The Path from The Soviet Union to Building Multi-Billion Dollar Companies — Max Levchin

You're having enough fermented foods versus the fermented food you've selected for yourself. So you relax some constraints and move on to spending less time thinking about them. I track sleep. I'm still an obsessive cyclist, so I track every metric you can imagine. I become more obsessive as a cyclist, which actually bothers me a little bit, but because I'm unwilling to believe that I'm ever going to age properly, I keep on looking for marginal gains by obsessing over the exact crank length adjustments I can put myself through just to return some of the marginal watts that I lose with age. I've gone from believing that all you need is cycling to widening my physical training routine too. So it's maybe slightly more brain allocated to what's the right day to lift heavy weights versus the right day to not do that.

True story.

So, I did have some fairly gnarly respiratory stuff when I was a kid. And I think my parents probably were scared into throwing everything and the kitchen sink into it. As a nice Jewish boy from Ukraine, I probably should have played the violin, but clarinet is a reasonable close second. They're like, 'Oh, but lung capacity, great, let's do that.' So that's how I got into clarinet. There was literally something that will feed the stereotype but also help with lung capacity development. And so I think my VO2 max at the time must have been like in the single digits because I remember just trying to open up my lungs and breathe in and it didn't work. And then like everything else I do, well, I can just obsess over measuring it. And VO2 max is definitely not a thing Soviet Ukraine really measured, but they did have the basic approximation—blow into a tube and see how far you can push the ball, all the physics equivalents of lung capacity measurements. So I was already looking at quantifying stuff from a lung capacity perspective as a 7-year-old kid. All of this while living in an apartment complex that overlooked Kiev's only outdoor velodrome. And so my platonic ideal of a manly man who was a sports phenom was these dudes who would get on the boards of a Kiev velodrome and just put their watts down and go for an hour or two just ripping through this thing. I was like, wow, one day when I'm not a scrawny little kid with a clarinet, I'm going to be like that. So I used to sneak into the velodrome when it was closed and try to ride my bike there except the cranks would touch the woodboards. One of the worst accidents of my childhood was I had a giant wood splinter in my butt because I crashed at the top of the drum trying to do loops and slid down. Of course, these weren't well finished boards and it was outdoors, so the elements rained on it. Probably not PG-13 after this, but anyway. So I didn't grow up with a ton of sport in my life. Soccer is prevalent obviously in Eastern Europe, but cycling was right there on display every day. If I wanted to see someone who looked just full of life and vigor, I would see a cyclist. And so I always rode a bike as a kid and always rode a bike in college. And as I got into busier and busier life, this notion of how do you stay sane became an important part of the consideration—if you're going to work for three days straight without sleep, you've got to offset it somehow.

I sort of naturally looked at cycling and then back to Nellie. When I left PayPal after we got acquired by eBay, I definitely hit a very low point where there's nothing going on in my life that was worth staying up all night for. And she pointed out that I have this amazing bike hanging on the wall that I've had since I think before we met. She's like, why don't you get back on the bike and see if you can find joy in that. And she knew my stories of rolling around the Kiev drum. And so I got on a bike and it was one of these moments where I went for a ride with a bunch of people and they were all very serious road jewelry, as the saying goes.

If you're buying more of a bike than you deserve to ride, that's nice road jewelry right there. I was on a beater bike relatively speaking. I think I brought it from college actually, and I sort of dropped everybody on the first hill and I was like, oh yeah, I know how to ride a bike. I have big lungs. That clarinet really paid off. And so that was like, I should really invest some time into getting better at this thing. And that's how I got into it. But I am exactly what they call a hammerhead. If I see a road, I can't not go fast or at least try to go fast.

A couple of different things. It's a low impact sport. So you can ride a bike for six hours and no part of your body will be in absolute shreds the day after if you at least somewhat know what you're doing. So if you have the time and you're interested in it, it's not a self-destructive sport, which some sports are more and others are less. This is definitely one of the less self-destructive on the actual physiology. Part of why I like to go hard on the bike is it actually allows me to clear my head. It's one of the few times when I'm not constantly playing with some work-related concept or thinking through some future ideas or whatever. It's a moment where everything gets tuned out and it's just me and the pain in the legs and you're telling yourself go harder or manage your energy but there's not enough time or cycles left and also what if we went to this market now? So it's a pure moment of flushing whatever's left in the head. And it does keep you healthy. It's a great way to stay alive and be healthy. And then more and more these days, it is a lifestyle sport of CEOs and leaders and people who read the same kind of books. And it's another form of community where if you don't have the time to join clubs—I'm not a big foodie. I'm not a big drinker. I don't have a ton of time for protracted social events. I try not to go to too many conferences and yet I'm human. I enjoy human company and so riding with a pack of people who are interested in similar things is another form of staying engaged.

Part of the trick is if you can still talk and they cannot.

Cycling takes place over hours and so you can spend some time talking. There's also an important part of cycling culture—the coffee shop ride where you sort of hammer to a coffee shop and then everybody falls in their chairs and gets their pinkies out and sips espresso. And I love coffee almost as much as I like bikes.

I think it's Jens.

I think he's East German. Jens Voigt.

One of my proudest moments on the bike is I rode with Jens a few years ago. He used to hang around Northern California. He has a lot of fans actually. He's transcended the very specific sport of cycling into having a fan base worldwide because he's such a character. The shut up legs—I own multiple pieces of clothing with shut up legs that I did not personally produce. He's famous for that line, but he's very funny. He's now primarily a commentator in bike races, but he is a great embodiment of this culture of endurance. Pain with a smile, if you will.

I'm pretty worried, not to sort of go all dark immediately, but I think people without the benefit of—questionable benefit of growing up in the collectivist paradise, to quote the current mayor of New York—don't understand just how corrupting it is. The ideas of socialism are amazing. Me for my fellow men, share and share alike, do the right thing because it's the right thing to do, not because there's a financial incentive. All those things sound amazing and so it's seductive—the idea of a workers' paradise without the greedy lenders, capitalists, bankers—all the things we were fed as children in Soviet Union. At first blush you're like, 'Yeah, it makes sense. People who work hard should have more.' And those things sound pretty good, but they inevitably require a bunch of structural change that just does not work thanks to human nature. For example, one of the greatest mental images I have from growing up in Soviet Union is if you went to a government-owned store—and every store was owned by the government—you would very quickly notice that the people whose job was to sell you things from behind the counter were always very fat while everyone you knew in your life was always very skinny. And you think, this doesn't make any sense—how do these people who work in food stores are always really well fed? Because they're stealing—they get access to the food. And it expands that notion of everybody pooling all their work product into one big pool and then someone's in charge of fairly distributing it—from each according to their abilities to each according to their need, the socialist/communist motto. And it works great except people who are doing the actual redistribution get to keep a lot for themselves. And it doesn't matter how honest they begin that journey, by the time they get to real power, they become profoundly corrupt and steal and keep and redistribute primarily to themselves. That alone is an indictment of socialism that you cannot get around. It gets worse because markets, free markets, capitalism—inherently forces competition. If you believe you have a thing you can sell at a lower price than the other guy, you get the market's attention. You work on creating a margin by lowering the cost of production, finding efficiencies, making the thing cheaper for you and cheaper for the buyer so you can compete on price. None of that exists in socialism because it's all centrally planned. 'We're going to make this many widgets and then redistribute them to all the right people, full stop.' And what happens is there's never the pressure to improve and so you always make stuff at the government-mandated cost and the government-mandated price. You have natural stagnation. You have a system that rewards graft, gives power to people who are most likely to become graft-driven, and prevents anyone else talented from trying to innovate and improve the efficiency of the system itself. It just does not work. The death of the Soviet Union was a surprise to exactly zero people because we knew as people who lived there that nothing ever changed for the better. Everything was the same price, the government-mandated price. And we were still using phones that looked like they were made in the 1950s.

Yeah. Yeah. And so anyway, I can't say enough how the seductive story of socialism really appeals and I can understand why. But if I didn't have my day job, I'd spend a lot of time screaming from every street corner, don't fall for the trap. It's a surefire way of getting to no progress at all.

I'm confident I don't know every strand of this particular disease. It's a complicated issue and I think it is definitely the case that the brochure looks great. The storytelling of 'do you know someone who's poor? Do you know someone who deserves better? Do you know someone who's been laid off?'—it is the case that capitalism can be profoundly unfair at least to an individual. As a system, as a way of improving the world, it works amazingly well. We've not created anything better. But you get thrown off the bus if you build something that gets outmoded or outcompeted. And it can hurt. For the entrepreneur, it looks like failure, which we are wired to accept, but it still hurts. For a worker, it looks like being laid off, which could be catastrophic for a family. The idea of social safety net, the idea of welfare is natural. You hate seeing your fellow men and women starve or not do well. So I understand the definition of the problem and income inequality—there's a lot of people who are not doing well and every new disruptive change from the industrial revolution to the AI revolution, there's always someone on the receiving end of the efficiency because their work is no longer required. Socialism is a compelling story of how to protect those people. And I think there's definitely a lot to do to make sure we don't leave our fellow humans behind. My love of capitalism and free market does not obviate my humanity. I see it as part of my job to ask, can we do better? Can we provide the social safety net for everyone? We're creating more and more efficiency and value. There have to be ways of helping people who are thrown off the bus. The idea of let's concentrate it all in the hands of the government, give a handful of people the right to redistribute it all—doesn't work, and I lived to tell the tale. I do think philanthropy is profoundly important and the more disruption we're going through, the more important philanthropy becomes. I think religion provides a really useful set of frameworks around how to think about it. Every organized religion has a version of what it means to be philanthropic. On the other side, I think capitalism has created opportunities for products to develop where optimizing for the most profit at lowest cost ends up building products that are devolved from their societal ideal. And I don't think we have to compromise by saying, 'I don't actually care that this is harmful to someone because it's the most profitable thing to do.' My theory for a long time has been that you can build products—certainly financial services—that are optimal not just from the lens of most profitable but also most society beneficial. Affirm is the answer to that question. What we are doing here is trying to say, you can optimize a product beyond just making the most money. It's okay to make a little bit less money if you create something that's more societally important, because in the long term, consumers will not go to a cheaper product because it's so societally important. I am to some extent embracing the ideas of pro-social product engineering. I just choose to do it strictly through the lens of capitalism.

Well, there's this acronym BNPL—buy now, pay later. We kind of invented the genre about 15 years ago. The basic idea is at the point of sale, most of us pull out a debit card or a credit card. Those who pull out a debit card generally think of themselves as financially responsible—if I can't afford it, I'm not going to buy right now. If you pull out a credit card, the minority of us who are for all intents and purposes independently wealthy just say, 'I'll put it on my credit card and pay it off at the end of the month.' There's a huge percentage of Americans that pull out their credit card and say, 'I'm going to add it to the giant pile of revolving debt I already have and I'm going to pay interest on it and I don't really know how to calculate that and I have no idea when I'm going to be out of debt, but I need that thing.' And sometimes the thing is a new bicycle which we can maybe postpone, and sometimes it's a pram to put my baby in. The idea of Affirm was—and it's worked amazingly for 15 years—that what if there was a third way? What if you could have the same transparency and responsibility and clarity of a financial decision with a pay-over-time product that you do have with a debit card? What if we added the idea of I want to pay for this over time, but I don't want it to revolve and I definitely don't want to be confused as to when I'll be out of debt? Essentially turning every transaction into a simple plan. Pay for my baby pram over six months, let's say $100 at a time, doesn't break the bank. Exactly six months out, I'm done. Not revolving. No more interest. And as we described the idea, we thought—if you're going to do it right, how about you build it so there is no revolving possible, there's a fixed schedule only, every transaction is pre-priced, if you're paying interest you know exactly how many dollars of interest you'll ever pay, that can't change, and if you're late there shouldn't be any late fees. You should just be motivated to pay on time because if you take too long, maybe we won't lend you next time. So it's almost like a too simple, too black and white a way to lend money as an alternative to credit cards. It is available now at three-quarters of all e-commerce checkouts in the US and Canada and rapidly expanding into the UK. We've announced we're going to go to a bunch of European countries. We sort of purified it down to no fees, no revolving, simple schedule, everything pre-priced, everything super transparent. This year we'll do almost $50 billion of these transactions. So it's worked out.

It seems to be working. We're publicly traded, quite profitable. The last 10 quarters we grew 30-plus percent year-over-year or faster. It's both very big, growing really well, still never charged a penny of late fees, never charged a penny of revolving interest. We're still true to the idealistic nature. As I was describing the product, I talked to a bunch of banking people who laughed at me like, 'Okay, dude, you don't get it. You've obviously never lent money for a living. Late fees is where the profit is. Half the profit comes from late fees. You just cut your profitability opportunity by half.' Putting that aside, if you don't revolve, what are you going to do when people take too long to pay you back? You want them to take forever. You want them to make minimum payments so the principle keeps being big and interest compounds into principle. All that sounds crappy. None of this sounds transparent. People in Silicon Valley who generally don't need these services don't get how significant the cost is on your regular consumer. As I was trying to raise money for this idea, I would talk to VCs and they're like, 'I don't know, why can't they use their platinum Amex?' They don't have a platinum. 'Why don't they get a platinum?' I'm exaggerating, but only slightly.

Anyways, the difficulty raising money for this financial services company, unlike PayPal, was I couldn't convince anybody this was a thing that normal people would use.

So the earliest version of this thing is actually pretty funny. Abridged since it's a colorful story—and actually Luke is involved, Luke makes another appearance.

Right after PayPal went public, Luke and I—Luke is a wild and crazy guy. I think he talks me into it. I'm already dating Nellie and I'm trying to impress Nellie. I decide the way to impress Nellie is to get a really cool convertible. So Luke and I find this convertible. Mercedes just launched their first hardtop convertible. We fly to LA with a view to buy these two identical—we debate, we can't have the same color. So we get to LA, we go to the dealership like, 'We want to buy this cool hardtop convertible from Mercedes, a mechanical folding hardtop.' So it's like a transformer car that becomes a hardtop or a convertible if you press a button. Super cool. We get there and they're like, 'Okay, Mr. Nosek, here's your auto loan. Go. Here are your keys. Mr. Levchin, your credit sucks, so we're not going to sell you a car today.' I'm like, 'Wait a second. We literally just took PayPal public. I'm okay. I've paid off my student loans and everything.' They're like, 'Yeah, but your credit score is really, really bad. So if you want to buy a car today, you're going to have to wire the total amount of money.' My future wife, then girlfriend, is like, 'Ooh, okay, you didn't tell me that part of your biography. What did you do?' I started this company in college with Luke Nosek, who's in the room. He knew and I didn't that if you don't pay your credit card bills on time, it'll show up on your permanent record. So I had to wire money and as the dealership was wrapping up the day, they're like, 'Oh yeah, we finally got your wire. Here are your keys.' So we drove from LA to San Francisco on I-5, basically racing each other at ungodly speeds and got pulled over at 2:00 in the morning. The cops were very upset but actually let us go because they were similarly fascinated by these cool new cars. We didn't get into a lot of trouble, but got into a minor amount for going way too fast at 2 o'clock in the morning. But the story stuck with me because I was fine financially—better than fine. I was my mid-20s and never needed to work a day in my life. And yet the car dealer who knew who I was—he looked us up. He was like, 'Oh yeah, I know you guys. You're PayPal kids. Cool. You can buy a car. Sorry, your credit sucks.' Why does my credit score not reflect the fact that I'm basically independently wealthy now? And that really stuck with me. The story came up over and over again. And I was like, I should do something about this. There's got to be a credit score to be built around—came to the US as a teenager with no record of any kind and $600 for a family of five and now independently wealthy Silicon Valley entrepreneur. It was a five or seven-year window between those two events. Seven years. I got a computer science degree. I was imminently employable. I started a bunch of companies. Not all of them failed. And why did it not catch up to me at all? It seems to have had no bearing on my ability to borrow a fairly modest amount of money relative to everything else. That's been sloshing around my head for years. And I finally sat down with another Luke friend from U of I—what if we built a better credit score? We're all CS majors. We all have some number theory background, therefore machine learning, therefore AI. We built a score using publicly available data and some secret sauce. And like, oh, now we should make people lend money using our score. So I talked to some bankers—which is how I began the 'oh, you're doing it wrong, kid' conversations. Nobody's going to lend money using a score that no one else is lending money against. What if it's a bad score? If you tell an entrepreneur over and over this thing will never work unless someone does X, the natural response is, I will do X and see if it works. All right, I guess we're going to lend money. At which point I was like, I got to understand how this stuff works. I've never actually looked at lending carefully. That began the journey of, how do credit cards really work? How did I manage to mess up my credit at 18 by missing three payments? A year later I was like, oh my god, this whole thing is so broken and no one even knows it. Revolving on a $1,000 draw and finding yourself with a $3,000 debt a couple years later and you can't explain it. Or getting a 0% loan but not realizing that if you're a dollar short or a minute late, it will compound retroactively. There are all these things the industry built that are profoundly anti-customer. It's one of the only industries in the world where you and your service provider face each other and you're like, 'Hey, I want to borrow some money.' Great. I'm going to bet on you failing. I'm going to give you a loan, but I want to believe you're not going to fail completely, but you're going to be really late. Ideally, you take forever to pay the loan back because I make more money that way. Optimally, we sneak in a few things you don't notice. It's all in the fine print but I hope you don't read it. It's a very strange vestigial—capitalism is supposed to get to efficiency. This got us into a bad part of the decision tree. What if I took a bunch of steps back to the branching point where you could have done the more consumer-friendly thing and just did that over and over? What product would I end up with? Oh, ended up with Affirm. And that's how we got here.

I would pitch this idea in its various forms. First of all, I did have to fund for a while on my own, which I didn't mind doing because I was getting completely obsessed with the idea. That was a small price to pay. I would keep telling the story to a CEO or former CEO of a bank or somebody who managed a credit card portfolio, and they'd be like, 'You're doing it wrong. All the money's in the late fees, the money's in delinquencies, come on.' I had two ideas that both turn out to be amazingly right. By the way, every business has a moment of luck where people don't really want to admit to it. Affirm had two. And there were ideas I'll claim some ownership over, but I had no idea if I was right. The first one, I read this study that said millennials hate banks. It didn't really explain why. Just millennials hate banks. They hate a lot of things. Every generation has a list of things they want to complain about. I'm not a millennial. I'm a relatively young Gen Xer. But millennials were willing to participate in some study where they ranked all the things they hate. And the number one thing—like 70% of millennials—hated banks. So every time I would ask a banker, why can't there be a lending product that doesn't take advantage of people's unwillingness to read fine print and do exponential math, they're like, banking is the stickiest thing in the world. You bank where you live. You like your bank. It's where your parents banked. I'm reading a study here that says they hate you. All the people that are going to be shopping and buying stuff 10 years from now, they hate you. 78% of them think banks are terrible.

I believe that study. Like, okay, they surveyed like 10,000 people. That's a pretty stout study. So if I'm right and I do a thing that's even slightly better than the existing thing, I have a ready-made audience of people that are going to be like, cool, I hate the other one anyway. And at some point I need an explanation beyond just they hate it.

And so what I realized, and I think this is the part that was lucky because I didn't know any better in the beginning, but I've since sort of confirmed this to be true. Millennials were early teens during the great financial crisis. And so if you are getting booted out of your house in '08 and you're like an impressionable youth and you're asking your parents, like, what's going on? Why do we have to live in a motel now? Parents are like, because the bank, they just took the money, they took my house away. The embarrassment and the horror of telling your kids we're going to move to a smaller house or to a rental or we're going to live on the streets, god forbid. That's a horrible impression to have in your most impressionable years.

And so I think there's a lot of people who are actually quite genuine in the 'I am willing to try anything but the thing that my parents got the sharp end of the stick for.' And so that was kind of luck number one. And I really believed that from the very beginning and it turned out to be true. If we build something that's good, normally you have to market it. You have to convince people to trust you and all that. I think they don't need any of that. They just need the belief that there's a better thing out there that exists and they'll give it a try because they are predisposed to not liking the other part. And that turned out to be completely true.

The other part, so if you are unwilling to profit from all the sort of wacky externalities of financial services, you better be good at underwriting. So if you're lending money to someone, you are paving over underwriting mistakes or lack of underwriting or poor credit score.

When somebody shows up and says, I'd like to borrow $100, and you want to respond to them in real time. You have some giant machine learning model or a small machine learning model. Decide yes or no. And what that practically means is it has to take whatever available data that person is willing to share, be it public or private data, run it through some form of classifier or these days all kinds of interesting architectural sort of ideas exist in underwriting, and say yes or no. But yes or no is actually not enough. What you really want is a price of risk. So it's some, for simplicity, expected value or expected loss. If I give you money today, what are the odds? What's the expected value of you bringing it back to me with the appropriate amount of interest if that's what the price of the transaction is. And so underwriting discipline of doing that at scale, in our case completely in real time, that's the score. We built the underwriting score. The credit score was that.

And the theory that I had was credit scoring is hard and you would have to have the absolute best people working on it, and yet most of the time when you encounter people working on underwriting or credit scoring they're not as impressive as people I'd met in my wandering Silicon Valley up until that point. And like, why is that? It's an interesting problem. It's quite mathematical. There's some very smart mathematically inclined people I know.

And the best I could get to was it's embarrassing to talk about that you're working on loan and credit underwriting stuff at cocktail parties. Like when somebody asked you what do you do? You're a math genius. You went to school for applied math or computer science. What do you do? One answer is, well I went to Wall Street and I build real-time trading models. That's kind of cool. Yeah, maybe it's like not society important, but you're doing something interesting. And then all the other answers: I work for the NSA and I break codes. That's cool. That's what I wanted to do when I was in college.

If you're like, I make lots of loans and I make sure that they're underwritten well, and by the way, I make most of my profits by charging people late fees and sneaking in nasty terms. So there's got to be a latent pocket of talent of people who would absolutely work on underwriting because it's a really hard and really interesting problem, but they won't join the industry until someone shows up and says, I'm going to strip it of all the gunk. I'm going to make it completely transparent. I'm going to be super pro-consumer. I'm going to take a lot of pride in the brand that we have, which stands for transparency and honesty and all the good things you can do if you would just take a broom to the whole thing.

And so as I sort of formulated the product, I'm going to get my unfair share of really brilliant mathematicians because they're not going to go to Wall Street because it's kind of a soul-hollowing thing to do to squeeze pennies out of the market. They'd rather come to work for me and build underwriting models with me trying to help people in normal America borrow money and not feel screwed or not get screwed. And they're totally, exactly right. Like we have people who've been here for 10, 11, 12 years doing that job who are still like, I'm so proud of what I do. I'm a mathematician and I'm putting my big brain to work on making honest financial products. So those two things were good ideas but they're also a lot of luck.

First of all, the reason the growth rates are compounding at a almost $50 billion mark, 30-plus percent year-over-year, quarter after quarter after quarter is pretty awesome. It is a staggeringly good growth rate. It is made better by way of noticing that our credit-related losses are super consistent. Like it's easy to grow the top line. I've lent more money this quarter than I did last quarter. Great. But will it all come back is a really important question. And the way you measure the success of these underwriting models is you ask the question, okay, so you made a bunch of loans last quarter, last year. Average half-life of our loan is five months. So five months back, you get a pretty good picture of how good are you relative to today's macroeconomic reality. And if you look at our loss numbers or delinquency numbers, whatever metric you want to choose, they're really, really consistent. So we are in fact very good at underwriting.

And yet even at 50-ish billion, we are about a footnote. So the overall credit card debt in the US last I looked was like 1.3 trillion. So we're scratching at the total size of just the credit card debt and like not all credit card transactions become debt and there's plenty of debit card transactions. And so the overall size of commerce is enormous relative to where we are and so our growth is not a surprise in a sense. What may have been a surprise at some point was people actually like this product.

There are reasons to say that this product is harder to use. Credit cards and debit cards for that matter is the single best financial interface ever created. Like, if you think of the level of complexity that takes place when you take out your card and tap a checkout counter and just walk off with your cup of coffee, there's a lot of stuff that happens underneath. There's an acquiring bank, there's an issuing bank, they're talking through a network, there's all kinds of complexity just at a technical level, and then there's credit and blah blah. There's like layers and layers of many decades worth of innovation that make an incredible single transaction happen in like a tap and off you go.

And so when you introduce buy now pay later, which is this idea every transaction is separate, you're conscious of every transaction. You understand how long it's going to take to pay this transaction off. You're actually creating friction. You're giving people more steps. So, one counter-argument to why this thing makes any sense at all is like why would you want to trade the beauty of tap and go with this like open your app and do some stuff. And the answer is simple. The industry devolved and you don't really trust what's going to happen to you after you tap. So, you hesitate and you're like, am I going to really be able to afford it or is my card going to work? Am I going to get declined or am I going to go into debt that I probably shouldn't go into?

Affirm offers the antidote of you go into the app, you look at your purchasing power, you know exactly what we are comfortable lending to you, you get explicit approval, next transaction is guaranteed to work, and by way of saying we're not going to charge you late fees or change any of the pricing. We're actually telling you, you should feel very good about this transaction. We are going to make less money if you're a minute late. We'll make no money if you don't pay us. And we feel good about lending you this money.

What we sell, or what we offer to the consumer of this, is certainty and sense of control by way of creating some incremental friction. Fast forward a couple years, I don't have enough of a clear crystal ball to know exactly what agent commerce looks like, but I know a lot of it is happening already and I think a lot more will happen. The discretization of transactions and this incremental friction will be reduced.

You don't actually have to do all the manual labor we are putting you through right now because your agents will do it for you. And so today we're offering a thing in exchange for some friction. In tomorrow's world, we're offering the thing that obviously works. $50 billion can't be wrong. But the friction will go away or largely go away, which I think is just going to accelerate this whole approach to financial transactions.

And the way people think about money will change towards: I know exactly what the total financial state I'm in. I have agents that are looking out for me. They won't get me into debt that I shouldn't be in. They will get me out of debt the second I should be out. I will have a PhD in consumer finance embedded in my phone looking out for every penny I got. There will be no slop. And by God, no one's going to fool me again with fine print-driven business model.

The world in which you have AI looking out for all of your financial concerns is a beautiful world because we are already there by way of not having any dependency on 'you're too dumb to know what's happening to you. So just pay up.' That's how a lot of the industry works today. And so I'm very excited about the world where you don't have to change your business model because suddenly no one's getting fooled, because that was never a part of our business model.

It's very close. I think it's very dangerous to consider commerce as this uniform fabric that just kind of want a thing, buy a thing. One of the dangers, by the way, of building on mass market products in Silicon Valley, if you're not inventing the future, in which case you're green field, blank sheet of paper, if you're trying to improve something that exists and has existed for a while, it's essential you travel to where normal people live and see what they do because we are not normal in Silicon Valley. In global Silicon Valley. This is not limited to San Francisco and the Bay Area.

We think you just want a thing and buy a thing. And if that thing is $10, great. And if it's $10,000, well, you know, we get paid a lot. So, okay, just buy the $10,000 thing. And you care about speed and efficiency and less distraction. And normal people actually hang back and like, whoa, it's $10,000. First of all, that's an incredible amount of money. And by the way, for someone who is not within these hot beds of growth and opportunity, maybe $1,000 is an incredible amount of money. Maybe $500 is an incredible amount of money. Depends on where you are.

And the sort of, ooh, I got to think about this. I want to make sure I'm not exposing myself to a bad financial decision. I also want to make sure I'm getting a good deal. How am I going to pay for this? How long is it going to be on my personal balance sheet before I'm done paying for it? Like those are all questions people are very conscious of. And so for a lot of people in every part of the US, agent commerce is here when you tell Instacart bring me a sandwich or bring back your groceries, or DoorDash bring me a sandwich or vice versa.

That's like the threshold of is this enough money for me to sort of hang back and ask these questions? For most people on whom it is too much, they're probably not using these services just yet. For whom it's sort of like yeah, I want a sandwich. I want it now. And like I'm going to outsource the rest of this to an agent be it a human or a robot, don't care, like you have my credit card number, bring me a sandwich. And I think that's here today and we will see more of that become a thing.

And many things like real-time delivery or near real-time delivery. If you go to other markets, you have, I think in India somebody launched a six minutes or less delivery service which boggles my mind but their storage of mass market goods to buy has got to be just like an exercise in extraordinary logistical prediction. Anyway, so I think that stuff is here. It's coming and more will happen.

The 'I want to know what I'm getting into. I want to have agency in selecting a thing. My taste matters to me' is a thing that probably will always involve humans. 'I need a pair of pants' is not actually a thing. If you're a Silicon Valley engineer and you're too busy coding and you have a big hole on your butt, you gotta need some pants. Agent, bring me a pair of pants. Sure. For most people, like, I want them to make sure they fit and I want them to be my favorite color and I also want it to match the rest of my ensemble and so on. And it goes up from there. Like I need a bicycle but I have brand preferences and my components are important to me, etc etc.

And so AI will not replace the need for decision-making and thoughtfulness in consumption. It will obviate some pieces like who has the best price on the bicycle I want with the component tree that I prefer is a thing that you will gladly outsource. Like you need to make sure it's real. You need to make sure that the data LLMs are fetching for you is current. But that's what we're working on today.

But at some point very soon you'll say, hey Chad-bot of the moment or of my browser, of my desktop, I would like to buy a beautiful looking Italian-made bike with Shimano components. I want a good deal. I'm probably going to pay for it over time because it's expensive and I definitely want to pay no late fees and ideally I don't want to pay any interest.

I'm bringing my business to someone who should be so lucky because multiple people will sell me a beautiful Italian-made bike. And go do some comparison shopping for me. Bring me some images of beautiful bikes and I want to lust after all of them and then pick the one that gets my business. That's a great task for AI.

The building blocks for that will include something like Affirm where the AI will say, all right, so all these people offer some ways of paying. Some of them offer Affirm which has no late fees and by the way has negotiated a special deal with a manufacturer or the seller where they will pay your interest for you. So they're actually covering the time value of money. So the plan is interest-free, which a lot of our transactions are interest-free, exactly this way. They're funded by the retailer or the manufacturer. And we're grasping at that future right now. We're certainly building a lot of the pieces.

So, you know, I tend to be slightly ahead of schedule as far as the future I want to live in. I'm trying to pull it in here, but this is quarters, not years.

For the foreseeable future, there's so much to build. The thing that irks me the most in today's media is the SaaS apocalypse, whatever. But the job apocalypse people are proclaiming is so goofy. There's so much opportunity to build so many exciting things for everyone, not just Silicon Valley startups. Everyone from the oldest companies to the youngest ones are like, oh my god, there are all these pieces of software that I had to buy from someone who did a bad job or that I could never buy anywhere and I wouldn't know how to start with. Well, now you have all these amazing tools that just birth it for you straight from your head.

Seems a lot closer than I would have predicted if you asked me just a few years ago. It's wild.

The world changed a bunch of times over the last 12 months, but I think like the Claude Code moment last December was like a big whoa, it really is here. The ability to produce something from a glimmer in your mind to a thing that actually works reasonably well and has only improved since was a big breaking point. The one before was obviously ChatGPT and so on.

I'm not paid by any of these brands, so I just want to make it very clear. I have some very strong opinions.

All right. So, first of all, this is primarily relevant to espresso. So, I love all coffee in all forms. I happen to prefer espresso as a beverage of choice, but I don't judge. You can have your lattes, your americanos, whatever. But from the point of view of espresso and espresso-based drinks, the single most important thing is the grinder. And so, you can get a fantastic grinder like a Niche or these $600, $700 grinders that will elevate your game to an incredible level. So, if you have a grinder that you didn't pay a couple hundred for, got to go do that.

If you want to get to like a demonstrably better grinder than some of the sort of the $600 range, go get something that looks like a KafaTec. That's like a $1,600 to $2,000 grinder. It's a great grinder. You can swap out burrs. So, switching out burrs is really important. But even if you're never going to replace your burrs, you still want something that's a flat burr grinder. And a good one will cost you on the order of $2,000 plus or minus. And then if you're like, you know what, no price too large. I'm going to get me the absolute fantastic home grinder.

Obviously there's like industrial strength stuff that's even out there. There's this thing called Weber Workshops that produces unbelievably expensive but unbelievably good stuff. So if you want a Weber, God bless. That's a great product.

No, different. Totally different. I had this exact question yesterday.

Anyway, so better grinder goes a very long way. After that, go invest in skills. So, there's an incredible number of online resources that'll teach you how to do a great job. I'm but a student of those people. There are many, many, many great videos you can watch on YouTube that'll teach you how to make a great cup of espresso. Before you buy anything else, just go watch all that stuff because it'll teach you a lot of things you don't know about. After that, you're probably going to want to buy a good espresso machine. The great news is that there are many, many very good espresso machines.

I don't want to begin religious wars here, so I'll state my preferences, but these are preferences. I prefer 58 millimeter portafilter diameter. Many people now swear by 54, 53 millimeter. You can go up or down, but I stick to the classics. I think the finest, most reliable home machine is made by La Marzocco. That is definitely in the multi-thousand range, but that's what I have on my counter and it's never failed me and it's beautiful both inside and outside.

There's lots of cool, if you're a super nerd and you really want to nerd out on the metrics, Decent Espresso is a fantastic product, really fun, but it's like an Android tablet that happens to be attached to an espresso machine. Basically, La Marzocco is old school. That is the logo you'll find in every self-respecting coffee shop. Like the long red typically spelled out La Marzocco, that's the granddaddy of them all. It's amazing.

And so on. And there are many, many other brands that will sell you a good espresso machine. If you're trying to go down market a little bit, like Breville will make a good cup every time. Breville.

It's very, very good.

Yeah, your obsession. So, I'm going to force your hand though on one here.

Chemex. Why Chemex? If you're going to go for a light roast, each type of roast, each type of bean speaks to a method. So, this is like another heresy, by the way, in the modern day. I like medium dark roast for my espresso. I love the honeyed viscous, feels like it's too thick. You can almost chew it. I love that in espresso. Like that's a great cup of espresso.

For non-espresso drinks, I actually love thin, high clarity, low body, just give me the pure essence of not diluted, but the bean in water. And Chemex are probably the best one of those. It's really the most controllable. Beyond that, French press is great. That's like a campfire type setup. And AeroPress, if you're going to go there, just get an espresso machine.

You're like in the territory where I probably like, you know what, no coffee today.

I've been known to exhaust the local supply of pre-ground pods in hotel rooms. So there's coffee and all the beauty of it and then there's caffeine and you know I need both.

The spice grinders as they're known, you can definitely make a cup of coffee with that too. But that said, by the way, if you must have a thing that is not an electric grinder, get a manual grinder. Those are very cheap.

Lido, if I recall correctly, is a brand that'll sell you like a very, very high-end, but still an order of magnitude cheaper than any of the stuff that I threw out. I mean, it's a little bit of a good workout, but you'll make beautiful coffee with that because there's a really nice correlation to the lower the RPM, the less kind of you're damaging, if you will, the bean as you're grinding it. So, you can actually get some amazing taste out of manually ground beans. It does take like 10 minutes per cup, though, so you got to be ready.

Yeah, switch arms every once in a while.

AeroPress, great campfire attraction.

So I grew up drinking not just instant coffee which was hugely popular in Eastern Europe. I occasionally was exposed to chicory coffee which I don't know if you ever tried that.

Except there's no coffee in it. It's just sugary root.

Yeah. I mean it's like kind of sort of tastes a little bit like coffee maybe. Pour enough hot water into it. If it's hot enough you're like ah it doesn't smell right but whatever. In Soviet Union was one marketed under 'coffee light.' The coffee beverage. Something meant to evoke some plant component that both coffee bush and chicory would share. So like, well, you know, they both have it, so it's good enough. It was disgusting, but it'll wake you up.

There's that.

Yes. So, here's a couple. There's not a panacea, but I will preface this by saying I have extremely low degree, by which I mean like zero to negative, of respect for business books. Vast majority of them are far too long, present company excluded.

Business books that are verbatim anecdotes of people who've done it are actually high value because I think it's very hard to distill what is inherently a collection of unique experiences. So like you can edit it down, but you can't make it generalizable.

So, I'm a huge fan actually of what you do in your books because there's always so much color directly from the source versus 'I think this person meant X when they talk to me and I'm not going to print anything the same thing they said.' Anyway, that said, I'm a big fan of business books that have been distilled to business essays because then I feel like the author did the work of like I'm going to generalize. I'm going to really generalize it.

So, there is this book called Seven Powers by Hamilton Helmer. If you haven't read it, it's a really worthwhile distillation of what it takes to build a competitively lasting business. It talks about why network businesses are longer living than non-network businesses and what brand actually means. Like people, oh you know, I'm going to build a brand. Like why do you care about having a brand? And so it's a great book. Distills a ton of these kind of things that you kind of think you know, but like they need names and terminologies.

It's idiosyncratic in the sense that like I didn't think of the word power. I didn't think of the term power until I read his book and I found it to be a great kind of distillation. It's very short. It's slightly shorter than Master and Margarita. So anyway, that's a good book.

Another good book that is off the beaten path, but a great one called A Failure of Nerve. It's a book about leadership and it really postulates this concept of differentiated leader. So as a CEO, as a founder, a lot of times you find yourself at odds with your own team where you're making an unpopular decision. You're firing a beloved employee. Whenever there's doubt, there's no doubt. Then you go do it and you're like, holy crap, what have you done?

So you find yourself in these moments where you have to persevere before and then after a decision that you made and many, many flavors of it. It's a good book that kind of teaches you how to think about it, how to tolerate the stress that comes with it, how to sort of put up with the pressure you're going to get and not lose your humanity and not become a tyrant, but also not be someone who is easily bowled over into like, okay, okay, I'll reverse my decision because then you lose the confidence of the people you're supposed to lead.

So, that's a great book. I am sure it's in your list and everyone's list, but if you haven't read it, if you're trying to start a business, you should read Influence by Cialdini because that is probably the most important social science book published in the last 50 years.

There you go.

I love biographies,

I think Chernow's biography, so Chernow wrote a bunch of these like tomes, speaking of enormous books.

Hamilton,

Mark Twain is a great one.

Titan is probably the one that's closest to like business advice. It's a bit of like how to be ruthless basically, but that's besides the point.

Exactly.

Those are great books. There's a great book called A Mind at Play, which is Claude Shannon's biography.

Yeah.

American polymath and mathematician. Yeah.

Fun fact, you know Claude Code named after Shannon.

He is a great example of someone who worked on some very serious stuff and was always playful. He's like Richard Feynman of Computer Science. So, Feynman was famously brilliant. And there's a great quote from Murray Gell-Mann. Somebody asked him like what's your algorithm for solving really hard problems. Like encounter really hard problem, step two, give it to Richard Feynman, step three, receive solution.

Shannon was kind of that in computer science, proto-computer science obviously, he was building this in the '40s, '50s, '60s and he managed to remain fun. He managed to have fun. He was just this unbelievable fountain of fun ideas and he was like a hardware tinkerer and made card counting devices and all kinds of hilarious stuff.

Exactly. And so that's a good book about him. These days it's so hard to read books because there's so much content online in podcast and blog form.

I have like a 500-page graphic novel that I'm going to be lugging to New York City because still the digital experience just doesn't do it. But you know the burden I bear.

It's not there. We'll be able to replay our dreams before we get that.

Affirm.

No, my work is out there. I try to build in public. If you're interested in Affirm, obviously affirm.com is where you find that. I tend to go deep, so my project list is short at any given moment, but levchin.net, @lev either one is the list of those things.

Great to see you. Safe travels.