The Exchanges

Q I mean, let's take it in kind of chronological there. You said one about stealing the processes that work. What did you steal that worked, and how did you translate that to whatnot?

A So I think, like, when I look at a huge part of the success of both Google and Facebook, it was really about empowering engineers and allowing engineering to, like, go out and build and solve real user problems. Google is incredibly user problem focused and incredibly engineering focused. Probably taking that to more of an extreme, if I'm being honest. So today we have about 50 engineers in the team. We have one product manager. And so our engineers drive the company and drive the product roadmap and drive the decision making. You know, we've had to, you know, adjust things a little bit. All of our engineers actually go through a product interview so that they can make sound product decisions and judgments. But when you have the people who are actually executing and building the thing, driving the decisions, they're way more empowered and you have fewer people in the equation so you can get things done way faster.

AI assessment note: “empowering engineers and allowing engineering to, like, go out and build”

Q I mean, let's take it in kind of chronological there. You said one about stealing the processes that work. What did you steal that worked, and how did you translate that to whatnot?

A So I think, like, when I look at a huge part of the success of both Google and Facebook, it was really about empowering engineers and allowing engineering to, like, go out and build and solve real user problems. Google is incredibly user problem focused and incredibly engineering focused. Probably taking that to more of an extreme, if I'm being honest. So today we have about 50 engineers in the team. We have one product manager. And so our engineers drive the company and drive the product roadmap and drive the decision making. You know, we've had to, you know, adjust things a little bit. All of our engineers actually go through a product interview so that they can make sound product decisions and judgments. But when you have the people who are actually executing and building the thing, driving the decisions, they're way more empowered and you have fewer people in the equation so you can get things done way faster.

AI assessment note: “So today we have about 50 engineers in the team. We have one product manager.”

Q have, and 90% of the time also being right, or being good at least. A question that I have that you mentioned earlier, which is really interesting, is you put all engineers through a product interview. That's quite rare. Can you help me understand, what does that interview process look like? What questions do you ask? What are the signals that great engineers produce to suggest they have product nails?

A Yeah, we honestly just give them like a product design interview question, pretty similar to what like a PM would get. It can take a couple of different forms. Probably the most common one is like, walk me through the whatnot app. Here's an example. Tell me how you would redesign the feed to make it better for users and literally come up with product solutions. And then maybe even talk a little bit about the implementation details and the signals you're looking for there. One, it's funny how this one always falls off, but someone who's actually like Use the whatnot app significantly. If someone's going to be able to make sound product decisions, they need to have the initiative to actually use the app, be excited about it, and be into it. You'd be surprised at the number of people who can hardly speak to the product. You know, I think a lot of people come into the interview process like, oh, there's this unicorn company, it's growing really fast, I want to be on the rocket ship. And so they look at it from that perspective, but if you're going to work on something, you have to be really into it. So I think that's number one. You know, number two is just, like, extreme user empathy. So, like, they can articulate some of the problems that users on the platform have, then be able to devise a solution to those problems. Then the third, actually, I think is a comfort taking real own…

AI assessment note: “Tell me how you would redesign the feed to make it better for users”

Q have, and 90% of the time also being right, or being good at least. A question that I have that you mentioned earlier, which is really interesting, is you put all engineers through a product interview. That's quite rare. Can you help me understand, what does that interview process look like? What questions do you ask? What are the signals that great engineers produce to suggest they have product nails?

A Yeah, we honestly just give them like a product design interview question, pretty similar to what like a PM would get. It can take a couple of different forms. Probably the most common one is like, walk me through the whatnot app. Here's an example. Tell me how you would redesign the feed to make it better for users and literally come up with product solutions. And then maybe even talk a little bit about the implementation details and the signals you're looking for there. One, it's funny how this one always falls off, but someone who's actually like Use the whatnot app significantly. If someone's going to be able to make sound product decisions, they need to have the initiative to actually use the app, be excited about it, and be into it. You'd be surprised at the number of people who can hardly speak to the product. You know, I think a lot of people come into the interview process like, oh, there's this unicorn company, it's growing really fast, I want to be on the rocket ship. And so they look at it from that perspective, but if you're going to work on something, you have to be really into it. So I think that's number one. You know, number two is just, like, extreme user empathy. So, like, they can articulate some of the problems that users on the platform have, then be able to devise a solution to those problems. Then the third, actually, I think is a comfort taking real own…

AI assessment note: “Tell me how you would redesign the feed to make it better for users”

Q What did you believe that you've changed your mind on recently?

A Probably one of the things, I don't know if a total mind change, but the thing that I'm having to, like, adjust to is as we, like, scale the company, we really have to, like, empower everyone to solve problems, and I'm that type of person when I see a Problem. I just, like, run at it, and I've had to, like, slowly, like, I see a problem, I let it burn for a little bit, so we can get the team focused in on that problem, and actually solve it on their own. I know that is correct, but I haven't been able to avoid the urge in a lot of instances. That's one where I've probably now taken a more extreme view, where my initial view would have been, like, but it's a really big problem, fix it. Now my view is, like, okay, but if I do that again, the team's never going to solve it, and we're going to have 10 more big problems, and so I guess my view on, like, letting problems simmer so we get the organizational muscle to have people solve those problems, and I would probably give credit to our VP of Engineering, Ludo, for pushing me and changing my opinion on that.

AI assessment note: “my view on, like, letting problems simmer so we get the organizational muscle”

Q I do want to ask you, you mentioned Facebook there on the build fast and unstable infrastructure. When we chatted before, you said most people are not building great consumer products today. It was an interesting statement, so I do want to spend a little bit of time on that. Why do you think that is? What is it that Fang maybe teaches us that's wrong?

A I'll step back, like, and be a little bit explicit about what I mean there. When you look at the broader, like, VC market right now, almost everything's in BtoB. There are very few Few consumer companies that are growing very fast with material businesses and material metrics. I think the kind of meme around it is like the air has been sucked out of consumer a little bit by the fan companies. And I just think that there's probably an element of truth there, but I think there's a lot more to it. Again, relates to this point of not being able to hire PMs, like building great consumer companies is not a mathematical equation. It's like when I go and look around, like all of the materials about how do I build product? How do I think about this? There's so much stuff on retention and cohorts and metrics and this and that, and that is just like not how you build consumer products. It comes from deep empathy of users and ability to devise good solutions. The metrics come after that. Once you build something great that people love, then you can start to look at the metrics, make sure the businesses, you know, you can scale the business, find issues and diagnose issues in the product later on, but the metrics are a second order phenomenon. I think in the past year, there's probably been like Three really interesting consumer companies. It's been Clubhouse, OpenSea, and us. Two out of th…

AI assessment note: “building great consumer companies is not a mathematical equation”

Q What's the hardest thing about being a first-time founder? Moving to you, personally, in leadership, in yourself, what's the hardest thing?

A I think my greatest weakness, probably developing a team and dealing with some of, like, the hard situations that pop up. I think, as I said earlier, of founding team, I'm the softie, and so when we get to things that are, like, really hard decisions, like, hey, this person's underperforming, I'm the one that's gonna, like, Let's give it some time. Let's like work with the person. Then as you scale your company, what starts to happen is you start to bring in a lot of people who actually have much more experience than you at lots and lots of different things. And sometimes even they have performance issues or be stuck on a problem. I find mentoring and up-leveling them to be difficult. And so if I were to like sum it up, it's the, the scale of kind of people management, dealing with people issues and those things. And I actually think it's in part, largely Because I just want everyone to succeed, but it won't happen. And so there's just hard problems and hard decisions and hard conversations you have to have. And I think that's candidly my greatest weakness.

AI assessment note: “developing a team and dealing with some of, like, the hard situations”

Q me up so nicely. I always have these great schedules, which are very structured, and then I get into it, and I'm like, ah, fuck it. I'm just gonna go whatever way I want. You mentioned 50 engineers, you mentioned one p.m. You said to me before the show, it's impossible to hire good p.m.s. Why is that? What's the solution? Is one to 50 okay? Help me understand this.

A I'll share a little secret. Our one p.m., she did just go on maternity leave on Friday, so we're actually officially zero p.m.s. And 50 engineers. We are going to build a PM team. We want to build a very exceptional PM team. We'll probably always keep the ratios very high relative to standard companies because we want to optimize for speed. The more people in the decision-making process, the slower it goes. It's just like no way to like solve that. In terms of hiring PMs, the biggest pool of PMs today are coming from like the fan companies, Facebook, Google, et cetera. And they tend to be Strategy and numbers focused. You know, I strongly adhere to kind of like YC's principle of build something people want, and so it just all flows from users. At this stage, the greatest superpower you can have is like deeply understanding your users and your user problems. You know, I think at scale, it's really hard to do, which is why Facebook and Google rely a lot more on strategy and data, but if you are really grounded and empathetic to your users, you can make high quality decisions Without huge amounts of data, ultimately what that allows you to do is speed everything up. Add value to users exceptionally fast.

AI assessment note: “In terms of hiring PMs, the biggest pool of PMs today are coming from”

Q Final one for the quickfire. Do you ever doubt yourself and have insecurity? You are, you know, as we said, first time founder. For me, many things that I do, I'm like, fuck, I, I don't know if I can do this.

A I have moments of very high anxiety when it's unclear exactly what will happen. I'm a little bit further along in my career. I'm 33, you know, I've been working in tech for like 10, 12 years. I think it'd be really tough to be a young founder who just hasn't been through the ringer a little bit. I've been through the ringer a little bit. I've seen some things. I've had ups and downs. There's fewer things that phase me. Success is breed confidence. You know, we've been really fortunate in, you know, less than two and a half years. We've built a really, really large business. We are one of the biggest marketplaces per period of time. We're the fastest growing marketplace ever. That does give us a high level of confidence. When I'm overly confident or overly underconfident, I always try and like switch the equation around and try and rethink it. Am I being too confident? Do I have it right? What could go wrong? And in some of those cases, you can fall back on advisors, whether it's Matt, you know, other folks that, you know, help.

AI assessment note: “I have moments of very high anxiety when it's unclear exactly what will happen.”

Q Was there anywhere else where it broke? So that was one. Was there anywhere else where it was like, ah, Logan?

A You know, the move fast principle is a double-edged sword, as it was at Facebook as well, which is why they went from move fast and break things to like move fast and build things on stable infrastructure. You know, I think in February of this year, our app went down every day, literally every single day. Everything we do fast, but that we can't substitute quality. Setting that minimum bar, that B plus student, and so that's one thing that I think we didn't push on enough on in the early days, and definitely, one, caused issues in some of the business metrics at various points in time, and also burned the team out quite a bit. Team was dealing with sev after sev after sev, or like app going down after app going down after app going down. There was, you know, like our DevOps team in particular, they weathered an incredible storm this year. It's a testament to how great of a team they are, but we definitely put them a bit through the ringer as a result of that, and we've had to, like, Push much more on quality, both because it's good for our users, but also for the sanity of our team.

AI assessment note: “in February of this year, our app went down every day, literally every single day.”

Q me up so nicely. I always have these great schedules, which are very structured, and then I get into it, and I'm like, ah, fuck it. I'm just gonna go whatever way I want. You mentioned 50 engineers, you mentioned one p.m. You said to me before the show, it's impossible to hire good p.m.s. Why is that? What's the solution? Is one to 50 okay? Help me understand this.

A I'll share a little secret. Our one p.m., she did just go on maternity leave on Friday, so we're actually officially zero p.m.s. And 50 engineers. We are going to build a PM team. We want to build a very exceptional PM team. We'll probably always keep the ratios very high relative to standard companies because we want to optimize for speed. The more people in the decision-making process, the slower it goes. It's just like no way to like solve that. In terms of hiring PMs, the biggest pool of PMs today are coming from like the fan companies, Facebook, Google, et cetera. And they tend to be Strategy and numbers focused. You know, I strongly adhere to kind of like YC's principle of build something people want, and so it just all flows from users. At this stage, the greatest superpower you can have is like deeply understanding your users and your user problems. You know, I think at scale, it's really hard to do, which is why Facebook and Google rely a lot more on strategy and data, but if you are really grounded and empathetic to your users, you can make high quality decisions Without huge amounts of data, ultimately what that allows you to do is speed everything up. Add value to users exceptionally fast.

AI assessment note: “We'll probably always keep the ratios very high relative to standard companies because we want to optimize for speed.”

Q challenging? Because I see this quite a lot with companies that I invest in where common that happens, but they mess it up and then suddenly they've got a load of people that actually they need to move on and move into different roles. And then the culture is kind of further impacted and it's further. How did you deal with that sudden wave that you then have to move?

A Luckily for us, it was never a huge wave of people. I give my co-founder a lot of credit here. He's always had a lot of the way we hired our team was through referrals, both me and him, probably more of his referrals than my referrals, if I'm being honest. He had very good judgment in people, and so despite us not interviewing explicitly on the principles, we kind of intuitively found ourselves bringing on people who are well aligned, and so what ended up happening was that a small number of people who weren't aligned. The impact of that Ended up being smaller. I can also say I hate having to let people go. I like deeply care about everyone in the team. I think it's terrible when it has to happen. We try and do it really humanely. I'm a little bit me versus my co-founder. I'm kind of the softer one. Yeah, I am. You know, we don't like doing it. But what I can say is after every one that we have done, the team's actually improved quite drastically. And it's kind of surprising and kind of jarring. What we've tried to do, like one of our principles is transparency. We just try and be really transparent about how those things work, so it's not a black box to people. Which is to say, we obviously don't comment on an individual person's situation, but we comment on how we handle these things. Like, if someone's underperforming or it's not a fit, we always try and work with them. If a…

AI assessment note: “Luckily for us, it was never a huge wave of people.”

Q I love that. You mentioned that the dog fooding culture, which, which I agree and love. My question is around that. How do you think about when to listen to your users versus when to listen to intuition and gut and build what you and the team think is right?

A You know, you brought up this question of, like, how come hiring PMs is so hard? This is part of the root of it. You need to have acceptance Exceptionally good taste and judgment to make those decisions. There isn't a scientific formula for it, and so I think the best that you can do is deeply understand your users and be incredibly empathetic to them, and also deeply understand the business. Things do trade off against the business. If you have people with sound judgment, you kind of unleash them to make those decisions, and they can make them. If they're really good judgment, most of those decisions will be good. 10% of them might be bad, but we live in a world of two-way doors, so you can walk those back, and you can learn from them. And then over time, you're basically just tuning your judgment. You know, we're pretty lucky because we built a product for us, and I do think that is like a hack to the system.

AI assessment note: “There isn't a scientific formula for it... You need to have exceptionally good taste”

Q Was there anywhere else where it broke? So that was one. Was there anywhere else where it was like, ah, Logan?

A You know, the move fast principle is a double-edged sword, as it was at Facebook as well, which is why they went from move fast and break things to like move fast and build things on stable infrastructure. You know, I think in February of this year, our app went down every day, literally every single day. Everything we do fast, but that we can't substitute quality. Setting that minimum bar, that B plus student, and so that's one thing that I think we didn't push on enough on in the early days, and definitely, one, caused issues in some of the business metrics at various points in time, and also burned the team out quite a bit. Team was dealing with sev after sev after sev, or like app going down after app going down after app going down. There was, you know, like our DevOps team in particular, they weathered an incredible storm this year. It's a testament to how great of a team they are, but we definitely put them a bit through the ringer as a result of that, and we've had to, like, Push much more on quality, both because it's good for our users, but also for the sanity of our team.

AI assessment note: “in February of this year, our app went down every day, literally every single day.”

Q challenging? Because I see this quite a lot with companies that I invest in where common that happens, but they mess it up and then suddenly they've got a load of people that actually they need to move on and move into different roles. And then the culture is kind of further impacted and it's further. How did you deal with that sudden wave that you then have to move?

A Luckily for us, it was never a huge wave of people. I give my co-founder a lot of credit here. He's always had a lot of the way we hired our team was through referrals, both me and him, probably more of his referrals than my referrals, if I'm being honest. He had very good judgment in people, and so despite us not interviewing explicitly on the principles, we kind of intuitively found ourselves bringing on people who are well aligned, and so what ended up happening was that a small number of people who weren't aligned. The impact of that Ended up being smaller. I can also say I hate having to let people go. I like deeply care about everyone in the team. I think it's terrible when it has to happen. We try and do it really humanely. I'm a little bit me versus my co-founder. I'm kind of the softer one. Yeah, I am. You know, we don't like doing it. But what I can say is after every one that we have done, the team's actually improved quite drastically. And it's kind of surprising and kind of jarring. What we've tried to do, like one of our principles is transparency. We just try and be really transparent about how those things work, so it's not a black box to people. Which is to say, we obviously don't comment on an individual person's situation, but we comment on how we handle these things. Like, if someone's underperforming or it's not a fit, we always try and work with them. If a…

AI assessment note: “Luckily for us, it was never a huge wave of people.”

Q I do want to ask you, you mentioned Facebook there on the build fast and unstable infrastructure. When we chatted before, you said most people are not building great consumer products today. It was an interesting statement, so I do want to spend a little bit of time on that. Why do you think that is? What is it that Fang maybe teaches us that's wrong?

A I'll step back, like, and be a little bit explicit about what I mean there. When you look at the broader, like, VC market right now, almost everything's in BtoB. There are very few Few consumer companies that are growing very fast with material businesses and material metrics. I think the kind of meme around it is like the air has been sucked out of consumer a little bit by the fan companies. And I just think that there's probably an element of truth there, but I think there's a lot more to it. Again, relates to this point of not being able to hire PMs, like building great consumer companies is not a mathematical equation. It's like when I go and look around, like all of the materials about how do I build product? How do I think about this? There's so much stuff on retention and cohorts and metrics and this and that, and that is just like not how you build consumer products. It comes from deep empathy of users and ability to devise good solutions. The metrics come after that. Once you build something great that people love, then you can start to look at the metrics, make sure the businesses, you know, you can scale the business, find issues and diagnose issues in the product later on, but the metrics are a second order phenomenon. I think in the past year, there's probably been like Three really interesting consumer companies. It's been Clubhouse, OpenSea, and us. Two out of th…

AI assessment note: “building great consumer companies is not a mathematical equation. It's like when I go”

Q You said there about kind of build things fast, ship it. How do you prevent chasms from happening or being created between engineering, trying to build stable infrastructure, Beautiful code, and then Jeff and the growth team building fast and shipping. How do you prevent that chasm with two kind of different mindsets?

A I think, again, it comes back to what I said earlier. It's we build quality stuff fast, and so there's no excuse to build anything that isn't of quality, with a couple exceptions. So if we were building, like, a new zero-to-one product and completely new business line, we're happy to, like, spaghetti code it, validate that there's, like, users love it. But anything in the core product that growth is working on, it's built to scale. You can build things to scale. In a lot of cases, particularly now given our size, working on quality and making sure it scales, you've got some level of testing in there, actually just enables us to move faster because we're spending less time on problems.

AI assessment note: “It's we build quality stuff fast, and so there's no excuse”

Q Final one for the quickfire. Do you ever doubt yourself and have insecurity? You are, you know, as we said, first time founder. For me, many things that I do, I'm like, fuck, I, I don't know if I can do this.

A I have moments of very high anxiety when it's unclear exactly what will happen. I'm a little bit further along in my career. I'm 33, you know, I've been working in tech for like 10, 12 years. I think it'd be really tough to be a young founder who just hasn't been through the ringer a little bit. I've been through the ringer a little bit. I've seen some things. I've had ups and downs. There's fewer things that phase me. Success is breed confidence. You know, we've been really fortunate in, you know, less than two and a half years. We've built a really, really large business. We are one of the biggest marketplaces per period of time. We're the fastest growing marketplace ever. That does give us a high level of confidence. When I'm overly confident or overly underconfident, I always try and like switch the equation around and try and rethink it. Am I being too confident? Do I have it right? What could go wrong? And in some of those cases, you can fall back on advisors, whether it's Matt, you know, other folks that, you know, help.

AI assessment note: “I have moments of very high anxiety when it's unclear exactly what will happen.”

Q How do you create a culture whereby engineers and team members feel that ownership and are excited by it to go big and to think ambitiously, but also aren't scared of ramifications if it goes wrong?

A Yeah, it's, you know, it's a really Tough one. I don't think we've nailed it entirely. One of the, the more interesting observations that I've had in building this company is the level of risk aversion. When you're coming from, let's just say a big tech company to whatnot, you kind of have to like, one of the biggest things that I see people having to do is they have to like shift their whole judgment model a little bit. When you're building a product at Facebook, for instance, you write up the product spec and there might even be like a section there about like key risks. And you're going to go have to, like, de-risk a bunch of those things. You probably spend the vast majority of your time actually, like, de-risking things. Whereas, like, the optimal way to do things, particularly at our stage, is to, like, invert that. What is the thing with the highest upside? As long as there are no existential risks, don't worry about them. Most of the risks that you enumerate are never real, and so instead you kind of have to just push something out there and understand what the real risks are. My co-founder's a little bit better at this than me. I'm a little bit, like, more risk averse, but it's amazing the number of things that we've pushed out and we thought would be higher risk. Something else would pop up and it just never does. I think there's a couple of layers to how you solve th…

AI assessment note: “I think there's a couple of layers to how you solve the risk taking.”

Q Grant, the best CEOs are the best resource allocators is a statement often said. You said they're about making bets. How do you think about the bets that you want to take versus the ones that you don't?

A I think that's probably true for a lot of founders. Like, we like to build new things. That's the joy. The joy is in building new things, which is in the new bets and bringing something from zero to one. I have a document with all the bets that I want to make at whatnot, and it's very, very, very, very long. Because you can very easily go in one direction or the other. I kind of think about 10 to 20% of resource should be on new and exciting things. I actually believe one of the bigger failure points for startups, they do not take enough. Again, it comes back to say they did not actually take enough. Like, you can basically fall on both sides of the spectrum, which are failure points. You can either take too much risk or not enough risk. If you look at, like, the really, really big and great technology companies, Amazon, Google, Facebook, Apple, it Etc. One of the common denominators is they continue to reinvent themselves and build new and exciting products. And the later you do that in the life cycle of your company, the greater the uphill battle and the inertia and the culture somewhere else. You know, as I said earlier, it's very hard to like get people to want to take risks. If we didn't have that as a core part of our culture from day one, and we got to a thousand people, it would be infinitely harder. The flip is you take too many, you can't focus, which means you can't …

AI assessment note: “no more than 20% of resource on new and exciting bets.”

Q Grant, the best CEOs are the best resource allocators is a statement often said. You said they're about making bets. How do you think about the bets that you want to take versus the ones that you don't?

A I think that's probably true for a lot of founders. Like, we like to build new things. That's the joy. The joy is in building new things, which is in the new bets and bringing something from zero to one. I have a document with all the bets that I want to make at whatnot, and it's very, very, very, very long. Because you can very easily go in one direction or the other. I kind of think about 10 to 20% of resource should be on new and exciting things. I actually believe one of the bigger failure points for startups, they do not take enough. Again, it comes back to say they did not actually take enough. Like, you can basically fall on both sides of the spectrum, which are failure points. You can either take too much risk or not enough risk. If you look at, like, the really, really big and great technology companies, Amazon, Google, Facebook, Apple, it Etc. One of the common denominators is they continue to reinvent themselves and build new and exciting products. And the later you do that in the life cycle of your company, the greater the uphill battle and the inertia and the culture somewhere else. You know, as I said earlier, it's very hard to like get people to want to take risks. If we didn't have that as a core part of our culture from day one, and we got to a thousand people, it would be infinitely harder. The flip is you take too many, you can't focus, which means you can't …

AI assessment note: “I kind of think about 10 to 20% of resource should be on new”

Q Hey, do you have to ask, what was the more challenging one, the buy side or the sell side?

A There's probably not a huge difference. If I'm being honest, I think it probably depends on the business. Certainly for our business, our business is very, very seller oriented. And so we do spend a lot more of our time on the supply side. When we do switch over to the buy side, we are able to do pretty amazing things that like move the needle. You know, we shipped the feature a couple months ago. I thought it was cool. It was kind of fun. We shipped it and then it had like, it increased our viewership on every stream by like, I was kind of shocked, but that is part of the reason why like, If you have a group of smart people, let them build stuff, let them do it fast. As long as they're oriented on user problems, they'll surprise you.

AI assessment note: “There's probably not a huge difference. If I'm being honest, I think it probably depends”

Q I love that. You mentioned that the dog fooding culture, which, which I agree and love. My question is around that. How do you think about when to listen to your users versus when to listen to intuition and gut and build what you and the team think is right?

A You know, you brought up this question of, like, how come hiring PMs is so hard? This is part of the root of it. You need to have acceptance Exceptionally good taste and judgment to make those decisions. There isn't a scientific formula for it, and so I think the best that you can do is deeply understand your users and be incredibly empathetic to them, and also deeply understand the business. Things do trade off against the business. If you have people with sound judgment, you kind of unleash them to make those decisions, and they can make them. If they're really good judgment, most of those decisions will be good. 10% of them might be bad, but we live in a world of two-way doors, so you can walk those back, and you can learn from them. And then over time, you're basically just tuning your judgment. You know, we're pretty lucky because we built a product for us, and I do think that is like a hack to the system.

AI assessment note: “There isn't a scientific formula for it, and so I think the best”

Q Hey, do you have to ask, what was the more challenging one, the buy side or the sell side?

A There's probably not a huge difference. If I'm being honest, I think it probably depends on the business. Certainly for our business, our business is very, very seller oriented. And so we do spend a lot more of our time on the supply side. When we do switch over to the buy side, we are able to do pretty amazing things that like move the needle. You know, we shipped the feature a couple months ago. I thought it was cool. It was kind of fun. We shipped it and then it had like, it increased our viewership on every stream by like, I was kind of shocked, but that is part of the reason why like, If you have a group of smart people, let them build stuff, let them do it fast. As long as they're oriented on user problems, they'll surprise you.

AI assessment note: “There's probably not a huge difference. If I'm being honest, I think it probably”