The Exchanges

Q How does that happen? Because that's not, yeah, not normal.

A Yeah, maybe the youngest tenure-track law prof in the history of the law school. I think a whole bunch of things went really my way on that one. I, I had an interesting Research agenda. I had published some articles as a law student and I was super keen on the academic path. I was really excited about the life of the mind and doing research and going deep on tax law. And so that's what I spent the ensuing decade really focused on. So from 2004 right through to when I started Blue Jay, I was full tilt into becoming the best possible tax law academic that I could be. And so I wrote dozens of Academic articles. I coauthored several editions of the leading tax book used to teach tax in Canadian law schools of the books called Canadian income tax law. And I coauthored with the lead author, David Duff, the second, third, fourth, fifth, and sixth editions of that book.

AI assessment note: “I had an interesting Research agenda. I had published some articles as a law student”

Q So walking back to that storyline, like you see this just before Chai TPT kind of this new model, you see it could lead to that vision. How do you approach it? Like how all in do you go on that? Cause you have a existing product with existing revenue. So how do you balance that?

A Yeah. It took some courage and some conviction. The courage and conviction was the conviction was doing what we're doing is not going to scale properly. We're just not going to get that breakout product market fit that we need in order for all of this to really work out. Partly it was that it was partly necessity. Like we're going to make this work. We have to get to some solution that can answer any tax research question people want to ask. The other one was, well, do we just forego all of this revenue? Cause Your point is a good one. Like we don't just want to tell our existing customers who are paying us good money for X and the models worked like they were providing value to the customers, but here's kind of how we navigated it. It was okay. We're going to put all of our existing tax research tools into kind of maintenance mode and we'll keep servicing the software. We'll keep updating it, but we're not going to invest in new feature development. We're not going to invest in building new models. People are going to get what's in there and we're going to take six months. I told the company we're going to take The first six months of 2023, we're going to focus all of our new development efforts in building this thing that can answer any tax research question in US federal income tax law. Why US federal income tax law? Well, the market's the biggest on earth. There are a lot o…

AI assessment note: “We're going to put all of our existing tax research tools into kind of maintenance mode”

Q There's three questions we typically end on. One of them is for you personally, like, when did you feel like you had true product market fit?

A I was doing a demonstration for a group at the Canada Revenue Agency. I think it was early, and we had just built the Canadian product and I was invited by somebody at the CRA to come into their Toronto tax services office and do a bit of a talk on AI and tax. And I did a demo. You know, I ran through a couple examples of questions, but then I turned it over to the audience and said, I put them on the spot and said, Does anyone have any questions, any tax research questions that they have? And a guy put up his hand and he said, I have one for you. He was sitting right at the front, um, at one of these round tables. And it was probably a group of like a hundred and fifty-ish people at the CRA, so a pretty big group. And he said, this is a tricky problem. It took us two weeks to, to do this internally. I don't have my hopes up, but can you try this? And he described the question and I push enter. I'm like, okay, like, let's see what Bluejay comes up with. And Bluejay started Bringing in the, the answer. This guy stood up and he like walked up to the screen and he was like reading it line by line. That's the answer we came up with. It was a two week research task and we had like a bunch of experts. And I remember driving back to my office at the law school afterwards. I like, I turned up the radio and I was just like pretty pumped because I was like, okay, like that was hugely suc…

AI assessment note: “I was doing a demonstration for a group at the Canada Revenue Agency.”

Q But this was just delivery or this was actually servicing cars?

A Servicing. Moving and servicing. They would call up other facilities and say, Hey, do you need an extra car today? There's a wholesale rates within the mechanic world. And so they would get a wholesale rate because they were a shop. And then they started to figure out it was also a capacity model on the shop side. So like a lot of mechanics would be like full up, but if they had brokered, 20 deals with the shops in the area, any one of them would want them a car that day. And then they would just charge a consumer rate. So When we first started, our consumers were paying the going rate, the average rate in in geographic region. And then we were brokering wholesale deals, which was anywhere between like a 50 and 30% discounts to shops already in the area. And they would take it because instead of having a mechanic sitting around with his thumb up his backside, you know, eating donuts, watching pornography on his phone, he could do some work and make some money. And the, the shop would not lose money, but they wouldn't necessarily make money on the transaction. So it was more of a capacity model on the shop side.

AI assessment note: “Servicing. Moving and servicing. They would call up other facilities”

Q Talk to me about the B to B transition. Like, how did that happen, and what did that do to the business?

A So, we're doing consumers. It's cruising along fine enough. There's a lot of issues in that statement. And then DoorDash contacts us, and they're like, hey, will you repair our driver's cars in Phoenix and LA, just as a test case. And so we're running around the office, high-fiving each other, like, you know, cranking out, like, more beers and Red Bulls, thinking DoorDash, like, all of a sudden we're gonna have access to, like, thousands and thousands of cars without having to go out and, like, Find each individual one. So we get really excited. We launch in LA and Phoenix, and then we quickly figure out that DoorDash drivers are just not our ideal customer profile. One, they're frugal, and pretty much most of them are driving cars that are, like, a little more beat up, and so they don't really want to put money into them, and then it's also the way they make their money. So they don't want that thing to be, like, out of service for even 12 hours.

AI assessment note: “DoorDash contacts us, and they're like, hey, will you repair our driver's cars”

Q talked about problems like that, that really kind of surfaced up? I mean, because you have the macro problem that you see, KKR, Blackstone, etc. But they, I don't know that as small, most SMBs see that until maybe it's too late, you know what I mean? But what, were there other conversations that actually surfaced like the day-to-day issues some of these people were facing? That they knew about?

A Oh yeah. I mean, uh, there was like one business, uh, they're using their personal card to fund their SMB. Look, like you, he had a per, he had a personal Discover credit card. He would hand his personal Discover credit card to his employees to go to, go to Best Buy, go to this thing, go to this thing. And he would sharing his personal credit. Here's the crazy thing. With the personal credit cards, um, utilization affects your credit score. So this dude was putting 30,000 dollars worth of purchases, and his credit score started nosediving. So his mortgage payments started rising, and then the loans for his businesses all hiked in terms of pricing. And so, look, we were, like, underwriting this guy's business when he told us this entire backstory, because we were his first S&P card, and we're like, dude, like, why did you not apply for a business card? And like, He was like, you know, man, like I didn't really understand the difference between a personal and a business card. And we were like, dude, you've been running an HVAC business for three to five years. Like, look, what do you mean? Like, you don't know what a personal versus business card could look for some people, they see it as points in cashback. They don't see it as anything else. And I think that's largely what the SMB economy looks at cards for. And like, that's what makes this compelling is what We're trying to sh…

AI assessment note: “there was like one business, uh, they're using their personal card to fund their SMB.”

Q And then, but you saw that, but what, are you just building breaks for, for SMBs? Like what, what is it that, what's the, what's the idea?

A Yeah, exactly. Like, I think, like, the main alpha that we're trying to uncover is, like, we're not actually trying to build for finance teams, and that's what Ramp and Brex do, right? Their whole focus is we want to build for CFOs, but, like, SMBs don't have CFOs. What we're trying to do is we're trying to be their finance team, and, and, and what does that mean? And it's sort of why, like, this AI agent play, it's, it's, like, what we're really focused on helping come to fruition is, like, I think like in organizations, there's two types of people. There's catalyzers and there's maintainers. Catalyzers would be like an engineer at a VC backed startup. They're building that new products and they're catalyzing the growth of your company. Then you have maintainers. Maintainers are, finance people are going to hate me, but it's CFOs. They tell you about your business. They're not going to two to three extra valuation tomorrow, but they're going to tell you, here's your cash flow. Here's your monthly statements. This is your accounting. Look, this is what your balance sheet looks like. We think maintainers are the ones that are going to get disrupted in this like AI revolution. And What better way to disrupt and maintainer than to be the maintainer for SMBs? At the end, most SMBs in America will never hire full-time finance functions. They, they're just not big enough to afford a …

AI assessment note: “What we're trying to do is we're trying to be their finance team”

Q And you charge what, like, per, I mean, there's interchange fees and stuff like that, but you also charge, what, per seat or per what, per card?

A So we, we, like, uh, charge, um, right now it's, uh, It's, it's technically custom. So like certain larger customers, we charge them like a really large fee for expense management, but like smaller customers, like small businesses, we charge about 60 to a 120 dollars for expense management. But honestly, that's like a nominal, uh, yeah, per year. Uh, so it's like recurring software. That is not like a ton of our revenue. Look, I think like software is about 1510, 15% of our revenue. Majority of revenue comes from interchange today. Um, I think, like, our philosophy is, like, uh, why charge a customer today when I can give them this platform? I could replace 10 things and then I could charge them more tomorrow. Um, I would rather take that route than, like, try to figure out the profitability game, like, tomorrow. Um, because, like, I would be super naive to sit here and say I could justify a 5000 dollar platform fee today. I can't do that. Um, and I think that's why I kind of admire Ramp, actually, is because Ramp realized this. Ramp would like, I'm not going to be able to justify a charge to a startup founder in 2020 when all we do is receipts and cards. Look at Ramp today. Travel, procurement, treasury, cards, um, like literally eight, 10 different things. Now what's Ramp? Ramp, it's charging now. For seats. Um, and I think they look at a hundred million dollars of software r…

AI assessment note: “It's, it's technically custom... small businesses, we charge about 60 to a 120 dollars”

Q When did you like, when did you learn to code? Like during, through that or were you already doing it before?

A Uh, that was the inspiration that got me, that got me started. And so that was when I saw that I was like, I want to, I want to teach myself how to code. I went to my school. I asked a couple of, uh, my, my seniors are like, Hey, how do I do that? How do you build apps? And they said, Oh, you're too young. You wouldn't get it. I mean, yes. Okay. You see a 10 year old scrawny kid asking you hard questions. You're going to be like, nah, man, but nobody says that to me. And so that's when I pulled actually my first all night, right? I remember vividly, I was 10 years old. I was like, I'm going to teach myself this thing. And so that was when YouTube had started to become a thing. So I went there, I started watching tutorials, started building things, and that was just the start of everything else.

AI assessment note: “that was the inspiration that got me started... I was 10 years old.”

Q And this was back, this is in New Delhi as well?

A This is, this is all, this is all in Delhi. And that was my, my first experience into just building a product. And I learned a lot more from him because at that time, right, like I've been this guy who spent most of his life behind a computer screen. I had written more code than I had spoken words in my life. And so, what I got from there was, he would be having calls with his LPs, and he'd just be like, Tane, just like, sit here in the background, just shadow me. I want you to learn how to do these things. And so, at that time, he's raising money from, like, Matrix Capital, Sequoia Capital, like, the top VC funds in the world, and I, as a fifteen-year-old, had the opportunity to just sit and listen him, listen to him.

AI assessment note: “This is, this is all, this is all in Delhi.”

Q And this was back, this is in New Delhi as well?

A This is, this is all, this is all in Delhi. And that was my, my first experience into just building a product. And I learned a lot more from him because at that time, right, like I've been this guy who spent most of his life behind a computer screen. I had written more code than I had spoken words in my life. And so, what I got from there was, he would be having calls with his LPs, and he'd just be like, Tane, just like, sit here in the background, just shadow me. I want you to learn how to do these things. And so, at that time, he's raising money from, like, Matrix Capital, Sequoia Capital, like, the top VC funds in the world, and I, as a fifteen-year-old, had the opportunity to just sit and listen him, listen to him.

AI assessment note: “This is, this is all, this is all in Delhi.”

Q it firsthand of hitting a million, hitting five. You can, you can stall out at any number at five, at 10, at 50. You're never done. Uh, so anyways, man, welcome to the show. And, um, yeah, maybe, maybe as a first question, like, tell me a bit about Uh, the background of Freshline, like what year are we talking in and kind of what was happening at the time?

A Yeah. So this was back in 2016. Uh, I was studying computer science and business, uh, in university. I was studying in Vancouver and, uh, uh, the Genesis really of the story and, and kind of, uh, the company comes from my co-founder. So, uh, his family had been in and out of the fishing industry, uh, in Northern New Brunswick for a couple generations. So super niche. And, uh, Him and I were, were, you know, we had started a few projects together back in high school, started a company together. Um, so we were very entrepreneurial and we had always kind of exchanged, uh, ideas and kind of thrown kind of different concepts at each other over the summers. And, uh, yeah, we were just kind of discussing his experience, um, in the summers in Northern New Brunswick, where, um, he observed a lot of these fishermen, lobster fishermen going dock to dock to dock. Uh, in each small village before going into the main city to kind of, uh, sell the rest of their catch. So what they were doing is they were going to dock to dock, selling bits and pieces of their catch direct to consumers, uh, before offloading the majority of their catch, um, in the main city. So, um, he being the curious guy that he is, my co-founder, he kind of went down to the fishermen and asked, Hey, why are you going through all this trouble to, uh, you know, Sell bits and pieces of your catch. And I think through that pro…

AI assessment note: “Yeah. So this was back in 2016. Uh, I was studying computer science”

Q read about zero to a hundred million, you know, all the time, and everybody's trying to outdo the last one that did it as fast as whatever. Um, but those are not, you know, even of the success stories, those are the exceptions. Um, so anyways, we'll go through all of that, but maybe just to start, like, what is, what is Lumio? Like, what does, what does Lumio do?

A What we did is we built a graph And that term, when we started talking about how we were going to build Illumio, you know, about 10 years ago, that term was not one that was often used. Now, literally everybody from Illumio all the way to Microsoft talks about the future of security is really being based on a graph. And a graph is sort of a very fancy and unusual way of describing a database that allows you to collect a lot of information, to understand not just the things in the graph, but the relationship that they have to one another. And then what we did was we built two things on top of that graph. Now more and more powered by AI. The first and what we're known for is obviously delivering zero trust segmentation. And that's what the company's been built on. We'll talk a little bit about what that means, but that's one of the two outcomes of having this very powerful security graph. The second one, and much more recently is we deliver now an analytics layer that we call insights that allow you to go inside Ask questions of the graph, understand things that may be happening in your environment that from a security perspective, either are risky or potentially risky. And so the way that we think about a Lumio is we think about it being built on top of this very powerful security graph, and then allowing our customers to do two things, look inside and see what's happening in th…

AI assessment note: “built on top of this very powerful security graph, and then allowing our customers to do two things”

Q And walk me through, like, what were you doing before Illumio that you even noticed all these things happening?

A So both my co-founder and I came out of the firewall space. Um, in my case, it was all network security. So firewall, IDS, IPS, a whole string of acronyms that basically mean looking at traffic, understanding how it moves and when it should and shouldn't be allowed to move. My co-founder was in a very large firewall company here in Silicon Valley and was an architect. So that means that he was writing code and architecting systems. I think the thing though, that's so much more important than what we were actually doing in our day job was that when we got together, he brought the deep technical and architectural point of view. And I've spent my entire career, including Illumio more at Illumio than ever before in customer facing roles. So sales and sales engineering and talking about how products are implemented and what competitive landscapes look like. So He was able to bring the technical problem and obviously a potential solution to the table. And I was sort of, we use a term around here at Illumio called the voice of the customer. I was essentially the founding voice of the customer. I, I was describing the problems I heard customers talking about when they tried to take their firewall and plug it in in public cloud, and it just didn't make sense or work. And that was the genesis of the conversation that became Illumio.

AI assessment note: “both my co-founder and I came out of the firewall space.”

Q macroeconomic problems on top. So it makes it doubly hard. Maybe take me back to like, you know, right before 99, how did you meet your co-founder and, and where did kind of the idea come from? I mean, Amazon had started, obviously eBay was around, like, did you guys look at those models and just bring them to Latin America or what was, what was some of the thinking?

A Yeah, so that's a great, uh, question. We went to Stanford for our MBAs. Uh, and that's where I met Marcos Galperin. He was a classmate, and we became friends together with another Argentinian, Martín de los Santos, that coincidentally today is Mercado Libre's CFO. We had lots of classmates in, in the business, and, uh, we went there during the boom days of the internet. Netscape had transformed the internet recently. Into something that was more for, for deep engineers into a consumer product. So everyone had access to, to the web, thanks to, to Netscape. Uh, so lots of applications started to, to be created on top of it. And as you well said, one of those was eBay and another one was Amazon. Uh, and as graduation was approaching in 99, I was working on a technology business difference to MercadoLibre, and Marcos was thinking in the idea of, of building. Initially, it was kind of the eBay of Latin America. He looked into different business models, and at that time, the three main businesses were, one, the Yahoo model, so search engine, one point oh, The second one was Amazon, and the third one was eBay. Those were the kind of most successful businesses. There were other, like, travel sites, and if you think about those three were the, the major ones, uh, and I think that Marcos, rightly so, uh, realized that maybe we didn't have in Latin America the technology that was needed …

AI assessment note: “We went to Stanford for our MBAs. Uh, and that's where I met Marcos Galperin.”

Q engineer, which is like the peak of engineering at Google. You leave, you start Rubric, which is now a twenty billion dollar public company, and three years in, you guys are a unicorn. Five years in, you're worth like Three billion dollars. You raised a couple hundred million, and then at the absolute peak where everything that you wanted to happen was happening, you leave. Why would you do that?

A Well, so that was 2019, and I left Rubrik to start Glean. And, and in fact, like, you know, at Rubrik, you know, we were very fortunate. We were growing fast as a company, and with that growth, we also realized that we were not productive anymore. As you know, there was a year when we went from 500 to 15, 1500 employees in one year, and that's three times as many employees. But when we looked at our lines of code, it was actually flat. Like we couldn't produce more with all the, all the, all the new people. And it turned out that like, you know, in, in a company, like as, as you get larger, the communication overhead and people knowing like how to do work, how to figure things out, who are the experts that can go and get help from all of these things become difficult. And we had this one big problem. Like everybody in the company was complaining about like not being able to find any information inside our company. And, and I, I'm a search engineer. I just said, you know, by training before Rubric. I was at Google for many years. And when people said that they can't find things, I said, yeah, like, you know, I also can't find anything because we had 300 different SAS systems. Our knowledge is spread all over the place. So we should go and buy a product, connect all of those things together. As we tried to buy that product, we realized there's nothing to buy. And, and that's when…

AI assessment note: “realized there's nothing to buy... that's the reason why I left.”

Q And I guess there's also a point of, like, the caliber of the engineers, I assume, that, that were using this?

A Right. I think that was, that was the big thing, was that it wasn't just, you know, it wasn't just, like, hobbyists. It wasn't, it wasn't like, ah, like random signups. It was, like, engineers of places like, you know, Netflix, Snowflake, ah, You know, ramp, like really good, you know, good companies that were, that were signing up for this. And, uh, we, we also like, we, we implemented it such that, um, if you got, uh, we did kind of a viral waitlist, waitlist mechanism, where if you, if you convinced like, you know, four teammates to sign up as well, then we'd bump you to the top of the waitlist and, um, and onboard you sooner. So, um, that really, that really helped us a lot to, to get, uh, to like encourage teams to sign up together versus, uh, versus just individuals.

AI assessment note: “engineers of places like, you know, Netflix, Snowflake, ah, You know, ramp”

Q Did you ever feel like, especially going through like those pivot hell periods, Like, maybe you, you wouldn't make it to the other side, and this is just gonna, like, not work out?

A Oh, yeah. I mean, there was, I think that, uh, you know, we've, we've now survived, uh, I mean, a series, both, both, like, internally and, and macro events, we've survived a series of, of, like, really unprecedented times, and, um, you know, we weren't sure, from the very beginning, like, we weren't sure if the seed round was gonna come together, if everything would shut down with COVID, uh, We, you know, we thought we were gonna, we thought we were likely to run out of money, uh, with, before the, the waitlist launch, and even for Graphite, and even, even then, I think we, uh, the, the night that we launched the waitlist was, uh, I'm not sure if you remember, like, the, the big, uh, like, AWS, US West II outage, um, in twenty-twenty-two. That, that was, like, the, when it was, like, down for hours and hours, like, That was the night when we were launching our waitlist, so we were, we were, like, sitting, like, scrambling, trying to figure out, like, you know, can we, can we move to a different region? Like, we never really thought, that was, like, the one thing that we hadn't, we'd gone through this whole checklist of, like, what's everything that can go wrong, and, like, you know, entire AWS region failure was just not something that was on the list. Uh, we also, uh, we had all of our money in Silicon Valley Bank when that, uh, when that collapsed, so, uh, yeah, we've had a …

AI assessment note: “Oh, yeah. I mean, there was, I think that, uh, you know, we've, we've now survived”

Q is now. Like today you have an idea you raised like 3,000,003 million dollars or whatever. Uh, it wasn't really the case back then, especially in, in, in Canada, but like, how much did you think in those, that early part of the seven years, like about, you know, should you raise a seed? Should you not? Like, how did you, yeah. How did you walk through that, that thinking?

A I mean, honestly, I, I, um, I'm probably, uh, an edge case here. I just didn't think about it at all. Cause I have no finance background and we were, we were Selling software, making money, and actually I was buying real estate for the, other than the first couple of buildings where we were, there were rentals, I was buying, I was buying real estate with our, with our profits and, uh, and just building it like a real, I felt like if anything happened with a software business, I wanted some assets for the company, you know, so I would buy the company, buy the, buy these buildings and, and, uh, Uh, and, uh, rented to light speed and ended up having some great real estate at the end of it. But like, I didn't even think about, about that until later when I, when I, when companies, as we went to like the, the, the bigger trade shows, when we, when, when Apple stopped doing the macro trade show and we started going to the retail trade show, we started to meet more VCs and they were like, okay, there's square that's coming up with a bit more on the, on the, you know, mass market. Then there's like shopkeep, which we eventually later acquired. And there was, We were seeing people that were sort of in our space get big investment, and, and we felt like, okay, then that's not going to be great for us if we don't have the same kind of, uh, firepower. We don't have the same kind of, um, ba…

AI assessment note: “I just didn't think about it at all. Cause I have no finance background”

Q What does that look like? Like what's an ex, what's an example of like high business acumen? Do you remember anybody that comes to mind, you're like, this person, there's just something they said or something they did that was like, yeah, this person gets it.

A I'll give you an example. I invested in this company called Power. Founder is Randy Fernando, and I helped him with his previous company, which was sold. He had a successful exit. Very, very sharp business acumen. He is a magnet of talent. He's able to attract amazing people. I mean, if you look at that, I'm like, really? You're gonna convince that guy to join you? I mean, he's, he blows my mind away every time I talk to him. The pace at which he learns is just a fast clip. I would tell him, point him in a few directions, and connect him to a few people. Six weeks later, I come back and talk to him. He has mastered something, because he's unlocked a whole new set of things that wasn't accessible to him previously. He is such a fast learner. So the ability to attract talent, the ability to learn, and the ability to solve problems and figure out which balls to drop. By definition, startups are chaotic. Startups are broken. The master of, uh, of this game needs to know which balls to keep up in the air and which balls to drop. That ruthless prioritization matters a lot. When you hit something out of the park, it really needs to go all the way. And there's no point hitting like, you know, one run here and like, you know, 10 feet out. That doesn't make a big difference in the startup world. The ones that areas where you choose to be successful, those areas, you need to be 10 X bette…

AI assessment note: “I invested in this company called Power. Founder is Randy Fernando”

Q Is there something typical? Like, would you say it's more like One meeting or more like six months? I mean, what, what, what do you think? Is there even a normal or is it just totally case by case?

A It's totally case by case. I can form conviction in less than an hour. I can, I may need a hundred hours to form conviction, a hundred hours, not just with the founder, but on my own as well, doing research on talking to people, talking to potential customers and do it. When I'm able to make a, If you ask me, when am I able to make quick decision? When I am on, when I have an informed mind. If I've noodled on the topic previously, I know that, oh yeah, this is a really interesting idea. I've researched this. I've talked to a bunch of people in the past. Those are topics where I can make very quick decisions. Most often I make these decisions in one or two meetings. It's very, very rare that it takes more than a few weeks for me to make a decision. Doesn't need to. If I can't, if I don't have the informed mind and I'm starting research today about something, it's going to take a few months for me to learn the various aspects of the market, and that's a waste of time for the founders. I encourage founders to find investors who have an informed mindset.

AI assessment note: “It's totally case by case. I can form conviction in less than an hour.”

Q And is that something you can see, like with Imogen?

A Historically, we could only see it through invasive tests, so you'd have to stick a wire or a catheter down somebody's heart artery. Obviously, not that many people are going to sign up for that kind of procedure. But in 2005, we, um, saw the advent of a new technology called a coronary CT angiogram, or this, it's performed on this, what we used to call the 64 slice CT. Now we have like 640 slice CT scanners and so on. And in truth, we didn't understand the vascular biology. We thought that all plaque buildup in the arteries was bad. Um, what we did was we utilized it as a tool in the context of research trials, clinical trials, To better understand what we were seeing and how that affected somebody's outcome. And could we improve the natural history of that outcome by intervening with lifestyle modifications, medical therapies, even stent and bypass surgery. And what we learned was a lot of surprising things. We learned that not all plaque is bad. In fact, that some plaques are extremely dangerous for you and the strongest predictors of who will have a heart attack and others are actually Very stable and almost protected against, um, having a heart attack in the future. And so what we said was, well, if some of the plaques are good and some of the plaques are bad, how do you turn the good plaque, the bad plaques into good plaques? And so we had done a serial study where we had…

AI assessment note: “saw the advent of a new technology called a coronary CT angiogram”

Q And, but I mean, on the analysis side, like why did it take kind of eight to 10 hours of manual work to, to look at that? And what's hard about even today applying AI to it versus applying AI to, you know, my hip x-ray, or is it the same level of difficulty?

A I think it's slightly more difficult. Like the heart, like if you look at this cat scanners, like the vendors who make them, like the GE, Siemens, Philips, Cannons of the world, like, um, they, you know, they've typically made their holy grail, like, Um, scanners that can image the heart accurately. Why is that? It's because it's the most challenging. The heart arteries are very small, and so you have to have something that has very high spatial resolution, like the megapixels on your iPhone, um, camera, and then you also have to have something that has very high temporal resolution, meaning the shutter speed on your camera has to be able to image a moving object And render it motion free, sort of like when somebody's running across like the field and you try to take a picture with a camera with a slow shutter speed, they just look blurry. And so both of those things had to come into play in order for you to be able to image the heart very accurately, very non-invasively with these CAT scanners. And so I think that the acquisition of the image is very difficult compared to static things like your hip, like it never moves. It's just going to be there. Right. And then the size of the Plaques are just very small sometimes, and so you really need very accurate AI algorithms in order to effectively do this accurately. There's, you know, a number of studies that we've done as a compa…

AI assessment note: “I think it's slightly more difficult. Like the heart, like if you look at this”

Q Yeah, like, how does that, maybe tell me more, like, how does that trial work? 7500 people, all asymptomatic, and you image some, and you don't image others, and then you, you do things about it. Is that, like, just high level?

A And then people act upon, uh, the results, doctors do, and then there'll be a follow-up, and we'll see which one is better. You know, all of the data that has been published to date would suggest that Imaging guided or imaging based evaluation on early detection of disease works. I'll give you some examples. Screening mammography. Screening colonoscopy. Screening pap smears. Screening lung CTs. All of those things to combat breast cancer, colon cancer, endometrial cancer, lung cancer, share the commonality that they all use some form of imaging, uh, in order to do early detection of disease at a point in time where it's treatable Easily rather than at late disease when people are suffering catastrophic events. And so we believe that the same concept holds true here. Heart diseases, coronary heart disease is a silent disease in the majority. We will never get to them if we, as doctors, just wait in the clinic because they die at home. And so we need to go upstream and screen the world to find those people before their events occur. In order to do that, we've got to prove it, right? And we've got to prove it within the context of a really well-performed, you know, um, very generalizable, uh, uh, randomized control trial, which is what we're doing.

AI assessment note: “And then people act upon, uh, the results, doctors do, and then there'll be a follow-up”

Q And, you know, let me ask you this. Did you, did you think at any point, especially given the nature of, of this, of, of clearly, like, was there a point where you thought it would break? Like where you thought just not, you're just not going to make it to the next step for whatever reason, it's going to fail.

A Oh, there were some dark days. That's for certain. Yeah. And there was more than one dark day. Tell me about one. I mean, during the pandemic, right, like, we were 17 days away from being out of operating capital. That was a dark month, like, where, you know, and, you know, it's, you know, you, you think, oh, okay, well, you know, you'll just tell people, just be patient for 30 days. No, that's not what happens. Like, legally, you cannot do that. You have to put them on furlough or fire them and do a riff, and, and if you do a riff, you can't get them back, and who's gonna go on furlough with a startup that is, you know, can't raise money? It just, those were some dark days. You know, and then there are times where, you know, early on, like when we started doing the product, like About eight months in, we're like, these machine learning algorithms don't work. Like what's going on? And so like, that was a very stressful few months. And then we worked really hard at it. And then about three, three and a half months later, suddenly they started working really, really well. And so, you know, there's just like, and it's all different kinds of things, raising money. It's the product. It's the, you know, working with the customers. It's like trying to get through the CPT codes. Like, At any given point, I can enumerate for you probably a hundred pretty, pretty bad days. Like, um, but …

AI assessment note: “we were 17 days away from being out of operating capital.”

Q If you couldn't tell, that was all satire. Thought we'd, uh, switch it up here on the PMF show for April Fools. But, uh, that's, that's clearly your, your Brent Butter. You're very good at it. What is your real job? Like, what do you, I mean, this is funny, but like, how do you make money from it?

A I, yeah, so I read, uh, satire about the, the tech industry, um, which Jabrowe Capital is just a satire newsletter. So we make money on advertisements, just like any other newsletter and media business. Um, and then we have a consulting business that works with founders who are raising money. So I actually have built a bunch of companies. I have raised, um, a bunch of money for my own companies and none of which were actually scams. Although some of them, many of them did not work out. And a lot of truths were in that story you just heard. Um, but, uh, now I work with startup founders to help them raise money, but most of my time is spent on this, this satire brand writing, you know, funny newsletters. I built, I built fake products now. I've launched a product called Is My CEO a Fraud?

AI assessment note: “we make money on advertisements, just like any other newsletter and media business.”

Q for them. So walk me through that. I mean, like, what, how, how does Skip the Dishes get, because, you know, I call it the Uber Eats of Canada, which is kind of unfair, because I think you were, you were before that. Like, how does, What's the origin story there? Like who has the idea for skip and how do you guys kind of get that off the ground?

A Yeah. So, so we had, um, we, we knew one another, like the founding team. We knew each other from, from university. We all went to the university of Saskatchewan and we were for the most part, uh, university athletes. Um, Chris, Josh, uh, Samir were on the track and field team. Uh, the older brother, Dan was a volleyball player on the, on the Husky volleyball team. Um, Andrew Chow is actually the, the odd man out, although he's kind of like a, like I call him like an intramural league, uh, basketball player. So he's, and he's actually like pretty good, but we all kind of knew each other from university and it was actually Josh. He was working in, in London at RBC capital markets. And I think that, um, the way that Josh works, he's always like, you know, absorbing what's going on around him. And one of the things, if you work in capital markets, these, these people are Are extremely efficient, driven, um, working a ton, uh, working on really, really big opportunities. And I think what he saw was that a lot of people were using a lot of time-saving, uh, services, and that could be on demand, you know, dog walking on demand, dry cleaning, uh, could be getting car services everywhere. They're just like literally setting up their lives so that everything is focused on work. And Josh was like, oh, Hey, like what are the other things that people are doing? Like cooking takes up a lot …

AI assessment note: “it was actually Josh. He was working in, in London at RBC capital markets.”

Q What was some of that? Like you mentioned the feedback, like what, what were like, what do you remember from back those early days that, you know, those early restaurateurs were telling you, especially things that were kind of insightful.

A Yeah. So, so for example, um, Here's a pretty easy one, but you know, we get a lot of, well, maybe I'll start on the restaurant side. So on the restaurant side, um, a lot of restaurants would say, I really, and this is actually an example of kind of like half listening, but also half, not necessarily doing what you're told, but kind of looking at what people do. So a lot of restaurants would say, Hey, I want to be able to print the order out. And we were like, okay. So then we build the ability for them to print the orders. And the other thing, but then we realized though, is that Paper is, is analog, it's fixed, but the process of preparing food needs to be dynamic because Depending on how backed up that kitchen might be, and depending on where that courier picking up that food is, those are gonna change. So if you print something that says, hey, the courier's gonna be here at 12 o'clock, but then what happens if the kitchen gets backed up, and now the order's not gonna be ready for 1230? Now that courier's sitting there for 30 minutes. So we actually had to, like, say, hey, we're gonna enable you guys to build a print, but what we're gonna do is actually very quickly move people over to a tablet so that they can See the real time that they should have the order ready and also communicate with the courier if they, if the order gets pushed. So that was an example of just like s…

AI assessment note: “a lot of restaurants would say, Hey, I want to be able to print the order out.”

Q How quickly did things kind of like grow?

A Yeah. And I mean, in the beginning, like you're growing off a very small number, but, um, the crazy thing about skip is that when you actually look over the course of the business, like we grew on average over 20% month on month, um, for like. Seven years. So it was like pretty crazy. And obviously you're gonna have like stronger months, weaker months. The hard part was actually growing when you have things out of your control working against you. So weather is, is, you know, it's very dependent on weather. So what would happen is in kind of late April, May order volume would just tank. So growing in a month where like no one's ordering food, it means that you literally have to work 10 times harder to get order volume in that month and versus like maybe December, you could like chill out and you're going to get 20% growth. So to kind of consistently grow meant a lot of people working really, really hard and looking for every way that we could continue to keep up with demand.

AI assessment note: “we grew on average over 20% month on month, um, for like. Seven years.”

Q two and a half million bucks, and I have to assume, like, raising, I mean, this is what I, what I gather, especially talking to a lot of, and I went through founder startups, so I felt a bit of it, this is back when, when founder startups still had a decent brand before all this stuff happened, but it's pretty easy to raise, right? Like, coming out of YC?

A It, it is, like, it certainly gives you, uh, kind of that brand recognition, or I guess you could call it, um, but at the end of the day, they're not, You know, we have 10 customers, right? They're not really banking on early traction. They're, they're once again banking on the team and the space. I mean, with craft, there was like a, a weak process. You know, we, we met, uh, once and then, you know, they did their diligence for about four or five days and then had a real deep dive where, you know, the partner, Jeff, uh, really dug into the business. I mean, it, it felt like a YC interview and that it was just, uh, Pretty hardcore line of questioning. I would say, I think we did a good job on them. Then literally the next day we got, we got term sheets from them and we're able to close around.

AI assessment note: “It, it is, like, it certainly gives you, uh, kind of that brand recognition”