The Exchanges

Q Okay. Yeah. Super cool. So let, let's jump into this really quick. Tell us what the application does. This was the first one quiz or what's it do?

A Yep. If you're familiar with photo math, it's kind of a similar user experience where a student can take a picture of something that they're stuck on, like a concept or a question that they're working through or a practice question they're working through and instantly get thrown into an AI tutoring session where they go back and forth with the AI. They're given step-by-step guidance and explanation. They're even given like pointers to resources on the web, um, so that they can like further, further their learning. Um, and the idea that was super like innovative here from back in the day was that, uh, we basically like I was like, mimic this photo math, like, 24 seven tutoring experience, um, and took it to every subject, uh, possible, every sort of topic possible. Um, and yeah, at a high level, this is pretty much, uh, what Quizzard, um, is going for.

AI assessment note: “a student can take a picture of something that they're stuck on”

Q Wild. Okay. Let's jump into the backstory here because you originally weren't going to do this. I mean, Palantir was pretty heavily recruiting you and your team had already sort of launched a V one, I believe in January, 20, 23. Tell us the backstory. When did you join?

A I joined in kind of, ah, but after launch, the first product had gone out, it had gone with, ah, a TikTok video, got a million views, um, overnight. It turned to 10,000 users. Um, and there was almost an instance of the need to like continue scaling from there. Uh, one of my closest friends, actually the first person I met on campus, Mike, who's, uh, like my co-founder and the CEO, um, Michael. He was the first guy. Yeah, Michael. He was the first person I met on campus. Um, He had seen me do a lot of work on, like, scaling AI stuff. Specifically, I'd worked at a hedge fund for about a year doing a lot of AI research. Um, but also I'd done some random stints of, like, I guess, like, I, I won a, uh, Stanford Tree Hacks price, uh, using, like, LLM's controlled drones, and so he's like, we need you to help us, like, scale our AI stuff. Um, and I, I'm not gonna lie, I was, like, earlier than I thought it'd be in terms of starting a company. I definitely was on this, like, I'm gonna go to Palantir to figure out, like, what I build next kind of route. Um, but when you get the opportunity to build with your friends, I think it's like an easier yes than ever.

AI assessment note: “I joined in kind of, ah, but after launch, the first product had gone out”

Q Why not? You guys are, a lot of your growth is because you're building in public. I mean, you're recruiting people. The reason you get such high responses on your LinkedIn posts is because people know you're at six million bucks in revenue with a five-person team. Why not lean into that?

A I think for us, like, we've seen, we, like, when we launch into a new space or we launch a lot of our strategies, like, a lot of it is, like, I guess for lack of a better word, we're pretty much in a competitive market. You know, there's a bunch of other folks also that build like apps as well, or they're launching a bunch of domains that we are going to be entering that, you know, people are building entire teams around that single app. Um, our concern, and we've seen this before is like, um, when we launched, for example, marketing strategies that do really, really well. So for example, like in the fall of 20, 23, uh, we, we ran this like man in the street campaign where if you search like Harvard, Boston, Columbia, NYU on Tik Tok, Uh, we were the first three videos, if not the first five. Um, and so, you know, like the moment that took off and did really, really well, then we started seeing people like, um, kind of try and milk that alpha after, and it sort of like fizzled out and lost its alpha. So I think like, in some sense, I'd liken ourselves to like a trading company where there's a lot of alpha. It needs to be milked ASAP. There's an arbitrage company to jump into. And the moment that signal is known, it's sort of like loses its value. So I'd say that like, we are public with every app that we are public about. I guess what I mean by that is like, Um, our plan is to a…

AI assessment note: “the moment that signal is known, it's sort of like loses its value”

Q This was the one it just, it just took, oh, I mean, did you do anything or was it just total blind luck?

A Um, there was a bit of, like, for context, like, um, my co-founders had spent a lot of time in, like, the consumer space. Like, um, I'd say, like, so Ashraf had previously run a, like, you know, like, sort of, like, meme page that was pretty large across, like, um, North Africa and, like, France. Um, Mike had done a lot of consumer stuff in the past, and, like, Ben had also done a lot of work in marketing. He'd actually run a, like, Discord community around trading for, for a while that was putting in a lot of revenue. Um, I think To some extent, it was like, we were like one of the first in the market. If you can imagine, like, Chatsipity was launched the end of November. This was launched in January, like, a few months after. Not many people had heard about AI. Um, so a lot of this stuff was very, like, novel. So I'd say we got in at an early time, but there was definitely a lot of thought in terms of, like, how can we simply catch the, you know, use, how we can create a very, like, clear hook in terms of this video. Um, and, and that's how we went with this, like, point of view, like, Chatsipity and Photomath had a baby, which clearly, like, exemplifies, like, Exactly what this tool is kind of coming with. And we go right into a product demo. Um, so people understand.

AI assessment note: “there was definitely a lot of thought in terms of, like, how can we”

Q hit. I'm going to, I'm going to go and install this. Right. So you can actually take us through the onboarding because this looks so simple, but I know anything like that works this well, took a lot of freaking testing. So what was yours as original pricing model as we watched through the onboarding experience here? Cause it wasn't, it wasn't requiring subscriptions. It was coins or credits, right?

A On Quizzard, it was coins for a very short period of time. I actually like the first, like, um, iteration of like the business model for Quizzard was everybody and all usage was free. Um, and like, there's actually an interesting way in which we were able to support that. Um, because, uh, but I'll get to end the study, but basically everything was free on Quizzard. Um, but you know, we put everyone's question in their queue and during peak times, you could skip the queue if you paid, um, for like, you use it, you know, you use a token to skip the queue. Um, and so during peak time is when you're Monetize then. Um, but we realized that that wasn't necessarily like the best unlock for us. Um, Mike actually like texted Nikita, I think on Twitter and, um, Nikita was like, uh, switch to weekly subscriptions and charge more or something like that. Um, who's Nikita?

AI assessment note: “the first, like, iteration of like the business model for Quizzard was everybody and all usage was free”

Q that's just really interesting testing here. Again, I think, you know, Sid, it's really interesting. I'm kind of on the lookout for founding teams that have built a process to run The most tests at the same time. Cause I think that's, who's going to win in this space. Whoever runs the most tests, the fastest wins. How many tests are you guys running right now across all your products?

A There's probably like, like six to seven on one and maybe like three or four or five on another. I think like for us, literally like, I'm so glad you touched on that because like, that's like a huge part of like the space entirely. Like every time we build a new product, like we build in our experimentation, um, like infrastructure, like as the first piece that goes into the software. In fact, like, So the way we operate now is we have a lot of, like, we have, like, the way we structure the company in the way that we think would scale is we have, like, three different arms right now. Um, we have our growth and marketing arm, we have our, like, product engineering arm, and then we have a platform arm. Um, growth and marketing is kind of, like, clear in the name and what it does. The product engineers are effectively hiring engineers who are effectively, like, CEOs of the product, i.e. they're living in the metrics, and their day-to-day is to live and own and, like, breathe their apps and, like, make them successful. Um, And so this is run by my co-founder Ashraf. Um, and then the third like org that I recently started is called a platform org. Now the interesting thing about this org is a little more detached from a day to day of like, of a single product. It does include the, you know, the traditional pieces of like, um, you know, general platformization of any tech company whe…

AI assessment note: “There's probably like, like six to seven on one and maybe like three or four”