Why AI Agents Could Finally Reinvent the Credit Card

Key Takeaway

Audit one recurring purchase or planned big-ticket item today: identify the exact product you want, compare the total cost, delivery reliability, and payment terms across sellers, then choose the option that best matches your priorities. The episode’s practical lesson is that convenience dominates low-stakes purchases, while higher-cost purchases deserve transparent financing and a deliberate comparison of price, trust, and timing. Use AI to narrow options—but keep final judgment for purchases where quality or fulfillment risk matters.

Episode Overview

Alex Rampell and Max Levchin trace payment innovation from PayPal’s early rejection of anonymity to Affirm’s evolution from a “pay with identity” concept into transparent consumer financing. They explain why cards remain an unusually durable payment interface, why small frequent transactions create outsized payment economics, and why AI may transform the payment step before it fully takes over shopping decisions.

Full Transcript

Transcript of Why AI Agents Could Finally Reinvent the Credit Card from A16Z. Auto-generated from episode audio; may contain minor errors.

The card payment interface is the singular best user interface ever created. It is the world's largest market by any stretch of imagination, and there are no niches in payments that are smaller than $100 billion. Once you go really big, the numbers get small, which is strange. There's a lot of volume, but the large volume revenue opportunities in payments tend to be the smaller dollar amounts. There's always an opportunity to use another form of payment delivery device to satisfy a basic need. Convenience just trumps as the total amount you're trying to send goes down.

The best user interface ever created is the credit card. This may actually be finally up for renegotiation because AI is already there. It's just that you haven't yet trusted your agent to do as good a job as you would. Something else that you're surprised has not happened yet. I went to some cryptography-related conference and presented a new idea in digital payments and was literally booed offstage because they're certainly not anonymous. The big innovation at PayPal was, what if we don't care about anonymity at all? In that sense...

Payments may be one of the oldest categories in tech, but Max Levchin argues there are still no small markets inside it. In this episode, I sit down with A16Z General Partner Alex Rampel and Affirm co-founder and CEO Max Levchin to look back at how payments evolved over the last 25 years and where the next major shifts may come from. They revisit the early days of PayPal and the origins of Affirm, including the pajama problem, the first experiments with paying using identity, and the moment the team realized financing wasn't just another checkout option, but a way to materially change conversion.

They also discuss why the credit card remains such a durable interface, how 0% financing can have very different economics, and what happens when AI agents begin participating in commerce. Max is skeptical that we'll hand over every shopping decision to AI anytime soon, but he thinks the payment itself may finally be ready for reinvention. So you guys go way back, having co-founded Affirm. Well before then, you guys are pioneers in fintech. You've been thinking about the space, trying to make sense of the present, think about where the future is going for 20 plus years, 25 years, maybe even more.

I'm curious, given so much time has passed since you first got into this space, what has most surprised you about what has happened, what hasn't happened, what did you expect to happen in the early 2000s of how this space would play out? I think the rise of Apple Pay and Google Pay and the extent to which they've really penetrated, because it's very hard to change consumer behavior in general. It was a bizarre set of accidents, if you will, where there is this merchant liability shift because mag stripes were so easy to replicate that you probably noticed a while ago, your new cards had a little chip on them and you had to dip the chip and then the machine would be like, don't remove, don't remove, remove, remove, right?

That's because the chip was being rewritten. It was much more secure than a mag stripe. And basically Visa and MasterCard and then the EuroPay, that's what the E is for. What is ECM for in EMV? EuroPay? EuroCard? EuroCard? I don't know. The Europeans somehow got their say because they always do. They insert their cookies. Definitely not electronic. Yeah, it's Euro something MasterCard Visa. So it was called the EMV switch. So all of these machines, if you did not want to be liable, like I could go to Best Buy, buy a TV and then get home, use my TV and say, nope, never bought the TV.

And if I was using my mag stripe and Best Buy didn't have the machine switch over, then Best Buy would be like, oh, shoot, we lost all the money on that TV. And Alex gets the free TV. I'm exaggerating slightly for effect. So everybody had to get a new machine. And then it turned out that those new machines had a little contactless thing where you didn't have to dip the chip or swipe the card. You could just tap. Nobody was using taps when those machines came out.

And now it's ubiquitous. So that's something where, I mean, generally speaking, it's hard to change consumer behavior. But I feel like between COVID and the fact that all these merchants independently had to switch over their machines and the fact that, I mean, I wouldn't have guessed in, I mean, you obviously did because the confinity was Palm Pilot sending money. So you're ahead of the whole PDA wave. But I wouldn't have guessed that those three things would have come behavior. It's like change of consumer behavior, ubiquitous new merchant payment terminal, plus ubiquitous mobile telephony.

And that's certainly one thing that has changed the world a lot. How about you, Max? You were dreaming about what the future of money would look like a long time ago. But it would have surprised you about what has happened or what hasn't happened. It's very hard to step back from the soup if you're cooking in it for the last plus or minus 30 odd years. I think to sort of further illustrate Alex's point, one of the things that's really amazing and subtle about Apple Pay slash Google Pay and its interaction with the Visa MasterCard networks is Visa and MasterCard have a hard two and a half second limit on interaction between the network bashing bank and the merchant and the acquiring bank.

So the whole thing has to happen in two and a half seconds or the transaction gets retried or maybe just canceled out and really leaves very little room for any kind of clever innovation. And online, you can play games. If you're doing e-commerce, you're like, well, yes, we're going to submit your card to Visa, but first we're going to run some anti-fraud checks and we're going to do some other things to reduce liability. But offline, like once your card is presented, two and a half seconds, that's all you got.

And Google and Apple Pay have singularly time shifted the whole thing by creating secure enclaves inside their chips and saying, I already know your card. I can do all kinds of things before I actually have to talk to Visa and MasterCard. And maybe the shocking thing is that Visa and MasterCard have not yet introduced some new standards saying, actually, it doesn't have to be two and a half seconds at all. It could be 15 seconds while we go and get a bunch of issuers to bid on a better credit quality terms for you or any other type of innovation.

But the sort of hard-written rules of DHOC era are more or less intact. That's not 30 years, that's like 60 years. So I think that's probably a critical or more criticizing take on what happened. But the shocking thing about payments is that it is the world's largest market by any stretch of imagination. And anything and everything you could have possibly thought of being sort of deconstructed and being the nichiest little thing to poke around and innovate in always turns out to be $100 billion. There are no niches in payments that are smaller than $100 billion.

That's probably the most surprising thing I've seen. Although, it's interesting. Once you go really big, the numbers get small, which is strange. If I send a $1 trillion wire, the United States just announced we are $40 trillion in debt. Let's just say that Elon's Mars colony is worth $40 trillion and Ro Khanna decides to tax him $40 trillion and he needs to send a wire, that wire is not going to be very profitable. There's a lot of volume, but the large volume revenue opportunities in payments tend to be the smaller dollar amounts, which is surprising.

Because it turns out, all of the QSR, the quick serve restaurant, the payment opportunity there, that's massive. That's why Starbucks invented Starbucks Pay, because they wanted to stop having to pay. It's like these prepaid, they only have to pay one time. You put $50 on your Starbucks account, well, they only have to pay Visa, MasterCard, and everybody else one time as opposed to every successive time. What's surprising about payment area under the curve, if you will, is that the bigger the dollar is, the smaller the rake, which I guess makes sense, because you're not going to get 2% of that $40 trillion payment.

But everybody goes after B2B payments. Everybody has a clever idea for B2B payments. That's the only one that is probably the exception in the world. I was trying to find... It's lost in time, the original email between the two of us, which I think we have slightly different origin stories of how we actually met. But I was trying to find the original email between the two of us where we start mentioning what becomes a firm eventually. And it's dated April 2011. And in it, I think I asked you, tell me more about this Build Me Later thing, because I was working on something else while you were watching PayPal buying Build Me Later, and you sent me this really long, elaborate, sort of, here's exactly what it is, and all these ideas around it.

And we then veer off into this ping-pong of emails around Build Me Later for business, and conclude that it's actually a dumb business. There isn't an opportunity. It's funny that you mention that. I bought the domain PayMeSooner as a result. So it's like, ah, Build Me Later was a good idea. PayMeSooner would be an even better idea. Actually, the idea there, it's not terrible, because it's not payments, it's lending. So think of it this way. Imagine that big companies beat up small companies all the time to get better payment terms.

Better payment terms for them, right? So if I'm GE, and you're Eric Torenberg, and you're a little guy, I'm big GE, I'm like, I'm going to pay you net 90. So you invoice me, and I pay you 90 days later. Now, you have to pay your employees. So you go to a bank, you go to somebody else, you're like, hey, I'm Eric, I run this business, most of my money comes from GE, they're an awesome company, they're going to pay me in 90 days. Bank doesn't care, right?

They're just like, yeah, we're going to charge you 15%. Meanwhile, GE can issue bonds at whatever, SOFR plus 10 basis points or something. GE is paying 5% for credit, you're paying 15% for credit, but it kind of doesn't make sense, because your credit is GE's credit, if that makes sense. Now, there's something called factoring, where you can sell the invoice, you can actually sell the receivable. That's somewhat usurious. That's very, very, actually, it's not usurious, because credit laws don't apply, because you're selling a receivable. Paying me sooner was an interesting idea, because the whole economy runs on the fact that my business is waiting on payments from you.

The reason why I have to borrow money is because I don't pay you sooner, and yeah, we didn't pursue that. There's some healthy businesses these days in both accounts, payable financing and accounts receivables financing, so it's not a bad idea, but the revenue opportunity, despite being an unregulated lending space, is seemingly lower than the one in consumer, because the convenience factor is something that every participant understands, and there are plenty of players that will offer you a slightly cheaper term, which is also the root origin of why a $40 trillion wire to Rakhana is gonna be not a very profitable business, whoever processes it, yes.

What's another idea that, you're surprised, does not exist yet, that you were thinking at some point, hey, the world obviously should work this way, there's an obvious opportunity here, and for whatever conflation of reasons, just hasn't happened yet, or to use Mark's quote, there are no bad ideas, only early ideas. Is there something else that you're surprised has not happened yet? You know, we're forever in search of the great way to pay biometrically, and it manifests itself more in movies, where people's thumbs and eyeballs get removed for authentication purposes, and yet we're still paying with chips.

By the way, the card payment interface is the singular best user interface ever created, and it's not something that people haven't tried to do better. Remember, even before PayPal, there was this wand that I think MasterCard and one of the gas refueling station networks produced this thing where you could just wave this wand next to a gas station, and payment is secured, and you can fuel your car, and when I saw it, I'm like, oh my God, this is obviously gonna take over credit cards, and it didn't.

And it's actually, therein lies a really interesting lesson in payment innovation in general. There's some unobvious a priori, but clear a posteriori version of critical mass, and if you don't reach it, you're going to fail. There's not really a okay outcomes in payments. You either get there, and everyone needs to have your widget, or your network, or your whatever, or it's just gonna go into the annul of time. And this is a fine example of a thing that actually made a lot of sense. Of course, you wanna have a little wand on your car keys to just wave and get refueled, but it's just a little bit faster, not a lot faster than the credit card you have in your pocket, and that just works.

And so, I keep on waiting for some completely different way of authenticating the payer to the device, and other than the mobile phone, we haven't done better. Amazon actually just discontinued the whole thumb payment. I love, it's not a thumb payment, it's a palm scanner. The palm, yeah, yeah, yeah. I loved it. I would literally go to Whole Foods next to whatever hotel I'd be staying in. It was sad that they got rid of it. I know, I used to use it, too. It's actually, it's not even faster, it's just fun.

I know, it's probably slower, actually. It is slower. It's like, tell me my fortune, and these grapes. The fortune is you owe me money. Yes. That is the only fortune that you ever get. Well, that was, I mean, I think I called it thumb by accident, just given the almost name of a firm. That's right. When the crypto industry was becoming popular, or when some of the major projects were starting in the space, were there parts of that that seemed exciting to you, or a realization of a long-held dream, or did it feel, hey, orthogonal, or just not really?

I'm frequently late to trends. I'm not the world's earliest adopter of majority of things I encounter. I feel like PayPal was pretty early, right? Well, sort of. To give you full context for PayPal, I really wanted to build something very specific and quite different from PayPal that operated on very low-power chips, and to do that, I needed to work out how to make cryptographic primitives work on a very, very low-power chip, and we did, and then we promptly realized that it really couldn't do very much other than encrypt some very, very small amounts of data, and small amounts of data that need to be encrypted and decrypted quickly that kind of lead you to payments, which is how PayPal came about.

But right before all of that began, I went to the wind-down-slash-bankruptcy party for DigiCash, which was the original granddaddy of all digital payments, and it was sort of a somber affair. There were Stanford University grounds, and it was like pouring out of the 40s, and a bunch of very sad-looking cyberpunks telling each other how the age of digital payments is not upon us, and so in that sense, we were very late when I was listening to these people talk, whose idea was crushed. It was blind signatures, and David Chum was somewhere.

He was actually not present, but he was roaming the streets of Palo Alto despondent because his brilliant idea, and a truly brilliant one, was basically being proven to be not having a product market fit. And then we promptly, soon thereafter, started PayPal, and I went to some cryptography-related conference and presented a new idea in digital payments and was literally booed off stage because it was not nearly as secure and certainly not anonymous. And the big innovation of PayPal was, what if we don't care about anonymity at all?

In fact, people don't, they just need to pay for their coffees or their online purchases. And in that sense, I feel like we were, again, very late to that game. The game had played out like all the enthusiasts had left the building. And so cryptocurrency, not to be confused with cryptography. So I read the original Bitcoin paper and thought, wow, that's a really clever way of solving the business in general problem. And I sort of, as all the sort of math slash cryptography nerds was impressed with the approach.

But I don't think for a second it would become a currency or a payment method, and et cetera. And I'm not convinced it's a payment method now. And it's been quite some time. But as currency and commodity in a store of value goes, it's certainly proven to be extraordinarily successful. And so I think as I watched cryptocurrencies expand and become a thing and more and more of a thing, the use cases are coming out now where stable coins are upon us. And that's certainly a collection of very clear uses.

But none have, in my opinion, come close to breaching the, I'm going to use this to buy a cup of coffee, which is kind of, that's the canonical and most important, I would argue, use case for all forms of payments. So as currencies go, as store of value goes, fascinating. You know, what is the last time you spent a Bitcoin or a Satoshi on anything actually important? You know, putting aside people who make it a point of spending it just to sort of prove the point wrong.

And it's the most important. Buying coffee is the most important metric because of just the volume of it, because the frequency, the practical. There's always an opportunity to use another form of payment delivery device to satisfy a basic need. So if you're contemplating a $40 trillion transfer, you'll spend a lot of time figuring out the secure way, the fast way, the cheap way, et cetera. If you're passing through your nearby bagel shop and you need a cup of coffee, if your cryptocurrency wallet password is too long, you're going to look for change in your pocket.

And if you don't have it, you'll pull out a debit card or credit card. And so user interface, as the payment amount diminishes, it takes over, cost takes over everything. Basically, convenience just trumps everything else as the total amount you're trying to send goes down. Yeah. I want you to take us both back to memory lane. It's the beginning of a firm and how you guys navigated the IDMAs of what would eventually become a firm. You go here. Well, hopefully our stories match. Yeah. No, separate room.

This is like the prisoner's dilemma, right? It's just like asking if our stories match, we get to leave. If they don't match, then we have to stay here all day. So my recollection was I ran this company, TrialPay, which did alternative payments for digital goods. So you don't want to pay for throwing a sheep or doing something on some silly social game, something that was not hard, valuable, or fun. Who would do such a thing? But you're buying coins in FarmVille. You're doing something. You're buying poker chips, virtual poker chips, of course.

You don't want to pay. Get it for free if you sign up for Geico. There's a lot of economic value to Geico. You don't really care if you use Progressive or Geico. Ooh, I can get FarmVille coins if I switch to Geico or get a credit card or sign up for Netflix. So that's what we did. And I was somewhat persuasive and annoying, probably more annoying than persuasive. So I signed up every other social gaming company, except for this one called Slide. And Max ran, proudly or unproudly ran Slide.

And I went to the Allen & Company conference, which I almost didn't go to because my wife was very, very pregnant. So this was in March of 2009. It was like the first week of March. My son was born March 28, 2009, but her due date was March 16. So I'm like, I'm not going to go to this conference. And then the Allen guys were like, this is a very good conference. You should really go. I was like, my wife's very pregnant. He's like, you should go to the conference.

If she goes into labor, we'll fly you back. It's like, all right, that sounds good. I'm going. So I went to the conference. I think we met there briefly. I successfully did not persuade you. But then I think I followed up with a note in Russian. Yes, exactly. And this wouldn't have worked today because now, obviously, I would have used Gemini or ChatGPT to write this probably somewhat riddled with grammatical flaws Russian note. Actually, that probably would have given it away of not being ChatGPT. It would have passed Pangram.

But I took Russian in high school and college, worked there, lived there for a while. And I think you were like, why do you speak Russian? I think that was the response. But then we just became friendly. You sold to Google so I could not ever convince Slide to become a client of TrialPay, sadly. But I seem to recall you had a tweet, like, I'm looking for something to do, any idea. And that was my reason to reach out. And I think we had coffee at the Forbidden Building at Google.

Remember how there was one Google building that no guests were allowed at, but you just didn't care? Statute of limitations. You're revealing all my secrets. Exactly. Because I know how much you love Vic. So we had coffee there. And then that's where this, yeah, we were talking about family later, but it was like, you know, my recollection was, it's really hard to pay for something on a mobile phone. Mobile phones are becoming more ubiquitous. We should have a way of solving the pajama problem of you're upstairs in your pajamas.

You want to buy something. Your credit card is downstairs. How do you pay? Meanwhile, not just our mobile phones ubiquitous, but social networks are ubiquitous. And maybe there's a new underwriting mechanism of kind of going back to the old general store from like the 1800s of it's like you show up in the general store. You don't have cash, like both literally and figuratively. You don't have it on you, but you don't even have it, you know, back home. But you run the general store and you're like, okay, Max, I got you.

Don't worry about it. But today you're just like a cookie in an IP address. So I don't really know who you are. Or if you go to Walmart offline, it's like, you know, you get greeted by the very nice greeter. I don't think they have them anymore. But like you get greeted by the nice greeter. They have no idea who you are. You can't do the general store thing. But if you have 500 friends on Facebook, this was wrong, by the way. But if you have 500 friends on Facebook and they all have, you have 1000 pictures that you've uploaded, you're probably a low credit risk.

And if you're not looking for credit, this is the key thing. It's like a lot of credit offers are sent out proactively. So you have an 800 FICO. I know that you have $14,000 that's revolving for some reason on your Capital One card. I'm going to send you a customized mailer saying, why don't you go refinance with me, Bank of America. But if you go on Google and you say like, I'm out of money, need money, need money, credit, credit, credit, like you're probably a bad credit risk.

So kind of going back to the general store concept, people that aren't looking for credit that might not have their wallet with them kind of make it like the 1800s and pay with your identity. That was kind of how I remembered it. Yeah, I think that that matches. The to add a little bit more color to it. So the year at Google was definitely a challenge in a sense that I was still coming to terms that I ran for like over five years, a company in social media that I had no business running.

Like I worked very hard, headache financially, very positive outcome for a lot of people. But ultimately, it never sort of scratched the sort of entrepreneurial itch that I thought I was scratching. So one of the things that happened during that year, I spent a lot of time soul searching, like what am I supposed to do next with my life? I was going to start a company, but maybe maybe should be more thoughtful about what company I'm going to start next. And my wife, who's pretty much always right, said, you know, the hardest you've ever worked and the happiest you were was during the anti-fraud days at PayPal when you nearly died or we nearly died.

She was there as well. Um, those were the days when you looked exhausted, but you actually were obviously very happy. And, you know, I know she was telling me, I know you swore off doing financial services. I know you're never going to touch payments again, but just this once, give give it a chance. And that actually had some influence as to my responsiveness to because you work in a payments company. And part of my reluctance to deal with trial pay was like, I can't handle another near payments company.

Like I went into social media for a reason. I don't want anything to do with this stuff. And so as I was slowly getting real back into working on payments, we started talking about this. The sort of a social BML was one of the sort of shorthands we used discussing what it would be. The general store analogy was interesting because at the same time, I actually remember which one of us knew this because you also speak Japanese. And this could be the thing you told me, but I remember reading about social credit in different economies.

There's a notion of just pay me later or pay me next time in Israeli grocery stores, which is still common today. If you're checking out in Israel, apparently in a small enough town, you can just show your bag to the cashier and they'll remember. And then they'll get you next time if you forgot your money. This is even more formalized in Japan, where until very recently, you would give your business card to the storekeeper, even if they didn't know you. If you had a business card, they would write what you bought on the back of it.

And that was your total. The notion of I just put it on my tab is a thing that's obviously fairly American concept and so on. And so we were converging on this. I was talking to a startup at the time as kind of just learning about the industry that was trying to build a social credit score, kind of modeling it on the Chinese social score, which is used for a lot more than credit, obviously. And so we were swirling around this idea of payments and solving the I don't have my credentials with me, but I still need to get through this.

Another sort of metaphorical version of this was I'm watching TV. There's this really cool ad back in the day when people had advertisements in between things and TV. And I really want the thing that was promised to me at the price that was promised to me. But I could do it on my phone, but I'm definitely not gonna get off my bed. And so there's some area under the curve that you can capture by bringing these transactions closer in time. And improving conversion, obviously. And so that's the kind of the swirly origin story.

A lot of it, I think the two of us had a slightly different version of what would be fun. So I think as a payments guy, you were mostly thinking like, let's go do some payment stuff. As a person recovering from doing lots of machine learning in the service of social media, which was, you know, good and well, but wasn't really what I was meant to do. And lusting for the days of doing machine learning to fight fraud. My MO was, I just want to build a really cool credit score.

And then someone else will take care of payments, like lending payments. I don't really know anything about it. And I've done payments, don't need to do lending. But building a really cool score, that would be amazing. Because I remember just the reams of data we had access to at PayPal. We could maybe get some data. Maybe we could buy some data. We can mine Facebook for data. I'm both trying to get in touch with Mark Zuckerberg to see if he might give me some of his data.

And so a lot of that was sort of in my head while you were like, I just want to close some transactions. I think that maybe if there's any difference between the motivations, you were much more motivated by, let's go find a merchant. But the similar motivation was like, you know, trial pay. I think we were in the throes of selling to PayPal, but then they left me at the altar. Not your PayPal, the many, many generations later PayPal. And I was like, I never want to start a company again.

This is so painful. And you had been in the throes of that as well. So it was also like, let's get this thing going and we'll hire other. Well, but then you introduced me to Nathan and Jeff and we kind of formally incorporated. But I also like the other thing that I distinctly remember, and you'll you remember this in a very negative light because it's PHP. But we had a meeting. I kind of think of the origin of being the Allen & Company Conference in 2012.

Oh, yeah, I don't I don't find it a negative thing. No, no, I know. But it's funny. So we have this meeting. Like, so we're going through the list with Son. Remember, you're your old chief of staff, right? So we're going through the list. And it's like, oh, we both know Jim McCann, who's this very friendly guy who used to be on all the commercials for 1-800-Flowers or 1-800-Flowers. You can order flowers, order it right now for me or whatever the commercial was. So we ping him to get a breakfast meeting.

It was like 200 person conference. Anybody can be with anybody. But it's like, hey, why don't we have breakfast? And I pulled like an all nighter making this demo in PHP. I basically like cloned. It's just funny. Like, hey, I can do this in five seconds now. But I cloned 1-800-Flowers site and then made a whole like pay with your identity. So you do a checkout, like add the dozen roses to make up for the bad stuff that I did. Sending it to my wife, click pay.

Oh, she don't have my payment card with me. Oh, I can pay with Facebook. What does that do? It does Facebook connect and make sure that I have over 500 friends. And actually Facebook had, I mean, to your point, they had a lot of interesting data. They had their own internal flags around. Did they think this account was fraudulent or not? And fraudulent back then was more of like it was just created not as a real account. It wasn't a real person because an account is not a person, right?

Hopefully they're the same thing, but not necessarily. Because one person can have multiple accounts. But if there's one-to-one, so that was our fun demo. And then I remember Jim McCann was like, this is great. Let's do it. So two other colorful points from that. So it was a good breakfast. And then in the middle of it, he goes, oh, yeah, we used to do exactly this thing back in the day when the service members would call us and say, send me and my wife some flowers.

And no, I don't have my credit card with me. And we're like, oh, that's fine. Thank you for your service. We'll just get you on the next one. And he intuitively grok this idea that you can absolutely post pay a thing if you have enough trust. And we were basically there raising our hand saying, we'll take the responsibility for you. They don't pay you. It's okay. We're going to eat it. And so he was instantly smitten. He had been a great proponent of the product and a supportive person slash presence in my professional life ever since that breakfast.

And he had absolutely no reason to trust us other than, like, he has some degree of sense who we were. And it was like, you guys are great. You seem like you really know what you're doing. He was so excited. We have no idea. Well, then we got handed to this. What was his name? He actually now runs a successful startup. The guy that was his Amit. Yes. Amit Shah. Yes. Yeah. So Amit took over. We implemented, he was very, very excited. But, and I remember, I came up in Microsoft, I think I did this in Microsoft Word or Excel, like here's our pricing sheet, because we needed a pricing sheet.

And I was like, why don't we charge 7%? So in consumer finance for these things, there's a concept of an MDR and an APR. So merchant discount rate, because you're discounting what you pay the merchant, because you're advancing in the cash right now, you're not collecting it until later. So call it a 7% merchant discount rate. But I think there was like some table, I just literally made this up. And then there was an APR, but the APR here, there was no APR, it was just MDR.

But then nobody, like I showed this to somebody, I think it was Rob Pfeiffer, actually, before we hired him. He had like a real finance employee. Yeah, but I showed him this, he was like, ooh, free flowers, right? Because he's like, you just didn't have to pay us back. So the returns were not fantastic on that. So cynical. But yeah, I think he literally said free flowers. It's like, this is a great demo, but free flowers. Just generally, very smart, but very cynical guy. So we were not gonna be perturbed by some cynicism.

I love Rob. His sister used to work for me. So I remember interviewing him in the context of a firm, although I think at the time he was still Expedite. Yeah, he briefly bumped around as Expedite. Yeah. And then we finally renamed it. What was it? It was incorporated as Expedite Software Inc. Yeah, I think that's right. And so how did the idea evolve as you went from there? When did it start to really get product market fit, how did the space play out in a way that...

Every startup has the 40 years in the desert. It's just a given. If it's worth talking about, I exposed. I mean, sometimes people just quit too early or something happens to destroy it. But I've never run into a company where people start a company and they just hit product market fit and 24 hours later they're just selling and everything's going swimmingly. So you always have this period where it kinda still makes sense to show up to work, but you don't really know if it's gonna end in like a great nothing.

And we were definitely meandering through the great nothing period for a while. After 100 Flowers, we had a sort of a nonstop interactions with said Amit who was never happy with the conversion rates, with the user interface. And he was mostly right. Like he was kind of a grumpy, demanding guy, but he knew what he was talking about from the product and financials perspective and he was definitely not gonna pay 7%. And so we were sort of limping through that. We're generating a little bit of volume, but among other things, the notion of, hey, you can pay us later, showed up after you selected the Flowers.

And so he would say, look, you guys are cannibalizing my credit card volume. If I were charging nothing for it, of course I'd send you more transactors, but you're charging me more than credit card. So it's a dumb idea. I don't wanna pay anything for it at all. You guys ought to be ashamed of yourself basically. And we were sort of trying to convince other merchants, but we had one merchant that was essentially saying, yeah, I mean, these guys are nice guys, but it's kind of worthless.

And so it was going nowhere fast. And then we had another friend, somebody named Nils, who ran a company called Beautylish, which is kind of still around, but they were selling online cosmetics and beauty products. And the only difference that Beautylish did as they implemented Expedite, I think by then it may have already, I think it was named the firm by then, was that they told their shoppers, up-funnel basically, as they were selecting their shampoos and perfumes, that you could pay in three installments or 30 days later or whatever it is we were trying to sort of do then.

And that had an instant 30% increase in conversion. And that was the, oh, we know what this is. This isn't an alternative. This isn't actually solving the pajama problem. It is solving my budget is this, but if I could pay over multiple periods of time, it would expand a lot. And so the second we sort of understood the very beginnings of that notion, we turn it right around into a sales campaign and very quickly found a bunch of small-ish brands, primarily direct-to-consumers, who could not care whether we charged them 1%, 5% or 12% because they were trying to grow their top line and telling their shoppers that, hey, you could pay for this over time instead of having to cough up the cash right now.

It was transformative. And so I would literally get these love letters from merchant CEOs saying, and I'm still friends with a lot of them, like early adopters. Most of these people have now either sold their businesses or the business has failed or they've gone on to do something else. But I still run into people who were, like there's this woman named Tracy from Tradezy. Tradezy.com was a- Oh, yeah, yeah, yeah, yeah. If I had just saw Tracy- I remember her, yes. She's wonderful and she was an early advocate for us because her business literally- I forgot about her.

She would email me screenshots of her dashboards and say, this is the Affirm Effect. Here's a 35% pop you guys cost for me. Like, please do more. Well, there was a solution to the pajama problem and then it was also all the mattress companies. Oh, yes. That was- That was another- I just remember it's like, wait, there's a company called Purple and there's a company called Casper. I was just like, they just kept coming out of the woodwork and that was transformative. And I remember actually we were talking about this in the context of like, how can you charge a high MDR and get away with it?

That sounds bad, not get away with it, but like actually show sufficient value to the merchant. The highest MDRs are actually for-profit education institutions. So, some of them are like 50- I mean, I don't like these guys in particular, but like University of Phoenix. So, there's the private equity firm Apollo, but there's also like this holding company of all these for-profit institutions called Apollo Group or Apollo something. And University of Phoenix I think is the biggest one. And they just assume that most people aren't going to pay and the gross margins are so high because it's an online course.

They're like, okay, you can take 50%, right? But you're stuck with the receivables on the backend, which goes back to the fraud fighting that Max was talking about before because the gross margins of the mattress, like these mattresses, I don't know if you ever bought one, it's pretty cool because they didn't actually ship you a mattress that looks like this. It rolls out. So, you get this like cylinder, it's pretty small. Some of them are boxes. Some of them are boxes, but some of them, it's like you just open it and then it just like, it's like one of those old toys that you'd put in like your bath and then it would expand or something.

That was like the mattress. So, they had very, very high gross margins, so therefore a lot more flexible on the MDR and more importantly, nobody wants, like you're in your pajamas on your uncomfortable mattress. You don't wanna spend $1,200 on a mattress, but if it's like 30, like once you subdivide it, it just has a massive increase of conversion rate. Yeah, there's a couple of other sort of technical things that are probably worth tapping into in that story, but that was the next big leap. So, the mattresses was the precursor to the great expansion into DDC brands, which I'm sure we'll touch on.

But before we get there, a firm briefly flirted with being a payment provider to the online educational facilities, which we ran out of kicking and screaming because the reason they're willing to pay these enormous MDRs is because the quality of education is largely terrible. So, the reason people don't pay most of the time isn't because they're like, oh, I thought about it now that I have a master's degree in basket weaving, I just don't wanna pay for it. It's like, it turns out my degree is worth nothing.

Why did I sign up for this ridiculous thing? And so, we were, if you go back to sort of things like, what is it called, General Assembly, is that their? Oh, yeah, yeah, yeah, yeah, yeah. There's a whole sort of class of these, like, we will teach you how to code. There's another pre-EI thing, like, learn how to code? Just hashtag GPT to code for you. But back in the day, when it was still important to learn how to code, but you didn't really wanna go to college for it, you just wanted to take a six-week course, there were plenty of the University of Phoenix clones that.

Cloud Academy. Yeah, there's all these guys. And we saw that trend and thought, oh my God, like, these people are willing to subsidize education, which is cool and on mission for us. And yet, the loss rates to consumer dissatisfaction were so high, we lasted like half a year in that space. But back to mattresses. So, the thing that's really powerful about direct-to-consumer, or you can think of them as vertically-integrated businesses. So, there's some factory in Brooklyn, presumably, or who knows where. China. I thought the mattress guys manufactured locally, because...

I don't know, but they were all in Utah. Remember that? Yeah. I feel like there was like a Utah mattress town or something, but I think they just all originated there. Casper is in New York, or was it? No, but Purple, I think, was in. Purple. And then there were like other Purples that were... Yes, I think one of the weirder things. The diaspora of Purple. The manufacturing was actually all consolidated. So, they ultimately sort of paid very little for these mattresses that were kind of all the same.

I mean, don't wanna reveal any industry secrets, memory foam is memory foam. Anyway, and so, their primary value, or their value equation was really interesting. So, first of all, they bought this memory foam in a box and sort of compressed and shipped it to you with a cool effect. The margin there were gargantuan, because you are, in fact, just sort of taking petroleum and turning it into foam. But the other thing is the replacement cycle, and this actually goes to a, I think a Harvard Business Review article that came out right before Casper was founded.

And all these guys, all these really entrepreneurial people read the same article, which essentially said this. People replace mattresses once every seven years. If you are the company that sells that mattress, it is the most important thing in the world. And if you miss it, the next opportunity is seven years from now. But the margins on mattresses, even before the sort of the foam and the everything else, are like 80% or some eye-popping number. And so, suddenly, four or five companies were born from that article basically saying like, oh my God, so if we just figure out how to market a mattress that you will either shorten your replacement cycle, so you buy the next mattress six years instead of seven years from the last one, or you just decide this mattress is the best mattress, or the coolest one, or the one that pops out of a box, that is like, oh, there's so much money in it.

And one, it'd result in a saturation of mattress industry, obviously. But more to the point, it created this enormous premium on, I will just compel you to buy a mattress by any means necessary. So telling you, I will give you a three-year loan at no interest at all, powered by a firm, of course, is like a small price to pay, and a cool marketing campaign. And so, there were multiple events like this in a firm's history where we saw an opportunity for the merchant to significantly contribute to the MDR, which allowed us to eliminate the APR.

In other words, consumer got a true 0% load. The backdrop to this is actually really, really important to understand, and one of my sort of things that I rage against, have, and will continue to until it's made illegal or eliminated by other means, is the fake 0% load. So if you ever have gone to a, fill in your favorite department store, there is a sign somewhere, these days somewhat less prominent, that says, get a branded credit card from, fill in the blank, 0% APR. There's an asterisk next to the zero, which says something along the lines of, assuming you make a purchase for the next 12 months or 24 months, if you're a penny short or a day late paying your principal, the interest accrues retroactively to the beginning of time.

Basically, the day you got that card, you swiped it for $1,000, you wake up and you owe $3,000 two years later. So this is called a deferred interest credit card. And a huge opportunity we tapped into directly from that rage was this idea of, our 0% is gonna be a real zero. When you go to buy the inflatable mattress or the mattress from a box, you're gonna get a 0% loan. And even if you're a month late or a year late, we're not gonna change the price because we are so sick and tired of these ridiculous people who are lying to you about zero with an asterisk.

There will never be an asterisk on a firm zero. That is the origin story of why we don't charge late fees, we don't do deferred interest, all the sort of gimmicks and gotchas that the industry loves to throw at people. We've run away from all of that primarily as a point of like, you know what, when you decide to do this for three years with us, you will never get screwed, you will never be surprised to the negative. And so that was another sort of big balance up point.

And then every mattress company was like, oh wow, these guys are doing it, hope we should do the same thing. Because fast forwarding a bit, a lot of people hear a firm today and one quick thing they think is, buy now, pay later. What does this not fully encapsulate or what does this miss about the company today? In its entirety. One of the things in the email thread that I just re-read in the origin origin of Affirm was your claim that advertising and payments are converging.

Which was the thesis behind trial pay and thesis behind a bunch of different things. Actually a failed project within PayPal called PayPal Shops where I was trying to, personally I built this thing where I was trying to remarket what you just bought, that information should give us a sense for what you might buy next and all that. And all these things by the way have turned out to be true. Like they're just 15, 20, 30 years before their time. But Affirm more and more today is means for merchants to not just satisfy demand.

So somebody comes in and says, look I'm gonna buy this thing but my budget is this big, I want the shoes, I'd love a bag. But Affirm is there to say, actually we can help you finance this on transparent terms. A lot of times, a third of the time on no interest at all. You should get the bag and the shoes, it's safe. But that is the brand promise of Affirm and it works tremendously. Well, tens of billions of dollars a year worth. That is now becoming more and more about creating a platform for merchants to tell their shoppers, hey, we are launching a new thing, we have this desirable new product we are selling.

We've transacted with roughly 50 plus million Americans alone and we're now live in four countries and growing pretty quickly. And so we have shifted to just satisfying demand to helping merchants create or guarantee demand. And the convergence of payments and advertising is upon us, although 15 years later is how it's predicted. You're too early to try to pay, I'm certainly too early to pay. Well, the other side, I remember, maybe it was in that thread or not, I should find that email. But most companies have very high customer acquisition cost.

And like the joke that I make now as a VC is that when I see 90% of consumer companies, I'd kind of rather buy Google stock or Facebook stock because that's where all the customers come from. And I distinctly remember many, many conversations around how do we have negative CAC? Right, and I think, to me, one of the coolest things about a firm to this day is it has negative customer acquisition cost. I mean, it is paid to acquire. or a customer, and there are some companies that you would think they look like this, but they don't because they're basically white labeling something.

But going back to the mattresses, let's just say that you're late on one of your payments. Casper, the friendly white ghost mattress company, because Casper's a white ghost, doesn't want to get, or a cat or something, we had a cat named Casper when I was growing up, they don't want to send you a dunning notice saying like, get off my mattress, you haven't paid me back. They want that to be a third party. Versus other companies where, yes, what is the difference, B2B2C is a really interesting category, it's business to business to consumer.

I tried doing this with trial pay, because it's like, all right, you're playing a Zynga game, go get free coins if you signed up for Geico, powered by trial pay. Nobody knew who the hell trial pay was, so we actually tried sending messages to consumers, like, you know, we got, I think we got blocked by our email provider, because nobody, we actually had a right to email these customers, but they had no idea who we were. Whereas actually having a financial relationship with 50 million people with negative CAC, I mean, it really is, not to pat you on the back, it's incredible, like, it's so hard to do.

We never see anything like this in the venture land, or, you know, like, because, yeah, there are lots of companies that can get to scale, I mean, Casper got to scale, but lots of CAC, and that's why a lot of these business models, they tend to degrade over time, because all the economic value goes to Google and Facebook. So if you can have negative CAC, and then you can launch other products as well, right? Because you actually own the customer, and most important, the brand wants you owning the customer.

Like, this is what I didn't get right at TravelPid, like, Zynga didn't want me owning their customer, Netflix didn't want me owning their customer, I just connected the two, but Affirm, the merchants actually want you to own the customer. It's very, you already have a burden of supporting a product and dealing with unsatisfied customers, or technical support issues, or whatever, a company that tells you 12 times for a year-long loan, or 39 times, you know, if it's a three and a half year loan, hey, you're late, or you gotta make a payment, or all the communications that come from your lender is just another type of burden that the brand is kinda like, hey, you know what, Affirm, you guys can handle this, this'll be great.

And so, the relationship with the customer is a key component of the foundation of Affirm, and our ability to develop new products is really built on that. One of the other really sort of cool, to sort of going all the way back to the credit scoring thing, so we are the only company in the industry, really, there's plenty of competitors who are trying, and some degree of success, but mostly not really, who will go into longer-term loans. And by longer, I don't mean mortgage for 15 plus years, but three and a half years is a fairly long period of time relative to the average buy-now-pay-later, which is like six weeks.

And to do that, you have to underwrite. You can't shortcut the, I'll just look at your FICO score, or I'll just sort of, you know, I'll look at your Facebook friends, like none of that works. You actually have to do a real, very sophisticated degree of machine learning work. There are two cool things about it. It's hard to do, and so it's just hard to compete with us on that front, but the other side is it gives you 39 shots on goal or 12 shots on goal to tell this consumer as you send them billing notices and as you have conversations with them, as they pay you to upsell them on a new service.

And so, these long-term loans, which are actually quite hard to manage and quite hard to maintain, reasonable default rates and delinquency rates on is the price you pay as a provider of financing to upsell consumers on more and more services. And so, our business has been fueled by these long-term products, even though they are probably the most challenging part of payments. Yeah, we've been covering a firm history while we're doing a historical episode. I also want to ask a question about PayPal history. Okay. Sorry. I'll keep it straight.

There's been, a book came out, Founders recently, there's been sort of a lot of revisiting of PayPal, both in its impact as a pioneer in payments and fintech, but also in terms of the people that came out of it and sort of the, it feels like the highest concentration of or highest hit rate of people who would go on to do incredible things. The only comparable thing I could think of today is not even a company, it's the Teal Fellowship or something. I can't think of a comparable company to PayPal today.

I'm curious if you can. And all with, or many with extremely strong strengths, but also very strong personalities and some strong weaknesses as well. And so, I'm curious if there's something you think that goes, as that story is being told, of its impact in the industry, how early it was, and sort of the people that came out of it, if there's something that's underappreciated or you think not quite grokked or any additions you want to make to how we perceive that company in that time period.

Sort of an old answer and a new answer to this question, which you might expect I've answered once or twice before. So, the canonical answer that I gave in a book, which by the way is very good. So, there have been many, much ink spilled on the history of PayPal. The founder's actually a very good book. Like, and part of it is because it's well written, but part of it is because Jimmy, the guy who wrote it, interviewed every single character in the story extensively. He spent a lot of time talking to all of us.

And so, it's at the very least grounded, in fact. So, the answer I gave for his research was, as we were interviewing, certainly the early team and even the later team, we kept on selecting for entrepreneurs. We would ask, like one of the stock questions I would ask, we would say, what are you gonna do after PayPal? And the answer that I welcomed the most was, oh, this is the last one. I'm gonna start my own afterwards. So, this, and they're just like, oh, if you're coming in here to kind of get your final graduate degree in entrepreneurship, of course we'd love to have you.

Like, you know, if you're that smart, you're that good, and you're that ambitious, you know, please come along. And so, that's kind of the why I think so many people that came out of PayPal proceeded to start companies. And like, literally, you know, hot second after, we all dispersed YouTube, and Yelp, and Peter's first major fund, and LinkedIn. And so, that was not an accident. We were attracting entrepreneurs by design. The other part that I think has actually not been talked about, and maybe subtle, and sort of my own cut of this, but we knew the people in the room intimately.

Like, we literally, you know, exchanged pheromones by hanging out in these sweaty rooms, brainstorming, and screaming at each other, and spending a lot of time as whiteboards, and declaring the other person's ideas idiotic, and so on. We knew the true version of that person, the sort of the extremely, I don't know, the ultra-philosophical Peter Thiel, and the, you know, go-all-in-every-time Elon Musk, and the, these days, gubernatorial David Sachs, you know, all these people, like, they all have kind of the presentation layer, and then there's, like, the, kind of, what they are like at a dinner party, and then, like, what are they when they're really stressed out, and just, like, head hurts, don't know what to do, and we have a real problem we need to solve.

So, if you know kind of the true base version of who we all are, as you watch one of these people go for something really big, like Elon taking humanity to Mars, which is extraordinary, and yet, you know, he's a human being. Like, I've seen him sweaty, and tired, and grumpy at the company kitchen 1,000 times, or Peter, when he is in his doubt, and calls me from some fundraising trip, saying, I think we might run out of money. You know they're not gods, you know they're humans, and that inspires you to actually, like, you know what, if that guy, who is just a normal dude, I spent a lot of time with him, he's normal, and he's going for this big of an idea, I should do the same, I should strive too.

And so, I think the combination of, like, these are very ambitious people, they all wanted to be entrepreneurs, but we also knew each other when we were young, and filled with doubt, which many, I think, still are. I think that that sort of gives you confidence that it's not all, you know, bed of roses, it's paved with bumpy parts. Is there anything we didn't get to that you wanna make sure that we get to while we're here? I'm probably less optimistic about agentic shopping, and I'm very optimistic about agentic payments.

Yeah. I think the notion of robots will buy our Friday night outfits is misguided. We want to know what we look like long before the robot delivers it to your door, but I think the friction, this notion of, so the best user interface ever created is the credit card, this may actually be finally up for renegotiation, because agents are, in fact, smarter than pieces of plastic, and even pieces of plastic with rewritable chips. And so, I think that's a lot of conversation around agentic commerce is probably looking 20 degrees off, where, you know, how can I convince you to just trust a robot to buy the right thing?

Like, I don't want to, like, I love buying bike parts, and I spend way too much time looking at two slightly different bike parts, because I just, you know, have to participate in the purchase. But the part where, like, okay, that one, the end, right now, that one is the beginning where you pull out your wallet, or you, and so, I think we're about to see some interesting innovation in payments user interface. So, can I ask a question on that? So, I mostly agree, but I think it depends on, there's like the, I need to research what to buy, where AI is just yet another tool at your disposal.

Like, you ask your friends, that's one, it's the bike part, I'm asking you, I'm also asking AI, I'm asking both of you, but now I have the SKU that I want, and I don't actually care which one of the 19 different places that I get it from, and I kind of think about the world, like, you have people that care more about money than time, you have people that care more about time than money. The way that I kind of think about agentic commerce, again, none of this is happening right now, so it's trying to prognosticate what's going to happen, is here is the UPC, here is the SKU that I actually want, go buy it for me at the lowest cost.

It's kind of the thing that people that have a lot more time than money do right now anyway, and like, that's why I always point to Camel, Camel, Camel, it's like my favorite example of this, because it's one of the top 100 websites in America that probably nobody with over $100,000 a year of income has ever heard of, but it's a way of saying, I know what I want, you're not helping, the robot's not telling me what bike part to buy, telling me what soda to buy, I know, just give it to me at the lowest price, but it takes way too much time, and sometimes you can just observe one group of humans that already have something as table stakes, and then porting that to another group of humans.

I guess, do you think that that form of agentic commerce, because it's actually related to agentic payments as well. I have nine different cards, I don't remember the Amazon Chase, issued by Chase, card terms for a non-restaurant purchase of something under $500, and if I've used my whatever, it's the same thing, like people that care more about money than time, they will figure that out, people that care more about time than money, they probably won't. I think directionally, that'll make sense. I think the adoption curve is probably going to be fairly slow, and these things are always sort of easy to predict to be very slow, and sort of the usual adage as well, slowly at first, not all at once.

I'm not sure I can handicap that well. The reason people don't, so there's plenty of people who live in this somewhat murky world of sometimes I have money, sometimes I have time, and it has more to do than anything with the price and the degree of care you put into it. I don't always just buy bike parts. Sometimes I buy milk, and milk is built. So I don't buy bike parts. I mostly buy bike parts. But I think what really happens for a lot of people, it's sort of like, well, yeah, I'll find the cheapest price, but then I'm not sure how to handicap the probability of that item actually showing up at my door at the time of my desired outcome.

Even if I'm okay buying some bike part from some faraway land, and it's sold from three different retailers, I'm gonna go with the most reputable one, versus the one that's promising me the cheapest, or sometimes the fastest. I think we're probably a few steps away from AI grasping both the preferences, but also handicapping how these preferences would change based on the input, which, in fact, maybe AI's already there. It's just that you haven't yet trusted your agent to do as good a job as you would.

Like, you can eyeball a site and be like, that looks like it was edited in the 90s. I'm just not quite sure that I'm gonna spend $1,000 for a new cassette, or $600, whatever the cassettes go for these days, in bike cassettes, the rear assembly. Well, Bike Closet is the site that I'm thinking about that looks like it was made in the 1990s that will sell you a very cheap cassette. Exactly, but do you know if it were used by someone? You know, teeth rubbed off and all.

Yeah, so I think we are in a difficult period where we're still figuring out. The other thing that, just to maybe end on a positive note, the thing that's really true that no one seems to be celebrating just yet is grocery shopping is 100% agentic. Like, for anyone who uses Instacart. Like, you tell your Instacart shopper, go bring me milk. You don't think twice when they say, oh yeah, you wanted whole milk from Organic Valley, but I found some other brand, and here it is. And you're like, oh, of course, like, it makes sense.

And so the AI inside that shopper's head is maybe not even as good as Gemini or Cache V is these days. But do you just go through Trader Joe's or Whole Foods or whatever you're shopping, and you figure it out, and it shows up, and 99.9% of the time, it is exactly what you wanted or better. And so we are already conditioned to allow some of these purchases to be fully outsourced. Payments and shopping and everything else. And so we're definitely gonna get there. It's just a question of how soon will we work through the quirks of like, well, returning this wrongly purchased by part or waiting for it forever and not really knowing what's gonna happen to it because it's in transit, but we don't know where it is.

I think that may be protracted. I mean, more protracted, people think. That's a good note to end on. We'll have to get a part two at some point when we talk more about the future. Max, Alex, thanks so much for coming on the podcast. Thank you very much. Thanks for listening to this episode of the A16Z podcast. If you liked this episode, be sure to like, comment, subscribe, leave us a rating or review, and share it with your friends and family. For more episodes, go to YouTube, Apple Podcasts, and Spotify.

Follow us on X at A16Z and subscribe to our sub stack at a16z.substack.com. Thanks again for listening and I'll see you in the next episode. As a reminder, the content here is for informational purposes only, should not be taken as legal business, tax, or investment advice, or be used to evaluate any investment or security, and is not directed at any investors or potential investors in any A16Z fund. Please note that A16Z and its affiliates may also maintain investments in the companies discussed in this podcast. For more details, including a link to our investments, please see A16Z.com forward slash disclosures.