm a loser. But it's like, actually, you already did something that most of humanity didn't even attempt. Yeah. And because of that, you're kind of treated a little bit differently in the job market, at least.

Yeah, it is a good life hack. So I mean, like, even if you don't want to start a thing, it's probably still worthwhile to like try to do something like unethical life hacks. Start a company and just put it on your resume or LinkedIn so you can get a job offer from other companies.

Yeah, I don't know. I think once I started seeing that, I was like, I don't really care anymore. Like once I was in college and I was— I realized that like, wait a second, like These people— I'm 19, and they're already offering me these things. Like, what the fuck? But it's like, I'm not anymore— I'm not smart.

Yeah, so this ties back to like the whole reason of getting a college degree is, for most people, it's so they can secure jobs in the first place.

Yep, yep, yep.

A lot of times, when you're a recruiter and you're comparing someone's LinkedIn or resume, someone who's been a founder is often more appealing than someone who has a college degree, because everyone has a college degree— or not everyone, but a lot of people have college degrees. not that many people are founders. And when you're trying to figure out which one to extend an offer to, odds are you're going to pick the person who's the founder, who stands out.

Yeah, I guess all to say, uh, don't start a company to get a job, that's for sure. Or don't, don't become a founder to get a job. But just become a founder just to get a job. I guess the lesson here is like, if you truly think about everybody, if everybody just thinks about their worst case, the worst case is like you have $0 and you're poor. Like, that was my worst case. So, but then as long as you trust yourself that, hey, even if I'm at $0 and like I have a month left of rent. Like, it's gonna be fine because I'm gonna figure it out. I think is—

I think at this point, having that mindset's like super important.

Yeah, just trusting yourself, trusting yourself, trusting the process. Yeah. Okay, cool. Um, if ever— we're gonna take questions for like 10 minutes. I think we're already like 10 minutes over time, so cool. Well, let's do like— let's do like 10 minutes of questions. So if anybody has questions in the chat, I'll be reading them here. I know a lot of people have been asking questions during, so— and you could probably go as fast as you want, by the way. Okay. I'm 18 right now. What would you say to 18-year-old self right now if you got the chance?

To my 18-year-old self. So, 18-year-old self in like 2015 would be different than 18-year-old self like in 2023. I would tell my 18-year-old self in 2023, build things that you think are useful for other people. I would probably tell him to focus on like AI as a field just because it's like one of the few fields where there's like this massive like overhang of capabilities, but not many like useful products being built. I would probably like try to focus most of my time on that. Like just find a way to like get in. Like you don't have to be a researcher. You don't have to like be the one doing the training of the models, but like just doing something there is probably gonna be pretty valuable and important.

Yeah.

Yeah.

You don't have to answer this, but have you ever taken shrooms?

I haven't.

I haven't either, bro. Yeah, maybe we should do it together.

Maybe. Yeah.

Maybe, maybe.

Maybe instead of IHOP.

No, no, we have to do after IHOP. All right, maybe, um, at IHOP. Uh, how does Lexica work? I think it's kind of interesting. How does Lexica work under the hood? Because it's actually crazy, the millions of images you're serving, how much money it costs. Like, that shit's crazy.

Yeah, so it's, uh, so there are 2 core parts of it. One is just like image serving that's really fast, and we're using Cloudflare Workers for that. Uh, we resize images into 3 different variants, so depending on how wide the image is on your screen, we serve a larger image, or if it's smaller, we serve a smaller one. Having like dynamic image sizing is really important because then you can serve like hundreds of images at a page on a page without it taking forever to load. The search is we're using this model called CLIP, which is a model that encodes images and also text into the same vector space. And then when you do that, you can effectively use KNN search to search for images and text. So if I give a query for cat, we convert that to the vector and then we just compare that vector to all of the other image vectors and see which ones are closest in vector space. And then we just return like the top 50 closest image when you search.

Yeah. What's your— how did you— because this could easily cost like tens of thousands of dollars a month.

Yeah.

How did you avoid that?

So we spent a lot of time. The first month it did cost like more than $20,000. And then we realized I was making a big mistake in how images were being served and just like spending some time optimizing it.

Yeah.

Just reducing image sizes to be a lot smaller.

That's a huge thing. Is there any one book that you'd be like, damn, that book is— it helped a lot?

Yeah, so there's this— I'm actually rereading it. I've read it like 3 times already, but I'm rereading it. I try to reread it like every year. It's called A Guide to the Good Life. And it's like this— it's pretty much a summary of like Stoicism as a philosophy. And I think that's a generally good philosophy to have as a founder where Like, the general gist of it is there are 2 things in life. There are things you can control and there are things you can't control that are out of your reach. For the things you can control, you generally shouldn't worry about them because you're in control of them. For the things you can't control, why worry about them? Because they're out of your control anyways. There's nothing you can do.

Yeah.

So, like, you can't really— there's really no valid reason to spend too much time worrying about one thing because Like, ultimately, you can do something about it or you can't. Either way, you'll be fine. It's a good book. I'm not doing it justice by speaking, by explaining in one sentence. It's a really good book.

I think someone in the chat put the title, so feel free to—

A Guide to the Good Life.

Feel free to rent the— Amazon gives free previews now. I don't know if it's old, new, or probably old, but you can get the preview for free, so check it out. Dude, everyone keeps asking if we've done drugs. We don't do drugs. It's kind of— what the fuck?

Do we look like—

I don't really do drugs. And neither of us, dude. It's funny. We've never done anything. Clean slate, sadly. Yeah. We got to change that soon. 2023 is the year, bro. How much money did you raise for Lexica?

We raised $5 million from Daniel Gross and Nat Friedman.

Yeah. So $5 million is a lot of money, right?

It is. Yeah. How do you— what do we do with it?

Dude, is that a lot of money now when you look at it? Because even for BuildSpace now, raised like a large amount too. It's like, how do you— does the money help? How much? And how much, you know, how much does it truly help? Like, for people at this stage who probably have not raised money at all and probably have this idea, they've— you all probably see all the time, like, $10 million, $5 million. Like, what does it mean?

There is a point of diminishing returns in like the millions. So like $1 or $2 million is like very useful. Like, when you get to $3, $4, or $5, you can only like move so much faster. Like, it's not a linear relationship between the amount of money you raise and how fast you ship things.

Yeah.

It's more like this logarithmic curve, or not logarithmic, it's like this curve where it's like, sure, you raise more and more, but are you really going to move faster than the company that raises $500,000 if you have $50 million? You might move a little bit faster, but definitely not 100 times faster.

Yeah. I think most people just need to have their needs met. One kind of hack that I have that I think I always try to stress to people is I always ask them why they aren't full-time on their thing. And it's almost always money, right? But then if we actually— like, most people don't have— their bill, monthly bill isn't that high. It's like most people who are starting stuff, um, if you have a family, completely different scenario, right? But if you're like single or you like, you have like, you know, a partner, like your bill is maybe $5,000 a month. Like $5,000 a month probably. Like when I was in SF, my bill was $2,500 a month. Wow. For food, for living, for health insurance, everything is $2,500 a month.

The cool thing is, uh, besides cost of living, like all the tools you need to run your company are like effectively Until you reach a certain scale.

Yeah, that's the crazy part. Technology companies are effectively free to run. Like, that's crazy. So it's like all you really need to care about is the monthly amount. And like when I actually figured out my monthly amount, which was $2,500, you know what I did? I literally went up to Furqan. I did $2,500 times 12, and I was like, Furqan, can you give me that? That was $25K, and that's how I started it, right? But it's like, turns out you don't need that much money just to work on stuff.

Like, I will say the reason we raised, uh, $5 million is because GPUs are expensive and we're doing a lot of training of models.

Yeah.

So, it depends on whether your business is capital intensive or not.

Yeah, yeah.

But for most software companies, they're not. Like, you can run them on like the Vercel free plan for like a pretty long time.

Yeah, this is a good question. I think both me and you kind of struggle still with this. How do you balance having focus on one idea when you have multiple curiosities?

That's a good question. I think a good answer is probably finding a way to incorporate those curiosities into your company.

Yeah.

So I am pretty interested in language models in addition to like Stable Diffusion. And I think language models are like incredibly interesting. But the cool thing is like we actually use language models for Lexica.

Yeah.

So that's pretty great. We get to like use this thing I'm curious about in our company. That's awesome.

Yeah.

It's a bit harder when the thing you're curious about is like not really related at all. So like if I was interested in like crypto, for example, it'd be kind of harder to integrate that into the company.

Yeah, similar. Same with Build Space actually. Like Build Space at its core is it's kind of like a selfish project where it's like because like Build Space is the perfect way for me to explore the shit that I think is awesome. Oh, like it's actually kind of selfish. That's cool. Think about it, right? It's like I got to focus on crypto. Now I get to focus on machine learning and crypto. Now now we're gonna have like four other things. Like I want to help people make EDM music. Like wow, that's great. Like I wanna I care about these things, right? I think most the people work at Build Space is pretty similar. It's like This is— this company was my hack to work on what I wanted to work on.

Nice.

But like—

That's smart.

Thing is, like, you having— I think most people look at multiple curiosities and multiple, like, being scatterbrained as a weakness. Like, oh my God, look at Airbnb, focused idea, focused founder, you know what I mean? But at the end of the day, most people are pretty scatterbrained. And, uh, ignore what the media tells you, I guess, about like how other— like, until you meet the person directly, don't have any assumptions about how they operate. But But yeah, I mean, it's just so— it's okay to be scatterbrained. In fact, it makes the thing better. But it's all about like, it's not a weakness. It can be a strength if you leverage it. So, but, uh, but yeah, yeah. BuildSpace is the hack. Lexica is your hack. That's actually how I see companies, man. That's how I see what starting a company is. It's like a hack for you to work on what you think is awesome full-time.

Yeah, that's a good way to put it. Yeah. It's really just a way for you to like not be degenerate and work on the things you want to work on.

Yeah, well, at the same time, creating like there's this incentive mechanism and it's called stocks and it's called like valuations and it's like that incentive mechanism exists to because there needs to be some real world value. Yeah, right. So it's pretty cool. It's like a hack. I just see starting companies as like a hack or not starting companies, but working on things, building things. All right, let's do like two or three more questions. Sharif, how do you deal with judgmental looks on people's faces like your relatives? when you tell them that you dropped out? Or yeah, actually, just generally, like, how do you deal with judgment when it's like you dropped out? Or like, you're working on Debuild for 2 years. I'm sure your dad called you up like multiple times. It's like, Sharif, it's still not working? Or Sharif, I don't get it. Like, the number of times my dad pings me and it's like, I still don't— why isn't BuildSpace bigger? Like, only 100,000 people? Like, only 200,000 people? And it's like—

That's a good question.

Like, how do you, how do you, uh, how do you stay grounded in that sense?

I mean, You kind of just have to have this inner self-belief that what you're doing is worthwhile regardless of what people think. And I think it is hard. It is hard when your family is like, oh, you're working on this thing? Why don't you go back to school or get a real job? But ultimately, you have to be willing to just not take it to heart. If you truly think what you're working on is important, then I think you'll be fine. It's really just like, this relates back to Stoicism and like the book A Guide to the Good Life. Like, you should care what people say to you, and especially if it's criticisms, but only to the extent where you can learn from the criticism. Yeah. And I think that's really as far as it should go.

Yeah, dude. It's like you shouldn't ever take it personally. Like, if that person looks at you and it's like, you've been working on it for 2 years? 3 years, and then you shouldn't feel bad about it. Like, don't take it personally.

Yeah, maybe they have a point. Maybe you should like ship faster.

Maybe they did. Maybe they have a point. Yeah, it's like, huh, like maybe I should be, uh, shipping. Maybe you're right, huh? But yeah, I think that's like one thing about just getting— about putting more years into this game, which we'll just call like the game of building stuff. Like, the more years you put into the game, one, one reason you get better at it is because you get better at listening to people. Yeah, because when you're 19 when you're 21, when you're like that, that at least me, I'm not generalizing all 19-year-olds. But like when I was 18, I was like, I was kind of just like a hothead. I was like, I don't care anybody's saying this is going to work. You're stupid. I'm the smartest guy ever. And like, and I wouldn't listen to anybody. Right. But now, now very different. Right. Because I'm like, you realize that doesn't work. But anyways.

Yeah. It's a balance between like listening to people's criticisms and learning from them, but also Having inner self-belief, it's like really hard to maintain the two and like walk that thin line.

Man, great question. Of the five million raise, how much is your own salary? I'll answer that for you. It's one one million dollars a month. One million dollars a month. So about five months of runway left. I hope you figure it out in five months.

More like two months of runway left.

Oh bro, that's crazy. Yeah. Yo, love love the new Ferrari. Thanks. I like this question. I like this question. Any advice? Yeah, any— I feel like this is kind of a strange one. Any advice on dealing with imposter syndrome?

On dealing with imposter syndrome?

So real quick, you probably still deal with it.

Yeah, I think most people do.

I woke up with it this morning.

Yeah, I think, I think it never really goes away. And I mean, like, it's— I don't have a good answer to this, really. I think it's just something you live with and like you learn to get used to it. Yeah, I think it helps when you're— when you feel confident in the things you're working on. It sucks when you don't feel confident in the things you're working on. It can be really hard to deal with. But yeah, I don't know if there's like a clear answer to this. You kind of just like deal with it. And I think almost everyone deals with it, which is kind of comforting in a way where you're not alone.

I think the best way to deal with it is by not taking it personally. Yeah. Like when you yourself are feeling like an imposter, like you don't take it personally. You don't think you're shit, right? You're just like, okay, I'm feeling like an imposter this morning. Bro, even this morning I was like, what's going on right now?

Yeah, use it as motivation to work on your company so you don't feel like an imposter anymore. But the thing is, the goalpost keeps moving, so you'll always feel like an imposter.

Yeah, dude, isn't that the worst part? I feel like that's the worst part of it all that nobody understands until they get there. It's like, once you've reached, like once you've raised $1 million, you're expected to raise $10 million. Once you have 10 users, you're expected to get 10,000. Once you've done all this stuff, it just keeps moving.

So you'll— Yeah, like imagine being Mark Zuckerberg, Like one quarter of the world using your products, but everyone thinks you're a fail.

Yeah, isn't that crazy?

It's nuts. Yeah, yeah, it's crazy. Dude, this guy has one quarter of the world using his thing, and investors are like, "How come it's not two quarters of the world?" Come on, dude.

It's like Oculus is trash. It's like man, the world is so. I guess the word is brutal, but it's I guess it's all about not taking it personally because once you don't take it personally, it's it becomes it becomes easier to like. Like, what's the word? Like rock in a way.

Yeah, it's— you just, you need to use it as fuel, really. Like, if you can take negative things and use it as motivation for your journey, I think you'll find yourself being in a much better place than if you use it as like criticisms or consider them as just criticisms.

Yeah. So interesting. Yeah. Let me see any other questions here. Oh my God. I mean, this is so funny knowing you. What is your morning routine? What is your morning routine for wellness? Do you do yoga? Do you work out? Do you meditate? Sharif, what's your morning routine for the people? How did—

what'd you do this morning? So this morning, uh, I, I woke up. By the way, you're—

I was getting dinner with— I was getting dinner with you and your girlfriend, and he was like, yeah, he woke up at 8. He slept at 8 AM today. But yeah, what's, what's, uh, in all honesty, what's like your morning routine?

Uh, so my morning routine is like very, very not interesting. It's just like wake up, eat breakfast, and then go to work. And that's really it. And like at work, I just like check emails and then start working. It's like not that interesting. I don't know why people ask about this a lot. Like I, for wellness, like I, I do need to exercise more. I don't exercise enough, but I hopefully wanna start exercising every day. Um, I like generally try to eat pretty healthy, so like I don't eat any sugars or I try not to. Um, but other than that, it's like, it's, it's really just like wake up and then eat some food and then move on with your life.

Yeah, dude, you're not gonna make it, I think, because you're not waking up at 5 AM, you're not taking a cold shower. Are you eating like 5 eggs in the morning? No, I'm eating— dude, I mean, carbs in the morning gonna ruin you. Are you taking like Nutrients? I take it, bro. It's over for you. Like, pack it up. Fuck, dude. Like, go, go roll on your Ferrari now. Like, it's over.

I'll return it to the dealership tomorrow. No, no, but it's—

I think it's correct. I think a lot of people like over, uh, over— I think both me and you don't have like the classic grind grindset type thing. But we're both extremely disciplined, though. I think that's what we both have for sure.

Yeah, it's also a marathon. If you're stressing yourself out every day trying to work 18 hours a day, it's not going to work out. You have to ultimately think of it as something that you can do. Your routine has to be sustainable for many, many years to be successful. There will be weeks or days where you kind of do have to work many, many days or nights in a row and work weekends too. Ultimately, it's like if what you're doing on a day-to-day basis can't be done for five years, it's probably too stressful or not not good enough. Yeah.

No, I think the best thing that anybody can do, even just even the people that are here, it's like just show up. Like I think you show up. Like even on days that you don't want to be there, you're still showing up. Yeah. Like you wake up and you and you show up at that computer desk and you're like, even if I don't get anything done, I'm gonna show up. That's good.

I mean, yeah, I respect that.

But like um. I don't know. Do you feel like that's you for the most part?

I do try to show up. I generally, even if I don't feel like working, I will probably still try to work.

So I want to stop the questions, but there's so many funny questions.

Okay, ask the funny ones.

Do you see yourself in 10— what do you— where do you see yourself in 10 years?

Being a VC? Oh dude, no way.

What? Hey, watch the tone. All right, watch the tone on the VCs, bro. Yo, sorry, sorry to all the VCs in the chat. All the VCs in the chat, I'm so sorry for this guy. I apologize.

And I— yeah. No, definitely not being a VC. I mean, like, no offense to all the VCs. I've met a bunch of them. And like, to be honest, very few of them were like actually interesting in any way to me. But yeah, I mean, they're like— I will say, like, without VCs—

It's pretty hard to be a good VC.

The whole startup ecosystem wouldn't exist without them. So like, I definitely— they need to exist for a reason.

Yeah, VCs get a bad rap because there's so few of them that are good.

Yeah.

Just like founders get a bad rap because 99% of founders fail, right? It's because only 1% to whatever percent of them are good and actually end up putting in the effort, right? The other whatever are just like— they're just not there yet, right?

Yeah. So yeah, most people watch the HBO documentaries on Elizabeth Holmes and extrapolate that out to every person who's a founder.

Yeah. There are some really good VCs. I'm sure you like your investors.

Yeah.

Big fan of my investors. So there's some really good ones out there. You just got to find them. What's a good VC? Founders Inc., baby.

Founders Inc., yeah. Founders Inc. Yeah, Founders Inc. is pretty nice. Good VCs. I mean, generally, I found angel investors to be a lot more helpful and interesting than VCs. Like, VCs for— VCs invest other people's money. Angel investors invest their own money. So I think even though they invest their own money, it's usually smaller amounts, but they generally are more helpful and like have more skin in the game than a VC.

Yeah. Dude, it's kind of a deep question. Have you ever cried while working on something?

Have I ever cried while working on something?

Yeah. Like, have you ever cried out of this pure emotional, like, destruction?

Pure emotional destruction, like joy or like despair?

I don't know.

Cry. Ah, have I ever cried? Maybe. Maybe. I don't know. I feel like, I feel like if I did cry, like, have blocked it out and buried it deep inside of me. So I, I don't really know.

Yeah. This person also wasn't happy with your 10-year question answer, by the way.

My 10— so where do I see myself in 10 years? It's just not happy.

It's like, that doesn't answer where he sees himself in 10 years.

Hopefully still working on Lexica. I think that'd be pretty interesting. Hopefully we're building some pretty cool creative tools that a lot of people are using. I think that'd be a good place to be in 10 years. Yeah.

Yeah. Yeah, I've also— you never cried working on something? Really? Interesting.

I've cried at movies before.

I don't see yourself as a— I don't see you as a crier. I don't see you as a crier.

I think the last movie I cried on was—

Interstellar. Yeah, probably. Do you cry? I think he didn't cry during Interstellar. I don't know.

I feel like it was like the Toy Story scene, Toy Story 3, when they're on the conveyor belts.

Fuck, dude, that one hits.

Yeah, dude, that one was sad. I thought they were actually going to die.

Fuck, dude, I cried last week watching WALL-E. I just remembered, dude, I'm turning into like a little dude. I got to, got to man up. Just kidding. Okay, I guess like let's just kind of like wrap it up here. We kind of went over your whole like builder like life, and a lot of these people are either starting it, they're continuing it. All these things that you did from Vector Dash to getting into YC to working on crazy projects in college to to failing in D Build to failing on Vector. Like there's a lot of fails in this story. There's a lot of Ws. Like Vector Dash could be seen as a fail, but I just see it as a W, right? Where it's like oh my god he. Stuck through it. It didn't end up working out, but they made something that like you had like tens of thousands of people playing games like using it, right? Like it didn't— sure, ended up didn't working out and you didn't end up being Mark Z from Vector Dash, but that's fine because it's still like a W because you, you, you saw it all the way through. Like that to me is what a W is. If you didn't see all the— all it— if you didn't see it all the way through, then that's an L in my opinion, right? Um, all right, yeah, whatever. That's what I classify as an L. But I don't know, when you kind of look at everything you've done And now you see what you're working on now. Have all those experiences influenced what you do today, basically?

Yes, of course. It's like collecting all the Infinity Stones. Each failure is like another stone on your gauntlet. And it's not until you have all 5 that you can finally snap and start the successful company. I do think they're massively important because you learn so much from each one. The skills you have in each one of those companies—

That's the best metaphor I've ever heard, straight up.

Yeah, I mean, it's necessary. Like, it'd be great if like the first time you tried something, you succeeded like incredibly, but for most people, that's just not how it works. And I think it's okay. Like, as long as you're working on things you believe are useful and important, and you're like pursuing things that you find to be curious, and you're just like being true to yourself, ultimately, that's what matters. And whether you succeed or fail, sure, you might fail 10 times in a row before you finally make it big on the 11th try. No one really knows.

Yeah.

It could take one try, it could take 100. As long as you're enjoying the journey, I think that's really what matters.

What if you're not, though?

If you're not, then stop doing it.

And then what?

Do something you enjoy.

What if I don't know what to enjoy?

Try many different things. Go wide and then focus on the things that you find interesting.

But how do you pick if you have like 30 different things you don't even know?

Yeah. How do you do what you don't know? I don't know. Find people who are in similar situations as you.

Yeah. I think that's the biggest—

See what they're doing.

That's the biggest hack. When you don't know what you're doing, Or you have so many options. Yeah, man, the best thing to do is to be around other people that also don't know what the fuck they're doing. Because turns out the combination of what the fuck am I doing with many people in the same room, like even 3 people, turns out that helps a lot. It does. Yeah, because even like me and you, like we've had the same group. It's like me, you, Johnny Frickon, and like we just talk amongst ourselves. Like we're the— we're kind of like an echo chamber, like the same 7 or 8 people from like last couple years. But it's great because I don't know, dude. Sometimes you have down days. I have down days. Johnny has down days. Furkan is like— Furkan never has down days, bro. I don't know what the fuck that guy's on. That guy's just built different. But also has like 15 years on top of us, just like working-wise. But yeah, I don't know. I think that is the best hack. I get that question a lot. Like, all right, Farzad, like, sure, I can do whatever I want and I can explore these ideas, but I don't know what to pick. And I don't even know what I like. And it's like, man, I think the best thing that you could possibly do is to find someone else that also is in the same spot as you.

Yeah, it's really, it's really important to find people working on similar things, or at least like have similar goals and interests as you, especially when you're a founder, just because it's such a, such a grueling journey and doing it alone sucks. Having co-founders is also really helpful too. Yeah. But if you don't have co-founders, like the next best thing is having other founders who are also working on early-stage companies.

You know, this final question, the final, final one, because—

I'm ready.

I feel like we've been here for a bit now. How long have we been here? Like an hour and a half? Not that bad. Still got 260 people. Thank you all for being here for this final question. All of you, nights and weekends as too. Sharif, would you Thanos snap the world?

Would I Thanos snap the world?

Yeah. Would you wipe out half the population right now if you had the 5 stones?

I would not, no.

Okay, thank God, dude. I was so worried about it.

Dude, fucking— If I was Mark Zuckerberg, it may be another— might be another answer, because then I could say I have 100% of the world population using my products. And maybe, maybe his investors would be happy with that. I don't know, bro. Mark Z would Thanos snap. He would. Yeah, he would.

Jeffrey, if you're in here, I know you want to Thanos snap the world.

I hope you change your answer, buddy.

Um, yeah, anyways, all right, man, thanks for coming on.

Yeah, appreciate it. Thanks for having me.

Thank you all for coming here. See you on the workshop on Friday. So next weekend, next weekend's S2. Good luck, guys. See ya. Bye. Peace.