How FTX CEO Sam Bankman-Fried navigates crypto trading’s new normal

Transcript:

Lau: Imagine, and you can create — once the stuff of bankers now in the hands of the individual. That’s what’s happening in the digital world of blockchain and crypto. Decentralization of an asset is truly just imagination away.

Welcome to Word on the Block, the series that takes a deeper dive into blockchain and the emerging technologies that shape our world at the intersection of business, politics and economy. It’s what we cover right here on Forkast.News. I’m Editor-in-Chief Angie Lau.

There’s a Japanese proverb that goes like this: fall down seven times, stand up eight. Throughout its existence, Bitcoin has crashed, but it has always come back. Will it do it again?

One thing’s for certain: The cryptocurrency market is no longer just about Bitcoin. In fact, it’s not even about altcoins anymore. From meme coins to stablecoins, DeFi and NFTs, the cryptocurrency industry has been completely revolutionized by smart contracts. And in times of market volatility, now investors have the option to trade tokenized stocks or commodities like lumber. 

Who else can better explain all these creative financial instruments in the crypto industry today? In fact, in many cases, he’s leading it. The man that pioneered it all, FTX CEO Sam Bankman-Fried, who also goes by SBF, joining us on the show today. Sam, welcome to Word on the Block.

Sam Bankman-Fried: Thanks for having me.

Lau: All right, quick question, Sam, before we start. Did you sleep at the office again last night?

Bankman-Fried: It’s a little bit of a definitional issue. I guess to the extent that there’s an answer to it, the answer is yes. So it’s been a busy few days.

Lau: You could say that again, the busy few days have seen a trillion dollars worth of market cap wiped off. We’re seeing a little bit of recovery. I guess a lot of your team probably also slept at the office.

Bankman-Fried: Yeah, certainly the developers. I think it was sort of a late night. A little over a year ago, the March 12th incident. It was way smaller in notional terms although moderately larger in percentage terms on that crash, but it was way more brutal for us and for the industry in general. I think the industry because just a lot of businesses were maybe bankrupt at the bottom briefly. For us, because our systems are just not as battle-tested and they broke a lot more, so we had two weeks of just nothing but customer support backlog from that day. It was not an easy period. 

This was a busy last few days, we had a day of backlog, not a month, so it was busy, but it wasn’t anything like what we saw last year in terms of the fires we had to put out.

Lau: Battle-tested to be sure, but a lot of other exchanges potentially not so lucky, and certainly the people who are trying to get access to the accounts and make trades suddenly didn’t get access. Where are we in the industry? That’s got to be a black eye. Across the industry, truly a black eye, even if you at FTX just had a little speed bump.

Bankman-Fried: It’s one of these things where when an eye gets blackened enough times, you sort of punch it again. It’s almost not an update for a lot of people, because this is really busy day, number 17 in a row that half the exchanges crashed, and it’s the same exchanges every time — it’s not super random. If you know how much volume is getting traded on a day, you can have a pretty good guess at which exchanges are going to make it and which ones are going to be down for a third of the day. But it’s not a good look. 

This is something I didn’t fully realize before starting FTX. I definitely had a conversation with founding members and developers when we were building FTX, and I think what I said was, “Look guys, if we can get 95% uptime, that’s pretty good. Obviously, 100 is better, but 95, that’s not the thing to focus on anymore — that’s pretty close to 100, and we should focus on doing other things, maybe come back later.” That was not the right way to look at it. If you have 95% uptime that means you’re spending 5% of your day getting death threats, and that’s a lot to spend. Especially when there’s leverage involved but just in general, customers care really deeply about being able to access their accounts during the most volatile times. I definitely underestimated that when we started FTX. 

That was something that over the last year we had to circle back on in revising our plan. We got up to 99.99, but even that’s not enough. There are two more nines we needed, and it’s just over the last six months or so that we finally rolled out a set of features that mean that there is basically no downtime anymore on FTX.

Lau: That’s critical because in the most volatile moments, that’s exactly when people need to be guaranteed access. Do you think that this is an industry problem that we’ll continue to experience? Or do you think people are learning their lessons in the same way that you recognized [that] you got to get to 100.

Bankman-Fried: People are getting better, but really fucking slowly. Every day it gets one basis point better. In five years, I think the downtime situations will look better. Next time there’s a busy day, Coinbase is going down again — it just happens all the time. People are aware of this; they’re working on it. They’re not working on it at the pace that I think we sometimes feel like it makes sense to work on things. I think they’re working on it at a pace that seems fast from a large ossified company standpoint, but I do think people are aware and trying to make progress.

When something happens enough times in a row, you lose the explanation of oh, they just weren’t aware. At some point, obviously, they’re aware that it’s going to happen next time. So the question is, “Why isn’t it fixed? Why does it keep crashing then?” A lot of this goes to is: what is actually the expertise of these companies. Is the expertise building complicated technical systems? That’s not always the answer. Some of them have unbelievably valuable strengths, that doesn’t have to be one of them, but that’s not the only thing that matters. Another thing that’s worth noting here: is it actually hard to run a crypto exchange? I don’t think it’s the hardest business in the world, but it’s a more complex business than I think a lot of people would think of an exchange as being, because it’s not just a matching engine. If your goal is just to give a matching engine online, it’s substantially easier.

The really tricky thing is the integration with the risk engine, the liquidation engine that deposits the withdrawals, the transfers. The problem is that all of those are co-dependent on each other. If you fiddle with any parameter, any one of those, you have to recompute every other part of the person’s account. You run into this thing where between every single fucking tic that you process, because every tic changes the marked pricing of Bitcoin by a bit, and all of a sudden, who knows who’s going to liquidate — you have to recompute the world. It’s easy to get in these situations where you’re basically having to recompute every single part of every risk check in every deposit withdrawal check after every single thing that you process. And that’s a way more intense problem than just matching orders. I think that’s why you often see these throughput issues. It’s not easy to build up capacity; you can’t just buy more computers because you can’t paralyze it. It’s all interdependent.

Lau: It really means that those developers are staying up overnight along with you trying to keep up with these recalculations and these reformulations and program that to that level of efficiency.

Bankman-Fried: Yeah, there’s a lot of that. There’s also a lot you have to do ahead of time. If you actually want zero downtime, you can’t wait until it breaks and fix it. It has to not break in the first place. Some things can break. If withdrawals are down for 30 seconds, no one cares that much. People’s expectations are that it takes 15 minutes to do a withdrawal on average. That’s the sort of thing where if you have various road bumps, that’s okay and you can be fire-fighting that in real-time. You don’t have hours to spare but you have minutes. But when it comes to the matching engine, you have maybe seconds if you’re lucky. You do have to be there patching in real-time, but you’re really sad. You get sad when your customers are sad. You really need to have gotten it to the point where it doesn’t need patching in real-time.

Lau: So you’re saying that this was a stress test moment and you learned that lesson a year ago. But so did a lot of people. This is another stress test. The question is, it’s not how do you pass next time, it’s if your customers trust you enough to let you pass the next time. That’s really the question that the whole industry really needs to ask itself. I want to ask you, for all of the customers — the world is looking at crypto and there’s a lot of people that said, ‘I told you so.’ What do you think triggered this massive correction/meltdown/crash of Bitcoin?

Bankman-Fried: You’d think what triggered it is almost a boring question. Things went down a percent. Who knows why? Someone sneezed. The interesting thing is what happened to the next 29%? I think what happens is Bitcoin just wiggling around, there’s a little bit bigger wiggle than normal, and someone gets liquidated for US$500,000 position, and all of a sudden it’s down another 20 basis points and it liquidates someone else. Then you start to get this cascade, and that’s really what happened.

The bulk of what happened was once you got the first liquidation, it triggered the second, which triggered the third, which triggered the fourth, which triggered the fifth, and before you know it, US$20 billion has been liquidated. If you actually want to sell US$20 billion of crypto, yes, you’re going to crash markets if you do that in a day. That was the core thing that happened. For whatever reason, this process started and then once it started, it was self-sustaining. There’s this ratio that comes out, which is you can look at basically per percent that Bitcoin goes down, how many notional positions have new liquidations triggered. For how many was that where their maintenance margin threshold was? And then you can look at what’s the impact of that size of position. Is it more or less than a percent? And if it’s less than a percent, you have exponentially decaying liquidations that peter out. If it’s more than a percent, then each liquidation moves things down more than the previous one, and you have an exponentially growing chain of liquidations. It’s really on a knife’s edge threshold, and you go from 0.9 to 1.1, you go from yesterday being a 3% down day to being at 30% down day because there’s the other US$19 billion of positions that don’t have to get liquidated if you don’t drive things down that first 5%.

Lau: That scenario is very familiar to regular equity exchanges around the world, these flash crash moments. Then there are stop measures put in place where in a decentralized world when people are holding Bitcoin and cryptocurrency, was this a flash crash that we’re going to see recovery in? What could have been put into place? And also, did you see different behaviors from retail investors versus institutional and traditional investors, which is what we’re hearing from a number of sources?

Bankman-Fried: You can try and put speed bumps in, and I think it’s correct to have something there. Your goal is to catch misclicks basically. Your goal is to stop anything that wasn’t supposed to happen. If someone is just sitting there trying to sell, they’re going to sell eventually, you can’t stop them. The goal is just to catch errors, catch erroneous orders, and things like that. The thing though, that you get is that there are a lot of people who sold and didn’t want to sell yesterday or two days ago. They wanted to keep the position. The problem is that when things are down, you can try and stall, but it’s always super dangerous because what if things keep going down? You still need to liquidate the account, stalling didn’t get you anything except now the account might be bankrupt. And all of a sudden there’s a loss; someone’s gonna have to eat that negative number. It’s actually not super safe to try to wait too long before liquidating. You have to do it before the account would go under. That puts a lot of pressure on the exchanges. 

What you have in the end, I don’t think there is anything really fundamental that changed from three days ago today. I don’t think this is caused by the world changing its mind on crypto. I don’t think it’s really caused by economic fundamentals. I think it was 10% of that and 90% was just liquidations. So it was sort of an accident, but it was a huge accident, and it happened. What you end up with is this really, really weird situation where if you managed to turn off every exchange at once and just manually pick up all the markets and move them up 10% and set them down with absolutely thin volume, nothing there. As in, three, two, one, go, all of a sudden things recover immediately, because you no longer need to liquidate anyone. Technically, everyone’s above water again. The selling all of a sudden stops. The buying comes in and goes right back up. And the key thing is you stop yourself from having to process those sellers into liquidation permanently or for a while at least. You do get these weird effects where this is not the kind of thing where the initial conditions don’t matter. You figure out what everyone’s opinion is and then average them together and that’s how markets work. Some people are sometimes voting against their opinion because they have no choice, because that’s how liquidations work. I think that’s a really key part of crypto markets. It is true in every market, but it’s really true in crypto markets, way more true in crypto markets than in most markets.

Lau: There’s a psychological factor, there’s no doubt about it. Retail investors are sitting right next to whales and institutional dollars that really propped up the prices. Did you see them behave differently at FTX? 

Bankman-Fried: It’s weird. The words retail and institutional have this somewhat clearly defined meaning in equities. My uncle is retail and Fidelity is an institution, and no one’s going to get those confused. In crypto, a lot of the volume comes from people where it’s a little bit unclear. Two guys working together in a basement somewhere, combined age 40, combined wealth, 40 million, and they have a corporate entity. Is that an institution? Where the money came from, they bought the Ethereum ICO. Now they spent all day trading US$200 million a day in volume in crypto markets. It’s not really what you mean by retail, but it’s also not really an institution. It’s sort of a day trader blown up to epic proportions, but blown up in massive. Day traders don’t trade hundreds of millions of dollars in most markets. They trade thousands or tens of thousands of dollars, maybe a million, but US$500 million from a day trader. Yeah, that happens, it just does.

Lau: Yeah, it does. You’re totally right. That happens in crypto all the time.

Bankman-Fried: That’s one of those weird things. When you say what were different parties doing — I can tell you what the banks were doing — they’re doing fucking nothing because they don’t operate on that time scale. What you really mean by institution, like a highly regulated, customer-facing thing that has lots of funds, operate on a week time scale. They’re sort of like mostly sitting out in something like this, and they’re sort of like getting back into work today and be like, “do we want to consider putting a position on?” If you look at who is active during that day, you look at the HFT firms, you can kind of guess what they’re doing. They’re doing the normal thing. Probably had a good day. That’d be my guess. You can look at the people who put on directional bets and that’s a lot.

I think Three Arrows is a good example of this. How would you classify them? I think institution is probably the better word, but they have a really unique style of how they think about crypto. Just looking at how they tweet about things, I have no clue what they’re doing. But you look at that whole class of people and some of them probably just bought the dip for a massive size. I think some of them probably were the dip because they were getting liquidated, probably not the ones you’ve heard of. The US$20 billion had to be someone, right? It’s not all your uncle. They’re on both sides of that trade — they being this collective class of day traders extraordinaire and/or crypto-specific delta-taking three-man shops, it’s super interesting. The answer, in the end, is that it’s an interesting, messy marketplace. What position you had coming in has a really big impact on where you come out.

Lau: How do you navigate this? I mean, for somebody who just takes a look at this world and you’re in it every day. You’re the oracle that notices these trends and then you create products. But as you’re in it, how would you characterize the points in which people should think about and use as navigational points? 

Because you’re totally right. This is not an equity market, this is not a traditional-styled when we talk about fundamentals, it’s like comparing apples to rocketships, frankly. That liquidity coming into the market is also super important for growth, so how does one navigate smartly?

Bankman-Fried: It’s interesting. It’s dynamic. It’s really dynamic. It turns on a dime. When we think about what to build, what products to be building here and how to design them, one unfair superpower that we have that ICE doesn’t have, CME doesn’t have, is our customers are tweeting at us all day telling us what they want. That’s just one example, but boy are they vocal and we appreciate that. We can often get a pretty good sign of what people would want just by listening to what they volunteer. And so I think that that’s one piece of this, and you will sometimes just hear a copy of random stuff and have no idea what to do.

Someone will be really pushing you there like, “you need the MATIC/BTC market. That’s what your exchange needs.” And we’re like, “Wouldn’t be bad to take some of matching engine capacity, probably at a few million dollars a day volume on the fence.” Someone will feel really strongly about that and occasionally they’re right. Occasionally, it’s like, “Oh God, that’s a market we need.” Clearly, that was actually what everyone was waiting for. That’s really our job: is above and beyond everything else identify that, identify the thing that everyone actually really fucking wants, even if they haven’t quite figured out how to ask for it yet. I think Trump futures was one good example of that. No one asked for that, we just built it and it had no traction for six months. Then election got near and all of a sudden, like traffic nearly just blew up on FTX just then.

A lot of this is trying to figure out what do we think people will want, what fits what they’re saying. And how do we take the feedback that we’re getting and turn it into something actionable? Sometimes we’ll get feedback that just seems like it makes no sense. They’ll be like “I really want to be able to do this type of like…” and that trade is nonsensical — you’re literally just burning fees. Like, sure, but why would you want to do that? We’ll go back and forth for 20 minutes, and it’ll be super frustrating. What’s actually going on is not that they want to do that trade. What’s actually going on is they don’t like using the product. They come away unhappy and they don’t know why. It wasn’t fun. It wasn’t clean, something was bad, something’s irking them, and they don’t know where to place it. They end up pointing at a random thing, it’s tangentially maybe related, not actually the thing that is really core bothering them, but there was something that they picked up on implicitly that was bothering them. A lot of our job is to pick that apart and figure out what that factor really is, figure out what actually is the thing that we do here. 

One of the more powerful examples of this was the sub-account switching bar. A lot of people really complained about not having a particular type of isolated margin on FTX. And we’re like, “Yeah, we have it.” And we just went back and forth and we were very confused. Eventually what we realized was it took four clicks to switch sub-accounts, and that’s too many fucking clicks. Yes, you could isolate your positions if you’re willing to click four times to switch which position you’re accessing, but no one wants to do that. And so in practice, that person is asking for something else that would’ve made no sense because the way we’ve done it also wasn’t any good. The actual solution there was just to put a bar on every trading page, and it just lists your sub-accounts as a button for each one. 

Lau: And they were happy.

Bankman-Fried: Yeah. All of a sudden they were like, ‘Oh yeah, this is all solved. This is great now.’ We’re like, ‘That’s not what you asked for.’ And they’re like, ‘I’m happy now.’

Lau: That is major EQ, Sam.

Bankman-Fried: It took us a long time. We had a lot of really consternating conversations before we figured out what was going on there.

Lau: And you step back and you went, ‘What are they really trying to say?’ So you’re part couch therapist, part solution oracle, part translator — because what they’re saying is not what they actually mean. Then you come up with the solution that answers the thing that they were trying to communicate but didn’t know how. Is that how you came up with lumber futures? Where did this come from?

Bankman-Fried: Honestly, that one was a dead obvious one from some perspective. We just looked on Twitter, and half our Twitter feed was people asking for lumber futures. There just came a moment where the world all got together and like, “fucking lumber, that’s what we’re trading today.” And everyone was asking explicitly for lumber futures, and we’re like, “Alright, fine. We’ll do lumber futures. That’s what you want. We hear you guys.” 

Lau: West Coast of North America is super happy right now.

Bankman-Fried: One thing we’re saying, it’s not that people are particularly bad in crypto to understanding what they want. I think people often understand exactly what they want. But like everywhere, they often do and sometimes don’t. But when they do, those are the things everyone does. Everyone has realized that they want a Bitcoin future. Everyone wants to trade those so exchanges list it. We’re listening to people saying, ‘We want Bitcoin futures.’ Yes, obviously, we have Bitcoin futures. There are the things that you’re already doing, everyone’s already doing, which are the things that people know how to communicate. Often the reason it’s cutting edge instead of the discovered territory is because it’s one of those things that is hard to put your finger on.

Lau: To navigate the space feels almost too much if you’re not familiar with it. How do you wade in? Do you wade in or do you just kind of dive in head-first? What’s the right analogy?

Bankman-Fried: You can wade in; I don’t think there’s reason to. I don’t think you’re buying yourself anything by taking it slowly. I think you’re just going to have a mediocre experience. You’re not going to have as much context as you want. Honestly, you’re just prolonging that bad period — you’re just drying out the period of not knowing enough, and you’re not doing anyone a service. You don’t have to dive in. A lot of people have great lives that doesn’t involve knowing what the exchange owned by Paxos is called. That’s totally fine.  I think Biden’s got a lot on his plate. He’s probably doesn’t know itBit, but it’s okay. Certainly doesn’t know FTX, that’s fine with me. He’s got some countries to worry about. If you do want to get into crypto, if that’s what you decide on, there’s no point in half-assing it.

Lau: Dive in and get ready. Meme coins. Alt Coins. Is this a good thing? Is this a bad thing? Is it polluting the sanctity, the purity of crypto, or is it kicking wide the doors and telling people to come on in — water’s just fine, let’s have some fun and gamify a few trades. What do you think?

Bankman-Fried: It’s got a little bit of a bunch of those. Dogecoin was amazing for recruitment for crypto. It got an enormous amount of interest in crypto. It also got a lot of disgust in crypto. It is a double-edged sword. That came at a cost. I think in Dogecoin’s case, it was worth the cost. I think for some other meme coins it wasn’t because they only really had one edge. Most these coins aren’t getting anyone in, they’re just pissing people off, but some of them are and some of them are great coins that are making people happy.

I think the bigger thing is that this is not really crypto things, this is just a world thing. This happening in a lot of industries right now. You look at what’s happening in the stock market like GameStop, I mean, that’s a cryptocurrency. It’s not called one, I understand technically, it’s an equity, but like, come on. That’s a shit coin. We all recognize it, right? I don’t if it’s what the Pandemic did to all of us or if it’s what we’ve always secretly wanted or if it’s what social media inevitably results in. 

This very particular storm is one which stretches across industries. It’s not even analogous. It’s not like both crypto and stocks have their own charismatic guy who when they tweet, the thing goes up — it’s the same guy. It’s Elon Musk in both cases. It’s been a weird year for everyone. I think the big question is what’s next year like? Is this the new normal or is this what happened when everyone is bored, stuck at home, socially isolated? I think that is certainly a piece of what drove it.

Lau: I think the GameStop phenomenon, you’re absolutely right about it. Everybody in crypto, everybody in the industry and blockchain has been in this for a couple of years, recognized the characteristics. It just happened to happen in the equity space. 

But when the crowd moved, all the players who hadn’t participated in crypto, had no idea that there was a crowd functionality to trading, suddenly saw very specific calls flooded in with a certain position, and people lost their shirt and people also made a ton of money. The funny thing was it just wasn’t the regular people who were making a ton of money. 

Bankman-Fried: Some of the regular people made a killing, but some of the regular people lost a killing and it’s just chaotic. One of the big things here, one of the new things is, if you look 30 years ago and you ask what could really drive a stock to go up an enormous amount. If the mutual funds got together and all decided it was a blue-chip stock that could do it. They collectively buy US$30 billion of it and it’d go up a lot. If a few retail traders did it — bought a million dollars combined, it’s whatever then they would sell it like at a slightly good price that institutions make a little bit mind, that would be that and the stock would go back down the next day. To the extent, retail was moving markets only in reaction to things that centralized media services were producing. 

And then social media came around, whether it’s Reddit or Twitter or 4chan or whatever your favorite vice is. All of a sudden for the first time, you had 15 million day traders get together and talk to each other. You can’t have 15 million friends get together and get lunch and then go pull out their app and trade. You can’t fit 15 million people in a restaurant, especially during Covid, but you can fit 15 million people on the same forum. All of a sudden they each have a thousand dollars to trade. Now they have US$15 billion between them, and they’re all talking and they all get together and decide that they’re going to buy GameStop. Now you have a US$15 billion buy order coming in collectively in GameStop. That’s big enough to blow through the institutions that are just like, ‘This is 10% rich, I’m going to sell.’

Lau: But there are rules that prevent those mutual funds to talk to each other, for banks and different firms to collude with each other. There are rules, and that’s what the SEC is for. And so that is a very highly policed and regulated market to protect the little people. But what happens when it’s the little people? You and me. I’m not part of a huge firm with huge capital. I’ve got one hundred bucks, a thousand bucks. That’s a lot of money to me, but now I have 15 million of my closest friends and we’re in on it together. By any happenstance that is also colluding. But there is in the world in which we lie.

Bankman-Fried: It was never extremely clear exactly what the thing was that was and wasn’t allowed. For instance, it’s totally fine for a bunch of people to get together and have a factual debate about a bunch of companies and just say a lot of true statements with the goal of figuring out truth. That’s okay.  Maybe after you do that, everyone learns the truth and without trying to collude, everyone just comes up with the same impressions. These arguments were great, these didn’t hold up. Everyone just comes away and like 90% of people come away pretty bullish or pretty bearish. It’s not exactly collusion, that wasn’t the goal of it, but it is an activity that correlates their behavior. 

You get into this really tricky thing where really what the SEC is looking at here is is intent. A core piece of this is was, the intent to do research and figure out the truth about the world or was the intent to make sure you do what other people are doing, even if it has no relationship to the underlying products because you will all win to each other’s impact. But of course, the trading sometimes looks the same between those. Why were you doing it? With institutions for decades, you’ve seen the first. They don’t necessarily go and collude, but someone will write a research report on some company to lay out a bunch of facts about it.

Lau: It’s going to be in the newspaper or the radio in the olden days.

Bankman-Fried: All of the mutual fund managers read the same research report, and they see the same maybe legitimate set of bullish facts about some company, and they get to work and a third of all mutual funds decide to buy that stock that day.  It wasn’t trying to collude, but they’re acting off of common information and so, not surprisingly, they make correlated decisions. And you really see this. On Wall Street, I never really heard people getting together to collude to do a trade in a manipulative way. Despite that, we would wake up some days and the whole fucking world would be buying some ETF. Just like everyone. What’s going on? How can these seven brokers all have decided today that this fund that they never talked about before, they’re going to buy this? But there is real correlation. 

Lau: They probably saw somebody’s giant position and if that person was respected in the industry, somebody else followed on, somebody else went I can’t miss that then followed on, then half the market’s in it, so everybody else has to follow

Bankman-Fried: Exactly, Warren Buffett discloses in some form regulatory filing that he got really long on some company.

Then you look at the retail side of this and it’s just all over the place. Was it attempted manipulation or was it honest, wholesome communication of facts to try and get to the base truth of economic fundamentals? And the answer is there are 15 million people. They weren’t all doing the same thing. Ten thousand of them were literally colluding in obviously illegal market manipulation sense. Fifty thousand of them were all getting side-tracked on the same red herring, trying to figure out the truth, but it was irrelevant. One hundred thousand figured out the same moderately relevant truth, but the whole world already knew it. One hundred thousand discovered on something moderately new. In the end, maybe they all make the same decision. They’ll hash it out, and they see which way the wind is blowing, but the thing that caused that correlation was not one thing. It was a complete mess of things that really spans that spectrum.

Lau: But, Sam, you and FTX and the industry, you represent what is also part of this new normal. You’re transforming access to some seriously specific plays once reserved for specialized brokers and traders, to moms and pops, to the retail investor to your uncle, to me. Is that dangerous? What is the new normal?

Bankman-Fried: I don’t think anyone knows what the new normal is going to be. We’re all just trying to be no more than a few years behind the times if we can get away with it. I may have to get a Snapchat soon, because I heard that was a big thing in 2010. Being a little facetious, I spent a lot of time trying to get a sense of what’s going on in

Lau: Honestly, there’s probably a million people right now starting new accounts if they haven’t already on Snapchat just because you said so. That’s the thing. That’s what’s happening right now.

Bankman-Fried: I’m glad they’re paying me enough for that. They’re not — to be clear. They’re not. I’ve never talked to Snapchat in my life. 

Lau: We’re joking.

Bankman-Fried: I think it’s really high upside. I think it’s high potential. That doesn’t mean every step is going to be the right step. There are going to be some missteps there. I don’t think anyone is ever going to succeed or should even think of it as a possibility to try and succeed at playing that perfectly. I think the goal that people should have in mind with this is trying to push in a net positive direction. Every step of the way, try to move things, try to make things a little bit better. You do something because you think it’ll be good. And if it’s good, do a bit more of it. And if it turns out it didn’t seem so good for the world, then take a step back. You’re not going to make every decision, that’s how it is, that’s okay. But just keep trying on the margin, probabilistically move in the good direction and over time that’s going to add up. Over time, the missteps will get weeded out and the right direction will be discovered. I think that’s the best that we can do here to the extent that any of us are stewarding what is sort of an unstewardable force.

Lau: But you got some pros on your side. Obviously, you come from experience, but you’ve got some strong bench strength as well. You recently hired from Citadel Securities, Brett Harrison. You also added an HSBC executive, Jonathan Cheesman, to head OTC and institutional sales. Where do you want to shape FTX to go? It sounds like you’re picking from the old world to create a bridge to the new world.

Bankman-Fried: Frankly, I’m super excited about that. There are other things as well, though you don’t necessarily hear in the newspapers about A+ memester hired to make some really sick Twitter posts. They’re fucking sick. Blockfolio’s Twitter posts are crazy.

I think fundamentally what I’d say is one of the really biggest fundamental driving philosophies of FTX from the beginning was maximize long-term expected value and that probably most of the expected values in the upside cases like the far upside cases, which means that you’re trying to maximize your odds of getting one of those real tail cases. And if you’re doing that, you can’t restrict the business too much in terms of what it’s trying to do. If you shoehorn yourself into the exchange that tries to serve institutions, the exchange that caters to low engagement retail, the Chinese exchange, the Korean exchange, the Australian exchange, the German exchange, if any one of those is what you see yourself as, it’s what the world will see you as. It’s what you will do. You may do great at that demographic, and it’s actually a really effective way to get moderately big. It is way easier, moderately big by being super focused on one demographic so you can just make every decision with them in mind. The problem is that you kind of hit a wall if you’re not careful. You just got half of Germany, but what’s next? It’s a really good path to becoming the 30th biggest exchange, but it’s not clear what the path is from there to get to number five.

We started from the beginning, obviously, picked our battles, really did come at it from the perspective of, in the end, we want to try and be huge, and we want to try and be the best product for as many people as we can be. What that means is that we don’t want to neglect any of them and certainly not permanently. So we’re pushing on a lot of fronts. We’re doing a big institutional push, we’re doing a big retail push, we’re doing separate product-related pushes for the power users that have made up a lot of our volume so far. I frankly think the upside of those three demographics are not that dissimilar from each other. I think they’re all within the order of magnitude of each other in terms of how big they could get. I’d be pretty sad if we just lobbed off a third of our upside arbitrarily at the beginning. We’re pushing on all of them is the real answer.

Lau: I think there’s a universal truth to that, there is a common thread and that common thread is everybody’s new to this space, and everybody wants to get pretty serious about it. What’s next? 

I want to talk about Serum and DeFi and decentralized exchanges. Is this what’s next? You’re a centralized exchange, for lack of a better word, at the moment. Are you going to disrupt yourself, dis-intermediate yourself? And what is Serum?

Bankman-Fried: In the end, it’s the electron’s world and more power to them. Serum was originally the answer to a question. And that question is: ignore 90% of all outcomes. Only focus on the 10% of outcomes where DeFi is the most successful. The 10% of outcomes where it’s at least 50 times bigger than it is today. And in many cases, thousands of times bigger. When you’re selecting for those, when you’re selecting for the cases, where DeFi gets huge, what does the world look like? In some sense, that’s a hard question to answer, it’s going the wrong way. Tell me what the world looks like and maybe I can predict how big DeFi will be, but how do you go backwards there? But I think you actually can deduce some things. The single fact is incredibly powerful. How many transactions per second does DeFi have in that world, in the world where it’s huge? You can kind of answer that question without knowing anything else. You can just look at what’s it mean to get huge — it’s got to be at least Facebook. If all of DeFi put together is smaller than Facebook, it didn’t get huge. Facebook’s got about a million transactions a second. Who knows what Facebook is, how about Twitter that’s about a million. About Visa at 50,000. You can look at a lot of different applications here. You can get a sense of what do I mean by getting huge? What would it look like to have that big of an ecosystem? I think the answer is 50,000 to 50 million transactions a second is the range that that spans. 

Now you know that. And now, you actually know a lot about what the world looks like. You all of a sudden know things about blockchain, about the future of blockchain if you get to presuppose that things get huge. You know that somehow there are blockchains that have probably a million transactions a second to be able to support that. Then you look at these blockchains and, oh boy, that’s not a million. And the problem is it’s not a factor among many factors. If you’re five orders of magnitude too slow, no other set of factors combined can make up for that. You literally can’t do it. You can’t run anything. You have amazing gooey, but there’s nothing on the gooey, there’s no program writing on it. So it’s just a hard constraint. Then we sort of said, all right, let’s try and build out some infrastructure for that. Let’s try and build out some infrastructure that would make sense as a first step towards a world with a million transactions per second on-chain. What does that mean? It means you can have an exchange, an actual exchange, the matching engine, not just an AMM, because you have the throughput and efficiency for both the decks on. It means when you’re looking at blockchain, you sure as hell better have a long-term roadmap to getting to somewhere in the ballpark of a million transactions a second. It’s about one blockchain ever built that does — Solana. Maybe someone will make a second one at some point that would be cool. It’s just like I guess we’re going on Solana; they’re the only ones trying to be big.

Lau: Not Ethereum, not anybody else. Everybody’s trying to be big.

Bankman-Fried: Well, they’re trying to be medium-sized. What does big mean? Take your pick on another blockchain and what’s their roadmap for transactions per second? When they say big, they mean a thousand. You mean getting to a thousand transactions a second, and that’s just not enough. That’s enough to run current DeFi. That’s what you can do with a thousand — current DeFi without incurring insane gas costs, you can’t do more than that. If you get up to a thousand, you can have one small to medium-sized company running on-chain, and that will take the entire blockchain, that’s what you get. Or you can have a bunch of tiny companies or a bunch of AMMs. You’re not going to have a billion-person consumer product, that’s going to overwhelm a thousand transactions per second immediately. When people say fast, they don’t think very hard about what number fast means. They say a thousand fast, but a thousand isn’t fast, a thousand is medium-slow. How many are trying to get actually fast, meaning hundreds of thousands, at least? It’s a pretty small list.

Lau: Next steps for you. We asked you when Coinbase did the direct listing, do we see an FTX direct listing some point in the future? And you even did the pre-IPO token on Coinbase, so what are your thoughts there for yourself? If you’re going to get that big, you’re going to need a lot more capital potentially. Would you go to the market for it?

Bankman-Fried: It’s an interesting question. We’d be crazy not to be doing due diligence on it. We don’t have concrete plans there, but we’re learning what we can and doing our homework. One nice thing about the business we’re in — and this applies to FTX but it applies to other exchanges as well, at least some other ones — is that we’re profitable. The expenses aren’t that big, all things considered. We don’t need to access public markets. We don’t need to get more capital. We can keep doing what we’re doing, and we keep trying to grow — we’re getting more capital just from the business. It’s not we’re on a timer and if we don’t list by then, we’re going to be dead in the water. That being said, that’s holding it to an insanely high bar. The real question isn’t do we need, but should we? I think there are uses of capital; we acquired Blockfolio last summer, [which] probably won’t be the last acquisition we do. We closed an acquisition this week.

Lau: Which was?

Bankman-Fried: Not public yet. I think not public yet. I’m pretty sure not public yet. 

Lau: I tried, I tried everybody. He’ll let me know when you’re ready.

Bankman-Fried: Those are big-ticket items, they can be super valuable. The other thing is trying to figure out what it did to Coinbase’s business. Ignore the capital, ignore the money. What did it do to their business? Did it grind them to a halt? Did it get them a huge number of counter praise? Did it get them a lot of trust? Did it give them a lot of ill-will?. Seeing that play out has been interesting. Even more generally, what we’re trying to get our heads around is what is the right business strategy for us in terms of building out a lot of the network and relationships and partners that are going to be really important for growing our business outside of the core-dedicated crypto ecosystem.

Lau: Because one day they might be the same thing. When what’s outside is inside and what’s inside is outside. It might just be what it is. We’re in the transformation, the metamorphosis stage right now, there’s no doubt. And Sam, if you’re a butterfly flying away. Thanks for letting us into your cocoon and sharing just how you see the world that you’re shaping right now. Before I let you go, I just had a thought. Can you pick something on your desk and show it to us? 

Bankman-Fried: Oh boy, so much crap around here. 

Lau: There are so many people who just see you like this. But what you did, I felt like we got to understand you a little bit better, and if you’re willing to share.

Bankman-Fried: I need to fiddle with things compulsively. I go a little bit stir crazy. Usually, it’s below the camera. You can’t really see my hands, which is intentional. If you see me in person, it’s incredibly rude. Sometimes I’ll be like talking to someone. I have a bunch of things on my desk. Recently, I’ve got some Magic cards here. I was kind of shuffle them while talking, and I think you see my shoulders moving sometimes, you can’t see my hands. That’s what’s going on. We played a little Magic internally, did a draft at the company.

Lau: Can you do one right now? What are your Magic cards? Do I have to pick one or something like that?

Bankman-Fried: The actual game is a bit longer, but right now it’s serving a secondary purpose, although maybe eventually. Its primary purpose is as something I can shuffle, and I think that can occupy my hands and free my mind from worrying about them.

Lau: Well, I think it is a perfect metaphor for how people regard you, which is part magician, part realist, but always moving. And we’ll leave it there.

Bankman-Fried: Thank you.

Lau: Sam Bankman-Fried, FTX CEO. Thanks for sharing with us here on Word on the Block.

Bankman-Fried: Thanks for having me.

Lau: And thank you, everyone, for joining us in this conversation. It was great to have everyone here. I’m Forkast.News Editor-in-Chief Angie Lau. Until the next time.