He got kicked out of Harvard—then grew to $5M ARR in just 3 months. | Roy Lee, Co-Founder of Cluely

Roy Lee (00:00:00):
If you want to build a company, then I think the answer to building the company always comes in doubling down on the thing that you are very good at. For you to be in the heart of the audience, you must start being in the mind of the audience. And to be in the mind of the audience, you must be in the eyes of the audience. And that means necessarily go viral and go viral necessarily means be controversial. The ability to snap your fingers, record a 10 second video and have that be seen by a hundred million people. This is never before been possible in human history, and you have no idea how powerful that can be. Cluely launches April 20th, and right now it's been almost three months now. And Cluely has gone from zero to five in that time. I think I might personally know the Twitter algorithm better than any single person in the world right now. Yeah, I mean, especially on Twitter, we literally signed million-dollar enterprise contracts just because people think my Twitter is funny.

Previous Guests (00:00:44):
That's product market fit. Product market fit. Product market fit. I called it the product market fit question. Product market fit. Product market fit. Product market fit. Product market fit. I mean, the name of the show is product market fit.

Pablo Srugo (00:00:57):
Do you think the product market fit show, has product market fit? Because if you do, then there's something you just have to do. You have to take out your phone. You have to leave the show five stars. It lets us reach more founders and it lets us get better guests, thank you. Roy, welcome to the show, man.

Roy Lee (00:01:12):
Thanks for having me, brother.

Pablo Srugo (00:01:13):
Dude, so I'm excited for this one, man. Obviously, I watched a bunch of the viral videos, downloaded the app and using the app actually quite a bit. And I must say, it's pretty freaking awesome, dude. Like, you know what I mean? It's just right there. I mean, obviously, I use OpenAI. I use a bunch of other transcription stuff, but just the fact that it's there always on, like I find even with. Sometimes with founders, I was talking to the founder earlier today and he was pitching me a startup. And as he was going, I'm like, you know, what are the right competitors? What about what he just said? You know what I mean? And it's just so contextualized. It's very easy to use.

Roy Lee (00:01:43):
Yeah, 100%. That's the whole thing we're building for, we think more contextual AI is inevitably the future. Everyone thinks it's, some people think it's VR, some people think it's glasses, some people think it's pendants, and I think right now there's so much money to be made in the computer. And right now, I think it all starts at the computer. So that's where we're at right now.

Pablo Srugo (00:02:04):
So look, I mean, you have a great product, but I think what you have even more so is insane distribution. So I want to go through, like, in very detailed ways how you do that. But first, you also have an insane story. So why don't you do that? Why don't you tell us a bit about, how it all started? You know, the stuff that happened with Harvard. The stuff that happened later on, like, yeah.

Roy Lee (00:02:20):
Yeah, sure. Senior year of high school, I got into Harvard Early Action. Later that year, I got rescinded from Harvard for a suspension that I received for sneaking out of a school field trip. Specifically, I get rescinded because they didn't report the suspension on my application. But that happens, and eventually I spend a year thinking this is fucking, like, God's gift to me. And he's sending me a message saying that I'm not meant for college. And for a year, I think like,

Pablo Srugo (00:02:51):
Did you even want to go? Did you want to go in the first place? Like, self-sabotage?

Roy Lee (00:02:53):
Well, yeah. I mean, I definitely wanted to. Well, I thought I was going to go to Harvard, but if I were not to go to Harvard. Then I figured, you know, at least God is showing me that I could have, and now he's taken away from me so that I could do something more important. But sort of, I pushed back on my parents a lot about this, and they sort of really wanted me to do the whole college thing. So, the compromise was I would go to California to go to community college here. Eventually try to make my way over to UC Berkeley to make my parents happy, but I'd be in the heart of technology. But I ended up getting into Columbia actually while I was at community college. And then, I sort of went to Columbia with the intention of leaving as soon as possible. As soon as I found like, a co-founder and a wife.

Pablo Srugo (00:03:34):
When did you decide that you wanted to. Cheers on the wife by the way. When did you decide? You decided you wanted to start a startup and that was the thing you wanted to do?

Roy Lee (00:03:41):
Yeah, probably. Maybe a tech startup was, maybe about three years ago. But a company, man, as soon as I knew that you could be, like, you could start a company. I knew that I was going to be starting companies.

Pablo Srugo (00:03:53):
So, in high school? Even going to Harvard, you knew you wanted to start a company coming out of it?

Roy Lee (00:03:56):
No, like going to elementary school. I knew that I was going to be starting companies. I was growing up flipping shoes, drop shipping. Doing all like the classic money making things. I didn't make money on software for a while, but as soon as I did. I realized, okay, this is the thing to be doing right now.

Pablo Srugo (00:04:12):
Got it. So you get into Columbia, what happens then?

Roy Lee (00:04:14):
Pretty much the second I land, I'm looking for a co-founder and a wife. I'll tell you, there's not many dateable girls at Columbia. It's pretty rough out there. So I didn't have too much success with the wife, but the co-founder.

Pablo Srugo (00:04:26):
Why the wife piece, by the way? It's not a normal thing most college students go in looking for.

Roy Lee (00:04:31):
I think people really undervalue, and perhaps this is like one of my bigger takeaways, but in almost all aspects of life. People really undervalue the social element. Most companies die because I think they get the culture wrong. Because it's not fucking fun to work at your startup where you're grinding 13 hours a day with people you don't even like. And in college, you have the unique opportunity to regularly be forced to interact with like a hundred people, thousands of people your age, and every single one of them could potentially be your best friend. And the person that you actually want to be building a company with. I think if it is true that. Well, so if it is true that this is the most socially lubricated environment that you will ever be in. Which it is, then you should take advantage of that to find the two most meaningful relationships that will be in your life, and that is the one with your co-founder, and that is the one with your wife. And other than that, you do not get the opportunity to go to fucking French class and talk to the pretty girl sitting two rows across. That does not exist when you're an adult. That only exists when you're in college and you have the final few years of your undergraduate life to make that happen. And I think most people really, really undervalue what it means to be. The benefit of socializing. When your company culture is really social and we really try hard to make it. Then ideas flow better, viral ideas flow better, it's funnier, everyone wants to work harder, and thee. Generally it's just a better, more fulfilling life. All this to say,

Pablo Srugo (00:05:48):
What do you think makes you? What do you think gave you so much clarity on the importance of those deep relationships? Because I think most people, they get they get the first part of it. Yeah, college is where you meet a bunch of people. So the goal is to meet a bunch of people and party, and all that other stuff, right? Versus meet the most important people you'll ever meet.

Roy Lee (00:06:03):
Yeah, I mean. I guess my time, I spent after I got rescinded from Harvard. I spent a full year just locked in my room and I talked to pretty much nobody. And if you ever meet me in person, you will see that I am probably one of the most extroverted people that you will ever meet in your life. This was mentally draining and it made me realize, like, Jesus Christ. I cannot believe how difficult it is to be lonely and isolated. And since then, I think I really stopped taking things for granted. Stopped taking the ability to socialize for granted.

Pablo Srugo (00:06:31):
Was that on purpose by the way? You locked yourself away for a year or just like you felt your friends were off to college and you weren't?

Roy Lee (00:06:36):
Yeah, exactly. My parents run a college consulting company. So very ironically, I mean like, you can't let it be known that the son of the college consulting company just got kicked out of Harvard.

Pablo Srugo (00:06:45):
Got it. So you're in there, you're looking for a wife, you're looking for a co-founder.

Roy Lee (00:06:48):
Yep.

Pablo Srugo (00:06:49):
What happens?

Roy Lee (00:06:50):
I talk, I asked about 50 dudes, probably, Hey, do you want to start a company with me? And 49 of them say no, but one of them says yes, and that person ends up being Neil. Neil and I get together and, even he thinks I'm weird at first. I guarantee you, like, over 60% of Columbia probably thought I was weird before, I got here. I would always go out and try to talk to people, and like, yo, you trying to hang out, you trying to get dinner at my place with like a bunch of random strangers and, nobody would be down. Look, people in my generation, they're so fucking COVID brained and TikTok brained. They don't know how to interact with real people, and when. I bet if I was, if this was the 1980s and it was just a bunch of white grandpas walking around. Everybody would be down, like, oh my God, look it's the extroverted Asian kid. And now in Columbia, I'm a fucking weirdo.

Pablo Srugo (00:07:26):
Dude, did COVID really impact? I'm actually curious. Tangent, did it really impact your generation that much? Because you're, I'm thirty four. You're what? Twenty four.

Roy Lee (00:07:33):
Yeah. Twenty one.

Pablo Srugo (00:07:34):
Twenty one, okay, even younger. Yeah, so what is the impact of COVID you find in your in your generation?

Roy Lee (00:07:39):
I mean, you could just tell. Like, people don't know how to fucking socialize.

Pablo Srugo (00:07:41):
Interesting.

Roy Lee (00:07:42):
And I feel like people don't know how to sit down and do hard, deep, meaningful work. And people don't know how to socialize, and talk with each other. And everyone is so fucking like, goldfish brain. They have the attention span of maybe like three seconds max. Everyone cheats on all their assignments. Ironic coming for me, but like.

Pablo Srugo (00:07:56):
Dude, cheat on everything. What do you mean?

Roy Lee (00:07:59):
They don't have interest beyond school, you know, like, you cheat on your homework and then use the time to do something more meaningful. But instead they just like sit there and scroll and they. I'm quite bearish on my generation.

Pablo Srugo (00:08:10):
You know, it's funny, though. Some of those those things you're saying, that they don't think about anything beyond school. That's true, my generation, too. I think, if anything, the only thing that maybe the big difference in COVID is a real difference. The other one is just, there's so many more outlays for that, you can waste. You can always waste more time through time. You know, right now, we're in the peak of time wasting. You probably have to waste even more time in the future. But maybe that's a big difference. You can really scroll through hours.

Roy Lee (00:08:33):
Exactly, bro. The COVID brain is, it's a terrible thing. Anyways, Neil and I get together. And for the entirety of my time at Columbia, we're just hacking on a bunch of things. We try and do a blind dating app. We try and do an AI note taker or an AI lecture note taker. We do a bunch of bullshit, and then we eventually land on. What we think is going to be our big idea. Which is this, we call this Drink, and this is a sales AI-generated sales leads for liquor distributors. And it's the most bullshit idea ever. It started with, we tried so many things, and at one point, we're like, let's follow the traditional YC advice. Let's go talk to users, and let's go pick a sector and find a problem. So we talked to bartenders, because we think it'd be fun to talk to bartenders in New York City. Worst case, nothing pans out. We'll be friends with all the popular bartender owners. Eventually, we iterate towards liquor distributors, and we iterate towards a problem that we think is worth solving. Which is, they can't find sales leads because New York has so many restaurants opening and closing every day. And that nobody can keep track of all that. So we'll just use AI to do it for them.

Pablo Srugo (00:09:32):
This is the distributors finding leads to sell their drinks into?

Roy Lee (00:09:35):
Yeah, exactly. 

Pablo Srugo (00:09:36):
Got you, okay.

Roy Lee (00:09:37):
That'll find me 10 new restaurants that I might be able to sell French liquor to. That's pretty much what we did, but it really didn't work, and the whole time my heart was actually stuck on a previous demo that we made earlier in the semester. That was a demo of a tool called Interview Coder, and what Interview Coder was. Was at first, it was just a LinkedIn proof of concept, like, hey, here's some cool technology that we found that lets you have an invisible screen overlay that could see your screen and give you answers during the middle of an interview without anybody knowing. And that's what we advertised. And all of a sudden, I don't know, we posted a demo. It went pretty viral, it did like half a million views.

Pablo Srugo (00:10:12):
On Twitter?

Roy Lee (00:10:13):
Yeah, on LinkedIn.

Pablo Srugo (00:10:14):
On LinkedIn. Okay, yeah.

Roy Lee (00:10:16):
And at the time, we just kind of put it to the side and thought like, okay, this is way too risky. Let's not do this, because it's going to get us both kicked out of school and blacklisted. But then Drink is, like, not working. We bust our ass, and after the fucking 10th failed sales call, I'm like, man, I'm shitty at sales. I don't know jack shit about liquor distribution.

Pablo Srugo (00:10:31):
What's happening, by the way? You're just getting like non-interest?

Roy Lee (00:10:34):
Yeah. I mean, the product doesn't fucking work half the time. Like, fucking 80 percent of the time the product doesn't work. The features are dog shit. It's not a real problem that we're solving. It was just lame. I did not enjoy it. My heart was already caught onto the viral demo that we made, and I knew in my heart of hearts. If you want to build a company, then I think the answer to building the company always comes in doubling down on the thing that you are very good at, and for me. The things that I'm very good at are extreme tolerance for criticism, and I think a general extremely high appetite for risk, and I think viral sense. And these are probably my three biggest superpowers. And Interview Coder just aligned with everything that I knew I was good at. I knew that I could record myself using this on a big tech interview. I knew it would go viral and I knew that I could tolerate all the hate that came from it. And knowing all this, I just say, you know what, fuck it Neil, like I don't care if you're going to stop co-founding with me because you think this is crazy. I need to do this Interview Coder thing. So I do exactly that. I record myself using it on Amazon interview.

Pablo Srugo (00:11:31):
How much were you driven by the way. By just the sheer going viral of it versus this is a hundred million dollar opportunity company side of it, you know, what I mean?

Roy Lee (00:11:39):
There's literally like a brain dump on my twitter from months ago. Where I would post like saying, I know that this is not a billion dollar opportunity. For some reason, there was some driving part of me that thought, man, I need this to go out and I need to show the world that I built this thing. I'm generally a pretty like, I consider myself a very insecure person. I'm really, strongly driven by the desire to prove the world that I am better than they think of me, and every single hate comment I see like it up. It like propounds, and in my head it just drives me to like, I got to show every single one of these guys wrong.

Pablo Srugo (00:12:09):
This is always, always? Or this is after the Harvard thing or always?

Roy Lee (00:12:13):
Always, always, but every single time. I'm, like, something like the Harvard incident happens. I'm just like.
 
Pablo Srugo (00:12:19):
You double down.

Roy Lee (00:12:20):
Yes, exactly. They do not know who I am.

Pablo Srugo (00:12:24):
And dude, what about the virality thing? Were you already going viral on random shit before? Or why did you have such a keen sense of it?

Roy Lee (00:12:30):
I don't know, I think. I definitely scroll a lot, but I think even with how I scroll. I feel, like, I have a much stronger viral sense than most people, and I can tell pretty quickly when something is going to go viral or has viral components to it. I think it's gotten way stronger since I've been posting so consistently. I think I might have a better read, and this is the crazy. The most arrogant thing I will ever say, is I think I might personally know the Twitter algorithm better than any single person in the world right now. And if there is a superpower that I have, then it would be that. Well, at the moment, truly, I think nobody else in the world has the degree I do. It'd probably be that.

Pablo Srugo (00:13:04):
So we're going to dig into that for sure. But maybe, yeah, let's keep going with the story then. So you decide, you got to do this interview. What does your co-founder end up doing? Does he join or does he leave?

Roy Lee (00:13:13):
For a while he leaves. He thinks that it's not worth it. It's too risky of a venture, he can't.

Pablo Srugo (00:13:20):
He doesn't thrive on hate, you do. So it's kind of an issue?

Roy Lee (00:13:22):
Yeah, exactly. So I'm like, you know what, fuck it. I'm just going to do it, and I pretty much do it. I launch a paid version on February 1st. We launch it, and then in the interim from February 1st to the whole entire month of February. I'm just getting shit, shat on by Columbia. I'm getting shat on by Amazon. I'm like, doing at the time, interviews with coders making, well, like 3K a month. I'm like, damn, this is. My parents don't even know that I just got kicked out of school. What the fuck is going on here?

Pablo Srugo (00:13:46):
Walk me through that whole story. I mean, I know you've shared it elsewhere, but like you, you know, what's? Is the viral moment when you post yourself kind of beating? Beating a tech interview and on YouTube or whatever? What's like, how does it all play out?

Roy Lee (00:13:57):
Yeah, so I record myself and I am using Interview Coder in a full Amazon interview. And I end up passing the interview with flying colors. He gives me a ridiculous hard question. I end up passing the interview with flying colors, obviously, because I'm cheating. I post it online on YouTube. Amazon takes it down three times for copyright strike, whatever the fuck that means. But in the interim, it goes viral every single time, and everyone on Twitter is talking about this. And I end up being literally the front page of Twitter. An Amazon executive sees this. They send a letter to Columbia saying, essentially, you need to expel this kid, or we're never going to hire you from your school anymore. It's a very thinly veiled threat. Columbia sees the letter. They loop me into a series of weird disciplinary hearings where they make up a bunch of bullshit. Like, hypothetically, if the data structures and algorithms class was fully online, and remote, and on Zoom, and they conducted interviews. Would you be able to use this? It's the most bullshit hypothetical. 

Pablo Srugo (00:14:48):
Right.

Roy Lee (00:14:50):
They looped me into that, and I ended up getting suspended for a year.

Pablo Srugo (00:14:53):
But in the meantime, you're tweeting about all this stuff as it's happening live, right?

Roy Lee (00:14:57):
Yeah.
 
Pablo Srugo (00:14:58):
So you're just getting more and more viral.

Roy Lee (00:14:59):
Yeah, exactly. I mean, at a certain point, you're past the point of no return.

Pablo Srugo (00:15:04):
Yes.

Roy Lee (00:15:05):
Something that most people don't realize is, it was way earlier on when I decided, okay, I am fucking like, I'm about to get fucked. I need to find a way out, and right now, the only escape route is raising some money from D.C. and going off. Fucking off the Silicon Valley. Truly, when you're blacklisted from Big Tech and you don't have any meaningful opportunities at the school. What else can you do but that? So it was really early on when I decided, all right, I'm just going to full send this thing. I'm going to try to bait the Columbia admin into expelling me. To make a bigger scene out of this, and I know it's coming anyways. I might as well just fucking make a show out of it, and I do that.

Pablo Srugo (00:15:40):
So you're getting this. Do you get like, besides the thing going viral, people commenting and some people hating, and some people being for you. I'm sure like, are you getting inbound? Like, literally from VCs about that product, the interview product?

Roy Lee (00:15:52):
Not about the Interview Coder product, but like when I went out to do my fundraise. It literally took like 24 hours. I think there was a lot of interest from VCs, like this is a. When was the last time someone got kicked out of an Ivy League to try and build a tech company based around the thing that they got kicked out for. Probably never in human history. So I'm like, alright, I think I have a pretty unique story. It's at the seed stage, kind of back based on the founder, and I think you back founders that are spiky. And this is like definitely spiky.

Pablo Srugo (00:16:17):
So when you get kicked out, what do you do? At this point, are you going in? All in on Interview Coder or how are you thinking? You kind of already knew it wasn't a huge opportunity. What's the play at that point? You've gone as far as you can off this thing.

Roy Lee (00:16:29):
Yeah. I mean, at this point we're thinking, we need to pivot Interview Coder into something bigger, and we have a few vague ideas. Maybe we can just do Interview Coder for sales specifically.

Pablo Srugo (00:16:39):
This is by the way when? This is March this year or March last year?

Roy Lee (00:16:41):
Yeah, this is March this year. Everything you've ever heard about Roy Lee has happened in the last six months. But yeah, I mean, we're thinking Interview Coder, maybe we'll do Interview Coder for recruiting and float the scripts, or maybe we'll do Interview Coder for sales. But in reality, the more I think about it. The more we realize this is actually probably the correct UX for all AI interaction in the future. Why don't we just capture the biggest mark that we can and just fucking swing as big as possible? We already know that we can go viral because it's what I'm good at and if we can guarantee, get like a hundred million views a month on something that I might as well. That I would prefer to get a hundred million views a month on like the way bigger thing. The way bigger thing being the, just directly compete with ChatGPT.com.

Pablo Srugo (00:17:19):
This is in March. So when do you launch, your current Cluely product?

Roy Lee (00:17:23):
April 20th. So we spent about, like, four weeks heads down building, coding the product. And we literally do the landing page in like 24 hours before launch and we ship the shittiest version of our product. It literally doesn't work when we ship it, because it forces you to install this external audio driver and that doesn't work. And like, it is so bad. It is unusable for the first three weeks. But then after three weeks, you kind of figured out, okay, we just built, like, an in-house audio driver that gets installed in the bundle. We figure that out. That helps a lot, and then throughout the time. We are adding so many things. I think one of the biggest benefits of going so viral is now, you don't really have to talk to users as much. Because you can sort of see user data, and the quicker you get to see user data, like, I'd rather see 70,000 user data points rather than talk to seven users. And I think this is, it allows you to iterate in the correct direction way quicker.

Pablo Srugo (00:18:16):
And when? Was your seed round before or after launch?

Roy Lee (00:18:18):
It was before. It was literally during the launch.

Pablo Srugo (00:18:21):
How'd you?  Oh, so you had the virality and you just go up on Twitter? Do you hit up, like, how'd you raise your seed?

Roy Lee (00:18:26):
Yeah, so we decided to raise a seed. We scheduled a bunch of calls, but I didn't. We scheduled calls for three days, Friday, Saturday, Sunday. So sorry, like Wednesday, Thursday, Friday. But I didn't even get to the Friday calls, because I started on Wednesday and by noon on Thursday, I was finished. And the round was way over subscribed already. We went in thinking we were going to try to do $3 million at a $30 million valuation, but people kept coming in. We had to close it at one point. So we closed at $5.3 million at $30 million, but literally, you get. Two more hours of investor calls, and we probably would have done $7 million.

Pablo Srugo (00:19:02):
But, how did you get that by posting on Twitter? Or did you get that by going direct to some VCs?

Roy Lee (00:19:07):
Yeah, I had some friends that I made throughout the process, like, Cory Levy from Z Fellows. He introduced me to a lot of investors, and I had a friend, Adrian Allen. Who also is an advisor, and he did a lot of intros for us, too.

Pablo Srugo (00:19:21):
Got it, and then how? So you keep growing, I guess, you launched on 420, you said?

Roy Lee (00:19:25):
Yep.

Pablo Srugo (00:19:26):
Okay, and then you grow for a couple months. And then your latest, your A 15, million from A16. That was a couple weeks ago?

Roy Lee (00:19:31):
Yeah, as soon as we moved to California and San Francisco and settled down here. The partners just kind of visited, swung by, brought some steaks, and then pretty much immediately they asked, hey, do you want some more money? Sure, yeah, I'll take some more money. But yeah, and I think A16 is a good name to have, especially for a guy like me. You kind of need the big guys to step in and give you a backing. Yeah.

Pablo Srugo (00:19:56):
And how fast have you grown?

Roy Lee (00:19:58):
Okay, so Cluely as a company involves Cluely the product and Interview Coder is a subsidiary company. Interview Coder launches February 1st and right now it's like hovering at around $2 million ARR. Cluely launches April 20th and right now it's up in like a little over two months, probably almost three months now. And Cluely has gone from zero to five in that time.

Pablo Srugo (00:20:22):
That's crazy, man.

Roy Lee (00:20:23):
It's like, I think it's not unheard of in the era of AI, but it's pretty quick, yeah.

Pablo Srugo (00:20:28):
No, but in the era of the world, it's still unheard of. You know what I mean? It's just like crazy shit that happened last year. All right, man, so let's go really deep on the viral piece of it all. Maybe we can start at a high level, but I wouldn't mind going through some of the key moments, you know what I mean? But maybe at a high level, I mean, you made a comment, like, you understand the Twitter algorithm better than anybody else. So maybe you can expand on that as a starting point.

Roy Lee (00:20:49):
Yeah, I mean. I think I've always had like, much stronger viral sense or much stronger attunement with short form algorithms. Content algorithms, just because I've spent. I'm like Gen Z, I've spent my entire life scrolling fucking Instagram and TikTok reels. And most people do not like, when their brain. When your brain is so plastic, most people are not watching fucking Nigahiga Smosh. At that, most people are like fucking 18 when they were doing that at least the, right now, the current decision makers of tech Twitter and everything. They're a little bit older, and they're not as content brain as I am. 

So I come into Twitter with this content brain immediately I think, man this place is fucking slow and stupid, and posting funny, interesting shit. Why does it not look more like my Instagram Reels page? Because the Instagram Reels page is literally a hundred times more controversial and crazy than anything you will ever see on Twitter. And it felt like the most PC like tone down. I knew I can sort of cause a ruckus here if I just apply my learnings from Instagram, TikTok to here. That's pretty much exactly what I did. The launch video was sort of, the craziest thing we could post. That would have gone sort of, viral on Reels, but obviously Twitter's not caught on, and that's what we landed on. And ever since then, I've probably done 3,000 tweets. And now at this point, I can sort of tell within plus minus 25 percent exactly how many impressions every single tweet I post will do. Generally, I think with Twitter virality, there's three buckets of virality. 

Informational content goes viral, relatable content goes viral, and engageable content goes viral. And I will niche down, and I'll specify on all three. So first, informational content. This is when, let's say, ElevenLabs is launched, and it's pretty interesting. Or DD Doss makes a post, and it's like a news report. I'm like, oh, wow, the latest model benchmark, blah, blah, blah. It's genuinely interesting, it's like tech news. That will go viral, but it's very hard to spontaneously just make something like that happen unless you yourself are a research lab and have a launch. The only way you can rely on informational virality to go viral, is if you genuinely have useful or interesting information to go around. Then there's relatable virality, and relatable virality is like shit posts or, you know, posting like shipping a product on Friday at 5 p.m. It's like, these are almost not converting at all. Like, for you to be relatable with so many people, you need to cast the widest net possible. The best way possible, there's no actual conversion funnel there. But the last thing here, and I think this is actually the broadest category, is the category of engageability. 

And I think every single tweet you make with the intention of going viral, imagine someone quote posted your tweet because they have such a strong opinion about it. They quote post your video, your product, or your tweet, or anything. What, can you see in your head? Can you visualize very clearly, exactly? What strong opinion they will have? If you cannot envision that, then your tweet will not go viral. And this is the reason why controversial things do so well. Because, very rarely will you quote post something and say, man, I really love this. In fact, almost always you will quote something and say, like, I feel so strongly negative about this. I have to share the world my negative opinion. I have to shut this down, and that is what feeds engagement. And generally, you can make things more controversial, but things don't have to be controversial to be engageable. For example, the whole 50 interns thing. I mean, this is like more, rather than controversial it's shocking. And you can arbitrarily give things shock value or engageability and the way that you word things. You can you can make things more engageable. And I also think engageability, like, sort of the two tenets of engageability are. I feel most people on Twitter will make these tweets and nobody can fucking understand what you're talking about. It's this 200 word essay on blockchain, crypto, whatever the fuck. Bro, I don't know, I don't understand. It has to be understandable, the average person, average American is so much fucking stupider than you think. You have to, if a farmer from Kansas picked this up in the daily newspaper, would he be able to read your tweet and understand exactly what the fuck you're talking about? 99 percent of the time it's no, and if you cannot market to a hundred million people. Then you cannot get a view, then you cannot have a tweet that's viewed by a million people. That's just the way the algorithm is. And then the second thing is, it's just, you know, you have to have that natural component of someone will engage with this.

Pablo Srugo (00:24:48):
So let's talk about this engageability. How much do you spend, time crafting a post? And do you worry about the start of it? Because you think about YouTube, right? And how much that intro matters, or even I post a lot on LinkedIn. You've got to click see more, so that means your first line's got to make them stop. How do you think about all that, versus controversial is fine as an idea? But then how do you put it into practice?

Roy Lee (00:25:09):
Yeah, I'd say like 10 percent of, the way my flow goes for videos and everything is I spend the first like entirety of the thing is 100 percent of time. I spend 10 percent of time thinking about exactly what I'm going to post. This is like, okay, I got it. I'm going to make a video about, like, 50 interns. And I spent literally 90 percent of time mulling over in my in my head. How can I make this more engageable with you with like, a person? Generally, if you have something to post, you can convey this idea in probably one or two sentences. There's almost no idea that you cannot boil it down to one or two core sentences. And after that, the only things you add. You should add for the sake of someone might engage with this. Let's add this scene here and add an extra two seconds, but it increases the odds very highly that someone will engage. I think this is a low-key sense that you have to explain. Generally, though, I think that everything in the world has the ability to go viral. Everything in the world has viral potential. Just most people don't have viral sense to funnel that.

Pablo Srugo (00:26:03):
Yeah. Maybe walk me through, maybe it's top of mind, but think about whether it's a new launch or a new feature that you're putting out, right? To make it really concrete and how you might think about like.

Roy Lee (00:26:12):
I'll give you a sneak peek of, probably the next launch video coming out. So we're doing Cluely for sales, and this is gonna be a launch video for, like, Cluely works for sales calls. And the whole thing, if you actually do an actual sales call. On an genuine demo of the product. Then nobody's gonna fucking watch your thing. A, it's not interesting, and B, unless it's super informational and you just have like the most crazy fucking product experience ever. And like nobody's gonna be super interested in that. So, how's a way that you can convey the message that you can get real-time AI in a call? That's what I want to convey. That's the first 10 percent. I want to make a video that conveys the idea that real-time AI in a call will help you make a sale. And then after that, I think about the craziest video ideas. What is the craziest thing that you might have seen? Like, a drug deal. That would be fucking crazy, if I was selling drugs. What's something else, that's pretty relatable and crazy? I think we have a lot of Asian audience. Like, tech is fucking all Asian. So I might as well. Well, what if I'm selling drugs like this Chinese mafia boss? And then that becomes a script and we flesh this out. So now the next launch video you'll see, is me dealing drugs. You think I'm selling cocaine to a Chinese mafia boss, but in the end, it turns out I'm actually selling software. And here you will see there are literally six points at which there's a twist or there's something interesting that makes people engage. And the entire video, is snippets of Cluely inaction. But more so than that, 80 percent of the content of the video is literally just gonna be like random shit for people to engage with, and if you look at it through that lens. You will see like how much work we put in to make it engageable.

Pablo Srugo (00:27:39):
Yeah, how much do you spend on a video like that? How much time does it take to make?

Roy Lee (00:27:42):
We're really trying to, well, right now. Post-production takes way longer than anything else. It probably takes like a week or two to edit the entire thing. Script writing, planning and everything. Maybe if we work hard and that's my sole focus, like, three days tops.

Pablo Srugo (00:27:54):
What about at the beginning, right? Like, now you have more resources. You should professionalize all of it. I'm sure you've got a bunch of interns doing a bunch of videos, and you can choose, and pick which ones you can release, whatever. That's to much stuff, right? But like in the earlier days, you know, Interview Coder days. How, you know, less resources, less time. How do you still put out something, that could go viral when you don't have the ability, maybe to to do that. Like, production value or how do you think about at that point?

Roy Lee (00:28:21):
I'll say the first launch of anyone's, anytime you launch something. There will be an informational component to your launch, and you can guarantee that your launch video will probably go more viral than any of your feature demos. Because it's like some novel, it's something interesting, there's information being spread, oh, here's a new product that's being put out. You can guarantee that you will get some impressions on your launch video, as long as it's not like a complete dog shit video. And for me, what that was, was the, hey, Columbia just kicked me out for building this tool. And that first thing, it got pretty significant traction. It did quite well.

Pablo Srugo (00:28:49):
How viral are we talking about? These videos?

Roy Lee (00:28:51):
I think it did a little under 2 million impressions. The long thread I made about that, and that was great. Not so good for me right now, but at the time it was super great for me. And then after that, I think what everybody drops the ball in is, doubling down on the momentum. And if you cannot keep doubling down, see where people are engaging. This is why I say it's important to see what the quote post says. Because the quote post is what you will respond to, and now every single post of yours for the next.

Pablo Srugo (00:29:19):
The quote post is what? Somebody else taking your thing and saying whatever they think about it?

Roy Lee (00:29:22):
For example, someone will take a screenshot of that and say like, "Crazy how this kid is speed running, ruining his life." And then for the next five posts I make, I'll like directly reference how I'm ruining my life or something. You know, like my life is not ruined or my life is ruined, and it'll just give more for people to engage with. It will give people more things to engage with. I think most people will make the one launch video, and even if it does go super well. They'll be like, fantastic, we had a viral launch video. Now let's get back to building. What the fuck? What are you talking about? You have the most hottest algo pool in the world right now. Everybody's hubmind, they're thinking about your product and you are going to spend this time not fucking doubling down on a potential distribution goldmine you have. Instead, you're going to like, fuck off. What are you possibly talking about? The benefit of a launch is that it gives you the right to try again and try to go viral again. And if you can go viral enough times, then bro, you you will end up like Cluely.

Pablo Srugo (00:30:11):
Well, yeah, I don't think most. I mean, the vast majority of founders just don't think about the power of virality. The other question I have, though, is do they have to? Like, you have now become this controversial figure, and you like it. And you thrive in that, you know, what I mean? Could you really get this kind of reach if you don't want that? Or it's kind of like, you either want to be this. Or it's just not worth it, because you're not going to get it.

Roy Lee (00:30:33):
Yeah, you can't. It's impossible.

Pablo Srugo (00:30:34):
Yeah, that's what I thought.

Roy Lee (00:30:35):
At the end of the line, if you are a big creator who's actually relevant. You are only big because of, truly at the end of the line. 50 percent of people in the world love you and 50 percent of the people in the world hate you. And you see this probably most prominently with the American political parties. I mean, there's a reason why elections are so close every year and why the American president has more power than anybody else in the world. Opinions are so divided and it has to be the case for you to be relevant. For you to be in the heart of the audience, you must start being in the mind of the audience and to be in the mind of the audience, you must be in the eyes of the audience. And that means, necessarily go viral and go viral necessarily means be controversial.

Pablo Srugo (00:31:09):
And then what about video versus text? Is it all video nowadays? Or are you still thinking about text in the same viral way? You're such a selfish person, I actually can't believe how selfish you are. Because you've been listening to the show, you listened to this episode, you loved it. You've listened to a bunch of other episodes and you haven't told anyone about it. You haven't told any of the many founder friends that you have about it. Think about how many founders have helped you out. When you're building your startup. So don't be selfish. Tell your friends about this episode. Tell them about the show and help me, help them.

Roy Lee (00:31:39):
Text doesn't really convert. It doesn't do that well. It's just, you know, maybe you can. The only reason for me right now, to make a text Twitter post is to keep my algo pool alive. But at this point, I have enough of a following that you can essentially eliminate the variance of, oh, this deserved to go viral but it didn't go viral. Now for me, I have enough of a base that, that's just not true. If I ever post a video that deserves to go viral, then it will go viral, and because I have sufficiently strong viral sense. Every single video I post will go viral. And then, yeah, after that. The only point of text tweets are to keep yourself alive in the eyes of the audience and keep your algo pool alive. But at this point, it's like.

Pablo Srugo (00:32:15):
What is it? Yeah, tell me more about algo pool. Because is that a real thing? My understanding was, if the content is great. The content is great, you know, what I mean? It should, you know, be counted by content. Is that not the case?

Roy Lee (00:32:24):
I mean, certainly not. I don't know if it's actually baked into the algorithm or if it's a psychology thing. But when you have an incident that goes super viral and, you know, you're on the front page of Twitter for being the kid who got kicked out of Columbia for building a cheating tool. Then all of a sudden, everybody is curious about you and you're hot on everybody's mind. And they're more likely to react to your post, more likely to visit your profile, more likely to see what you're doing and comment because they remember you. And the consumer brain forgets things very quickly. I'd say maybe 40 hours for any hyper viral moment to lose its relevance.

Pablo Srugo (00:32:55):
So it's more about. Yeah, it's less so about the actual algorithm. It's more about the fact that, if your name is hot. Then the next post, people scrolling, they're just more likely to stop. Because they recognize your name and they're like, oh, I want to see what this kid's up to next.

Roy Lee (00:33:07):
Yeah, I'd say that. I mean, well, I actually don't know the Twitter algorithm. I can only imagine that there must be something baked into there. Where if someone made a super powerful post. Their content gets pushed out more because the algorithm thinks that they'll make more interesting content. But yeah, I have no fucking clue. I'm not like a ML engineer at X. You have to ask Yaseen about that.

Pablo Srugo (00:33:24):
Walk me through the other video. Obviously, your classic launch video, cheat on everything. Walk me through all the thinking around that. Like, you had Cluely as the product, the idea. How do you decide to go with that tagline? How do you decide to go with that scene?

Roy Lee (00:33:36):
Yeah, I mean, people don't know this but I'm really. Even at the point of Cluely, I'm still riding the initial momentum of Interview Coder. That first post going viral. You have the fucking distribution potential goldmine that you get out of just making something very reactionary and then reacting to all the reactions over and over again cyclically. It will literally spiral you up into something where you never even knew existed. It started with me building the cheating tool for interviews, Interview Coder. That goes viral and then eventually I just keep doubling down. Until I build a product that has nothing to do with cheating pretty much and I'm just forced to call a cheat on everything. So that it can spark reactions and I can ride the algo pool of that initial momentum. People do not fucking understand. Nobody, I feel like nobody on Twitter gets how important it is to keep the momentum alive of the algorithm. Like, if you keep riding that boat, the boat will go so much farther than you think. Because everyone will have a unique reaction. Everyone will put their own blend on what they think it is you're doing. I see a few viral threads about calling me Mark Zuckerberg or something. So I immediately make the series A video, the fundraiser announcement. I make a social network shot for shot remake. Because I'm just reacting to the things that are said about me, and you can just keep doing this into infinity.

But, I guess, the actual launch video. I mean, we sort of knew that we wanted to swing really big with this next one, and we were in this mindset like, all right, what's the craziest thing that we can do? I know for certain that I want to make the next thing go viral. And it seems like Friend.com sort of showed everyone the playbook. You spend a bunch of money on a launch video, make it look really nice and cinematic. Give it a black mirror spin, and people will go fucking crazy. Like, what the fuck is this black mirror bullshit? And we don't even have to show the product at all. We just have to show the idea of the product. Which is AI in conversations. So, like, real-time AI that nobody knows you're using. So that's, sort of the idea that we got. Like, how can we do this in the most visual and crazy way? And obviously the most crazy, social contract that you don't want to be broken is a fucking date. And obviously we're trying to go viral in tech, so the date has to be with a pretty Asian girl. And I think something funny would be, to again, ride on the momentum of the Interview Coder thing. Would be to once again, point out how young I am and how I'm pretty much fresh out of college. Always harping back to that. So everything here is a combination of riding on my initial hype and also making it engageable. 

I work at Bananazon. It's a direct reference to Amazon, the company that I got kicked out of. We're on a date and I know the hard thing was like the art conversation was probably the only piece of that that wasn't like intentionally designed to, react to something. But if you go back and watch all the shots. Every single shot is screenshottable. You can imagine someone would quote post with a screenshot and say something. There's always an Easter egg in every single one of the Cluely clips. Even at the end, there's like a whole, oh, sorry, I got to go to an anime convention with my boys. It's like an Asian thing, we know that people will react. In fact, my one regret. I probably should have been more specific and talked about a trending anime that would have been more engageable. Like, generally you're Asian, you're in tech. You know what anime is, you will engage.

Pablo Srugo (00:36:37):
I do think the one way you think about it is so different is. I think, most founders are almost too literal with this stuff. They want to sell, the whole product is sell, sell, sell, right? Like, here's the product, how do we sell it? And you're just like, how do we just go viral, right? You actually look at your video, like, you said. I mean, the product isn't even really there and everything you're showing doesn't even exist. And yet it doesn't, it really doesn't matter.

Roy Lee (00:36:58):
Yeah, exactly. I think, sales is about a lot more than just being a little sale person hopping on a call and discussing how your product out like. There's so, especially with the consumer. The consumer is so fickle and you can go from being the coolest thing in the world to being the lamest thing in the world overnight. And, if you want to truly sell a consumer product, you have to be cool. And, I think it takes a lot. It takes a lot of knowledge, it takes a lot of consistent virality and you cannot ever be cool just by shilling your product. Nobody wants to watch an ad ever. So you need to consistently get a lot of engagement without people feeling like, I'm watching an ad. And I think this will be sort of like the future of marketing.

Pablo Srugo (00:37:34):
You think this works for B2B? Or really just mainly B2C?

Roy Lee (00:37:36):
Yeah, I mean. Especially on Twitter, we literally signed million dollar enterprise contracts, because people think my Twitter is funny. Pretty obvious that like, it's not just 18 year old Gen Z that are on short form now. Now literally everybody in the world is on short form. I think this is something that just the world, like the corporate world is not caught on to the fact that literally everybody is on short form. The CEO of every single Fortune 500 company is scrolling through fucking like Twitter, LinkedIn, Instagram, everybody scrolls. And if that's the case, the mastery of the short form algorithm is a mastery over human attention. Wars have been fought over this shit.

Pablo Srugo (00:38:08):
Well, I think the point is you want people to know you and have a point of view. And the ones that don't like you won't buy and that's fine. But at least there's some that do like you and they'll buy. The worst thing is like they don't know you or they're indifferent about you, in which case they're nothing. You know, they're basically zeros. Neutral, which is useless.

Roy Lee (00:38:23):
Sure. Yeah.

Pablo Srugo (00:38:24):
The other question I had is, you talk a lot about Twitter. What about like LinkedIn? What about, you know, YouTube Shorts? What about Instagram Reels? What else is working?

Roy Lee (00:38:31):
Yeah, probably Instagram and TikTok are our other two big levers. We're on LinkedIn. We post, but I don't consider it a huge drive for growth or anything. Everything on LinkedIn is probably downstream of my Twitter. And LinkedIn is like a fake social media. But IG, TikTok, we do quite well. I think we're averaging somewhere around 80 million views a month on there, and the way we do that. There's like three tried and proven things in short form, and that is a UGC content, influencer marketing, and clippers. So we have about probably like 400 clippers. We send them a content repository of every single podcast I've been on. Including this one, and they will just take interesting clips from that and clip it up.

Pablo Srugo (00:39:09):
You find that does well?

Roy Lee (00:39:10):
 It literally directly correlates with how crazy I can be on podcast.

Pablo Srugo (00:39:12):
I see.

Roy Lee (00:39:13):
But sometimes it'll be like morning and I'm a little less energetic. And sometimes I'll say something crazy about like fucking ancestry and women. And those more viral. This is not converting at all, by the way. The only point of this is for more people to hear the Cluely name and be like, what the fuck? Another Cluely thing on my for you page. UGC is probably our most converting. You can literally just like make up a hilarious, like make up some crazy fucking hook. Something that would guarantee go viral. Like, imagine I'm fucking Bank of America and, I'm releasing. I don't know, QuickBooks or I don't know, CISA. The trick, like, the key here. You can make a video, it's like, here's how to have sex with a hundred freshmen. I'm an accountant and I have so much time at work and I have all this work I'm going to do. So instead I use this Bank of America feature to offload all my work. And with all that extra time, I'm able to DM a bunch of freshmen so I can have sex with them. Like, what the fuck are you talking about? You're literally like these two things have almost nothing in common. But the hook, here's how to have sex with 100 freshmen. Like, what the fuck? That's crazy, that in itself is viral, and you can literally do this with any product. Just say something completely absurd at the start and make an extremely boost connection. That's sort of what we do with a lot of the Cluely products. It's very easy to do.

Pablo Srugo (00:40:15):
And you say it's converting. It's like, literally you'll hook somebody with something so tangent like that. But then ultimately, some will actually convert and sign up, and whatever.

Roy Lee (00:40:21):
Yeah, exactly. At that point, it's literally funnier the more you go and talk about the features. The products like this is so I can't believe a company doing this is everything about it is viral.

Pablo Srugo (00:40:30):
You think one of the, kind of, the negative takes where I would be like. This is like you can only be hot for so long. You know what I mean? If you push it like this, at some point you'll be like, oh, Cluely. Yeah, they just say crazy shit. I ignore it. Is that at all, I mean, your experience is not that. How do you think about that? Can you just keep this going forever? Just keep going viral?

Roy Lee (00:40:47):
Look at Mr. Beast, Jake Paul, Logan Paul. There's always more things that you can make more interesting things. The only time where this expires is, if we keep doing the exact same shit. I really don't think Cluely says like crazy shit. For example, I think the whole 50 interns thing. That's a pretty, it's a very novel concept, and if any company in the world did this, including Cluely. I think it would independently have gone viral. Like, wow, that's a very crazy experiment that you're trying. There's so many crazy experience that we can try, and even if you know Cluely is the crazy company. Still, if we do something that is like genuinely pushing the envelope, you have no choice but to react. For example, the fundraise announcement, you know, everybody knows Cluely is like the guys who do crazy things. But still, like, damn, really? A movie trailer for a fundraise announcement? That's crazy, you're comparing yourself to Mark Zuckerberg. It's like, as long as you keep doing interesting things. People will not care that you are the guy who does interesting things. They will just see the interesting things and that'd be interesting.

Pablo Srugo (00:41:38):
That makes sense. The other thing I wanted to ask you about. I saw this post on the UGC thing. How are you setting that? Well, actually two questions. First, just on the intern thing. Are you really hiring 50 interns? Or is that just a catch line and then you'll figure it out?

Roy Lee (00:41:50):
It depends on what you determine to be an intern. If you think someone who is a creator. Who makes content for us part time, gets paid for making content and the content directly converts into like, marketing expense. Then yeah, we probably have over like 500 interns. But if you think about interns, like Justin. The traditional brick and mortar, you come in and do work for the company, like that sense. Then we have like five.

Pablo Srugo (00:42:11):
Got it.

Roy Lee (00:42:12):
It's pretty arbitrary. But 50 makes people shocked.

Pablo Srugo (00:42:16):
Somewhere in between 5 and 500. And walk me through that UGC. Now that you've got the resources to actually set it up. How have you set that up and created this kind of army of just people posting stuff related to Cluely? What's the setup?

Roy Lee (00:42:28):
I think UGC is really simple. The key that makes UGC work is that. People get super underpaid right now for their ability to make viral content, and right now this delta has not been captured by anyone. So when you hit someone up and say, hey, are you interested in a paid content opportunity, and they're regularly making content. Almost every single one of them will say yes. Because they've never been paid for content before, even if they have like fucking 100,000 followers.

Pablo Srugo (00:42:52):
Yeah, this is true up to what? Because obviously, the power lobby. The one with 20 million followers making real money, but there's basically this long tail you're saying.

Roy Lee (00:42:58):
I'd say 100K, yeah. So if you can find someone like, 20K followers and a couple of their videos been going viral. Just as a, hey, paid content opportunity, you're interested in Smith. You just, all you have to do is make a fresh account and just post Cluely related content. Here's a short brief, and then you sort of just give them creative freedom. You fire them if they do poorly and you just pay them per video with view bonuses. Then they will rip for you.

Pablo Srugo (00:43:18):
That's it? So you don't actually have to do anything. They decide and then they're just judged by the outcome.

Roy Lee (00:43:23):
There's a lot of operational things that we do. Payroll for 60 people where half of them are coming in and out every single week. That's very hard to manage. We go through every video and we have in-house content coaches that will look at the videos. And determine a hit video, yeah. Like, you need to improve, try and talk this way. Try and look at the camera this way. This is more viral, things like that. When we see that a video is working and we track all this. When we see that a video is working, then we will distribute it to all the other creators and say, Hey, you guys all have to make this style of video now. Because it's obviously working. Let's double down on this format, and there's a lot of operational stuff that goes on behind the scenes.

Pablo Srugo (00:43:58):
And then how much of your time? Your brain energy is devoted to just new ideas that are going to go viral versus, you know, running a business that you have now? $5 million ARR and so on.

Roy Lee (00:44:06):
I'd say probably like 20 percent of my headspace is devoted to the whole marketing. Go-to-market side of things. But now at this point, we have an in-house film studio and sort of the only thing I need to do is say, hey, I think we need to do this sort of launch video. This sort of thing will go viral, and then I pass it off to a script writer. And he writes the script, and then he passes it off to a runner. Who will go out and get all the props necessary, make the scene, and he will work with our director of photography. They'll make the shot, and they'll pass it off to our editor. We'll edit the whole thing in-house, and the whole operation takes like maybe two weeks.

Pablo Srugo (00:44:36):
On the product side. The other thing I want to dive into, because this is something we hear about a lot. The 0 to $100 million ARR. How much of this is ARR versus somebody trying it out, paying for a month, churning out? What is? I'm sure churn is high just because you're casting such a wide net and a lot of people probably try it out, and then churn out. But how do you think about retention in your product and what does that look like?

Roy Lee (00:44:55):
Yeah, I mean. We're definitely working on retention a lot more. This is probably the make or break for us right now. Retention is quite good for people who can actually get to the meeting quicker. There's a lot of people who like, Granola who has not captured the entire market of a meeting note taker. That does not join the meeting and just like an all-in desktop meeting note taker. And there's so much left in the market to capture there. So retention, we know that post-call meeting note takers are like. These are good, granola's good. We know this, but the problem is that most people can't actually get to the product. It's quite a novel interface, and I think most people are unaccustomed to this pane of glass that shows up over everything.

Pablo Srugo (00:45:33):
You're saying they're running it. It's recording, but they're not really using it. They're not querying it as much as you'd want them to?

Roy Lee (00:45:38):
Yeah, yeah, yeah.

Pablo Srugo (00:45:39):
Where does the product go from here, by the way? Because that's the other thing. It's so simple, it's so easy to use. It's great, but then it's not clear how you evolve.

Roy Lee (00:45:46):
Yeah, well, I think it's data integrations and actions are the next big thing. Can you imagine that this doesn't just integrate with the web? Imagine it integrates with all your emails, your Google Drive, your Notion, and knows truly everything about you. And it can, contextually rag over everything. Now, when you have a meeting with a person, then it can. Literally, all the relevant information in your life is fed into this one thing. It can just spit out to you the one or two sentences that you need most in that moment, that's crazy. Imagine for enterprise, we integrate with your Salesforce, your HubSpot, your every single CRM, every single database of customer, case study, that you've ever done. Now you're in a call, now you can bring up exactly the most relevant case studies, exactly the most relevant consumers. That's insane value. It's never been done before, and I think the next step after data integration is action. Suppose you can just tell Cluely to do something and it'll be able to click around on your computer, and do things. That would be the biggest jump. And from then on, it literally just replaces your computer. And I think once we get to a point of truly replacing the computer and being a completely new graphical interface for technology. At that point, we have a very strong shot at being the winners of distributing the glasses, the pendants, the fucking brain chips, or whatever the next interface may be.

Pablo Srugo (00:46:56):
You're talking about a UI being language instead of, you know, windows basically.

Roy Lee (00:47:00):
Exactly.

Pablo Srugo (00:47:01):
And do you see the voice piece too? Like, when you go in a meeting and you always have Cluely is a person in there sort of thing and it just says shit like you. I don't know, I just imagine. I don't know, is that a real thing or is that just?

Roy Lee (00:47:11):
I don't think that's where Cluely will go. I feel like standalone AI humans is never going to be a thing.

Pablo Srugo (00:47:17):
Why is that?

Roy Lee (00:47:18):
I feel like every conversation is not driven by the context, bro. I feel like this is a human conversation that we are having, and even if an AI was her. Maybe it'd be able to give you information, but it wouldn't be able to. I feel like there's a very human component to all conversations that AI just misses and will always miss.

Pablo Srugo (00:47:34):
Yeah, I could see that. You know where I'm thinking about it in a conversation, you know, you have these team meetings and like you said. It knows all your salesforce, knows all your shits. You're talking and you're like, yo, how much is, you know what I mean? You need facts, like, oh, how much should we invest? This happens for us, for example. How much should we invest in a startup? And then somebody goes and looks it up. I mean, obviously Cluely could just put it by text. But theoretically, a Cluely agent could just be like, yo, you invested $2 million, you know what I mean? Whatever it is, but is that just like, this is not. You think text is more of the stronger place?

Roy Lee (00:48:01):
Oh, yeah. I mean, I think generally AI should be like augmenting humans. Cluely, hopefully it will augment humans with AI more so than it will like replace them. And I think, yeah, it's like a much stronger product. And I think it benefits the world more for us to build the augmentation component of this.

Pablo Srugo (00:48:20):
What do you worry about? Do you worry about competition? Do you worry about ChatGPT kind of putting something out like this or not really?

Roy Lee (00:48:25):
Not really. I think the other apps, like we move quite fast. And I think right now, the current state of the product is very different from what the product will look like in six months. And I think, sure, we might intersect with some competitors, but I don't think anybody sees the exact same vision that I see with the product. And I don't really see us as zero sum. And I also don't see us as playing in the space that we're playing in is much bigger than I think most people realize. And if you claim, if we're thinking about this as like, we're thinking about the market as being as big as we truly think it is. And right now the current state is just like a drop in the bucket. And we can also argue, hey, Cursor is competing with us. Because Cursor is also trying to do an AI that can agentically take actions for you. What the fuck, you're not competing with Cursor? Yeah, I'm really not worried about competition, I think.

Pablo Srugo (00:49:11):
I mean, ChatGPT being the only. The one just because it's a consumer app and they really do, I think. The question is have they, I mean, for me at least. Are they still on this AGI thing or have they realized that they're mainly a consumer app and they just need to take, you know, I don't know. I don't really know where they're going, and to an extent. It doesn't really matter to worry about them. But that's the only one that I would say, you know, could in theory do something like this.

Roy Lee (00:49:33):
Yeah, seriously and I think sometime down the line. Probably we'll have a data mode of our own and train data based on the user experiences. And at one point we'll just have the best actionables in the world. Because now we just know exactly which ones are best.

Pablo Srugo (00:49:48):
Cool, man. Dude, let's stop it there. Let me ask the last three questions, we always end on. The first one being, it's probably pretty clear in your case. But when did you know you had true product market fit with Cluely?

Roy Lee (00:49:58):
It's actually not clear at all. I think probably like two weeks ago. We launched live meeting notes and that was the first, like, retention tripled pretty much overnight. And I think that's the point where we realized like, okay, we have something that is genuinely working. But with Interview Coder, that day one for interviews. Like, bro, if it works. You have PMF, and at that point. To cheat on interviews, yeah. If it will give you the answer and you won't get caught. Then yeah, that's PMF. It is very obvious, day one we knew we had PMF.

Pablo Srugo (00:50:24):
Live meeting notes, by the way is pretty sick. I used it. I use it pretty often. I mean, sometimes you close a notebook. But just it's crazy how often you miss something, especially in a group conversation and you're like, oh shit, there's. So that was a good one. Second question is, have you ever. Whether with this or last or Interview Coder, like whatever. Any of those two companies, was there ever a point you thought shit was just going to crash and burn, and just not work out?

Roy Lee (00:50:50):
Everyday, like interview coder. Bro, It was so shaky. When we launched the entire database and fucking like I leaked the service to OPE. Anybody could go in and fucking nuke the DB if they wanted. It was crazy how shaky it was, and we were literally putting out the fires as they were coming on. And every single day I woke up thinking, damn, someone fucking hacked us for sure. Someone leaked all the information and how the fuck now. But fortunately, it didn't ever get there. Cluely, I think this is still an ongoing issue. We wake up, there's people claiming like, oh, I hacked Cluely. Man, you did not fucking hack us, and right now. I think I'm in a more comfortable spot where I feel more confident with the product and the vision. And like, oh shit, this actually works. We can actually keep growing this. But previously with Interview Coder, man. Every fucking day I woke up, just show me the cease and desist already.

Pablo Srugo (00:51:36):
You still wake up with your heart like pumping out of your chest?

Roy Lee (00:51:38):
Less so now, but I think now. We actually have some real grounding and a real shot at building something huge. But previously, before you sort of like. I don't think we're at escape velocity right now, but I don't think we are default dead.

Pablo Srugo (00:51:53):
Last one, what's your number one piece of advice for another early stage founder?

Roy Lee (00:51:57):
I'd say take more risks, and the opportunities in short-form content are much bigger than you think. I think in our generation, there are two very big opportunities, and only one of them has been fully realized. One being AI, which obviously every tech entrepreneur is on, and the other being short-form content. The ability to snap your fingers, record a 10-second video, and have that be seen by 100 million people. This has never before been possible in human history, and you have no idea how powerful that can be. That's what I would say.

Pablo Srugo (00:52:22):
Roy Lee, thanks for jumping on the show, man. It's been awesome.

Roy Lee (00:52:25):
Yeah. Thanks so much for having me, brother.

Pablo Srugo (00:52:26):
85% of people who listen to the show, just started listening to it this year. You're probably one of those people. In fact, the odds are 85%. Guess what? You need to go back. Because there are over 100 other episodes that you need to check out. That are just as good, if not better, than this one. Don't miss out.