WEBVTT - Understanding the Collapse of Sam Bankman-Fried's Crypto Empire 0:00:10.000 --> 0:00:13.440 Hello, and welcome to another episode of the All Thoughts Podcast. 0:00:13.480 --> 0:00:18.279 I'm Tracy Alloway and I'm Joe wisn't Tracy? Obviously, the 0:00:18.280 --> 0:00:20.520 big story, which we haven't talked about yet, is the 0:00:20.560 --> 0:00:23.800 collapse of the crypto Exchange ft X and the broader empire, 0:00:23.920 --> 0:00:26.439 which was built by Sam Bankman Freed, which includes a 0:00:26.520 --> 0:00:30.600 trading shop alimeter research. That's right. It has been really 0:00:30.760 --> 0:00:34.440 a stunning two weeks for the industry, in part because 0:00:34.479 --> 0:00:38.120 Sam Bankman Freed, or SPF as he's frequently referred to, 0:00:38.720 --> 0:00:43.159 had become arguably the face of crypto. He was highly 0:00:43.200 --> 0:00:46.560 influential in d C, he was a donor to numerous 0:00:46.680 --> 0:00:51.280 nonprofits and politicians, wildly rich, and of course we have 0:00:51.479 --> 0:00:54.880 interviewed him a number of times on this very podcast, right, 0:00:55.000 --> 0:00:57.880 so the last time we interviewed him it had already 0:00:57.880 --> 0:01:00.800 become something of an infamous episode. We're joined by our 0:01:00.840 --> 0:01:04.759 colleague Matt Levine, and during that conversation that he described 0:01:04.840 --> 0:01:07.720 much of defy or at least the concept of yield 0:01:07.760 --> 0:01:10.200 farming in a matter that sounded a lot like a 0:01:10.240 --> 0:01:13.120 Ponzi scheme. And of course he didn't describe f t 0:01:13.400 --> 0:01:16.280 X or l AMDA his own business as a ponzi exactly. 0:01:16.800 --> 0:01:20.759 We're still learning details about how it all collapsed. Nonetheless, 0:01:20.760 --> 0:01:23.040 in the wake of this implosion, I don't think there's 0:01:23.080 --> 0:01:25.400 any doubt that those comments are seen as just a 0:01:25.440 --> 0:01:28.720 massive red flag, right, And as you mentioned, this is 0:01:28.760 --> 0:01:32.960 all moving very very fast. It is a gigantic, really 0:01:33.200 --> 0:01:35.720 sprawling story that ranks up there with some of the 0:01:35.760 --> 0:01:40.200 biggest corporate disasters and potentially frauds of all time. And 0:01:40.520 --> 0:01:43.760 given its size and speed, it's a little tough to 0:01:43.880 --> 0:01:47.400 do a complete overview of the story right now. There's 0:01:47.520 --> 0:01:50.120 already been a lot that's been written about it and 0:01:50.200 --> 0:01:54.240 reported on at multiple media outlets, So we're going to 0:01:54.360 --> 0:01:56.840 try to do things a little bit differently. We're going 0:01:56.880 --> 0:01:59.680 to try to break things up a bit. First, today, 0:01:59.760 --> 0:02:02.160 we are going to be speaking with a market participant 0:02:02.160 --> 0:02:04.920 who used f t X and was caught out in 0:02:05.000 --> 0:02:08.720 the scandal. And then we'll be speaking to a whistleblower 0:02:08.760 --> 0:02:11.799 who was among the first to sound the alarm at 0:02:11.880 --> 0:02:15.200 something actually being wrong at f t X and Alameda, right, 0:02:15.240 --> 0:02:17.120 And then we're going to be doing a follow up 0:02:17.160 --> 0:02:20.320 episode tomorrow with Matt Levine where we'll try to understand 0:02:20.400 --> 0:02:24.240 more about what's happened and what is currently going on 0:02:24.440 --> 0:02:26.240 right now in the wake of the collapse. But for 0:02:26.280 --> 0:02:28.040 now we want to start with the gunny guy. Avoid 0:02:28.160 --> 0:02:30.800 here's the founder and CEO of Winter Mute, which is 0:02:30.840 --> 0:02:34.560 the largest crypto market making operation. It had been very 0:02:34.600 --> 0:02:36.960 active on f t X and he's going to help 0:02:37.040 --> 0:02:39.200 us understand a little bit more about f t X 0:02:39.240 --> 0:02:42.840 and Alameda's role in crypto market structure. So if Jenny, 0:02:42.960 --> 0:02:45.440 thank you so much for coming on the Odd Lots podcast. 0:02:45.480 --> 0:02:47.200 What do you tell us? What do you do and 0:02:47.240 --> 0:02:50.440 what is Winter Mute? Yeah, Hi, thanks for having me. Um, Yeah, 0:02:50.440 --> 0:02:52.640 I guess very quickly about the Wintermute one of the 0:02:52.720 --> 0:02:55.840 largest crypto market makers. We've been at since the Southern 0:02:55.880 --> 0:02:59.280 suventeen and I think at this stage basically is the 0:02:59.360 --> 0:03:03.680 largest the native market maker after Salamat's device. It's it's 0:03:03.680 --> 0:03:06.240 an interesting way to have to grow the ranks with 0:03:06.320 --> 0:03:08.760 all the with all that's going on, but basically we 0:03:09.040 --> 0:03:13.680 our businesses across centralized exchanges, decentralized exchanges have been active 0:03:13.720 --> 0:03:16.079 and defie since two thousand nineteen, and of course or 0:03:16.120 --> 0:03:18.680 to see as well. So it's a lot of different 0:03:18.880 --> 0:03:23.320 diversified activities. We are training billions dollars per per day 0:03:23.320 --> 0:03:27.360 on all kinds of venues and in general very deeply 0:03:27.400 --> 0:03:30.760 intertwined with the crypt taker system. Can you explain that 0:03:30.760 --> 0:03:32.959 a little bit further? What does a market maker in 0:03:33.080 --> 0:03:36.960 crypto actually do? And what is your business relationship with 0:03:37.040 --> 0:03:41.720 the various exchanges like f TX. Yeah, on centralized exchanges, 0:03:41.800 --> 0:03:45.960 that's fairly similar to how it works and traditional finance. 0:03:46.000 --> 0:03:49.080 We basically provide bits and offers all great nuclear so 0:03:49.160 --> 0:03:52.880 that people get access to liquidity. When it kin comes 0:03:52.880 --> 0:03:56.080 to like how we work with different exchanges, There's are 0:03:56.120 --> 0:03:59.240 no formal contracts as such. It's basically us creating an 0:03:59.240 --> 0:04:02.280 account just like everyone else, institutional accounts of like everyone else, 0:04:02.680 --> 0:04:08.920 and and then effectively sent orders via APIs that exchanges proved, 0:04:09.960 --> 0:04:14.160 so you mentioned they. You know, you've sort of grown 0:04:14.320 --> 0:04:17.560 in market share since the collapse of Lamet, which was 0:04:17.600 --> 0:04:22.280 a competitor. Can you talk about prior to the implosion 0:04:22.520 --> 0:04:26.119 of f t X and elmet, which of course will discuss, 0:04:26.240 --> 0:04:29.240 what was your understanding and what was your view of 0:04:29.320 --> 0:04:33.160 the f t X L meter relationship. I guess yeah, 0:04:33.200 --> 0:04:36.600 I mean alla meta part was pretty shocking because basically 0:04:36.560 --> 0:04:39.760 pretty much everyone in the industry, including myself, assumes that 0:04:39.839 --> 0:04:43.599 Alometer has been writing running a pretty big and successful 0:04:43.800 --> 0:04:48.080 operation across centralized and decentralized exchanges. For example, we've seen 0:04:48.120 --> 0:04:51.799 them being very active on Serum, the Solana Native decks. 0:04:52.000 --> 0:04:54.600 We suspected that they have a pretty decent market share 0:04:54.760 --> 0:04:59.400 on ft act itself, despite actually not being able to 0:05:00.279 --> 0:05:03.760 confirm it or denied, simply because one of the great 0:05:03.800 --> 0:05:05.960 things that was about FC access that they had a 0:05:06.040 --> 0:05:09.839 leaderboards for volumes and they were pretty public being there 0:05:10.000 --> 0:05:13.359 and saying that they have InterMute is whatever number it 0:05:13.600 --> 0:05:16.600 or number three or like whatever depends on the months, 0:05:16.680 --> 0:05:18.719 but you will would never see lamed in there, which 0:05:18.800 --> 0:05:21.880 was kind of annoying. But yeah, we kind of got that, 0:05:22.160 --> 0:05:26.080 and not everyone wants to show the presence his publicly 0:05:26.160 --> 0:05:28.680 is with it. And now I'm thinking, okay, probably the 0:05:28.800 --> 0:05:33.080 market chair was either very big or maybe potentially very small, 0:05:33.120 --> 0:05:36.560 but more likely very big because like one of the 0:05:36.560 --> 0:05:40.359 main because like I guess the main question is given 0:05:40.400 --> 0:05:42.279 the size of the whole that f t X faces 0:05:42.360 --> 0:05:45.600 and that they lend money to al Ameter, somehow they 0:05:45.600 --> 0:05:47.400 managed to lose those those billions, and one of the 0:05:48.800 --> 0:05:52.000 one of the ideas that's traveling around crypto Twitter is 0:05:52.040 --> 0:05:55.240 basically they were just very bad market maker, I guess 0:05:55.440 --> 0:05:57.680 because they were kind of forced to probably quit it 0:05:57.760 --> 0:06:00.440 an f t X and they basically lost of money 0:06:00.440 --> 0:06:05.080 over the years supporting that liquity. What was your experience 0:06:05.320 --> 0:06:09.679 of last week, Like when did you first think, oh, 0:06:09.839 --> 0:06:12.960 something is happening here. I saw you were quite early 0:06:13.000 --> 0:06:15.960 about pulling your money out of FTX dot US, but 0:06:16.040 --> 0:06:18.080 some of it is stuck in f t X, the 0:06:18.160 --> 0:06:22.359 international exchange. Like what exactly was your experience of the 0:06:22.400 --> 0:06:25.200 past ten days or so? Yeah, I guess when it 0:06:25.279 --> 0:06:27.960 all started with the z twits and when like the 0:06:28.040 --> 0:06:31.279 first withdrawals started in f t X, we basically withdrew 0:06:31.760 --> 0:06:35.240 roughly half of our capital simply because we basically our 0:06:35.240 --> 0:06:38.200 initial assessment was exchange is going to be fine, but 0:06:38.279 --> 0:06:43.000 it might be struggling with withdrawals for sometimes. So it's 0:06:43.080 --> 0:06:45.440 just not smart to look so much so much of 0:06:45.480 --> 0:06:48.159 our capital over there, because like the way we operate, 0:06:48.240 --> 0:06:50.720 we have a number of exchanges that we consider to 0:06:50.760 --> 0:06:54.000 be safe. That's where like we basically have a pool 0:06:54.000 --> 0:06:57.440 of capitals that we keep rebalancing across different exchanges. And 0:06:57.520 --> 0:07:01.960 also decentralized exchanges, so block chains, and we have a 0:07:02.040 --> 0:07:04.680 number of exchanges which considered safe and we basically park 0:07:04.880 --> 0:07:08.440 excess capital there every nouncer because it's quite easy, like 0:07:08.440 --> 0:07:10.880 an fc X was one of them, like great things 0:07:10.920 --> 0:07:13.720 that was about. Another great thing that was about fc 0:07:13.800 --> 0:07:17.000 access that you could actually like very different stable coins 0:07:17.040 --> 0:07:19.760 again versus each other. So it was it was a 0:07:19.880 --> 0:07:23.480 very neat exchange to park the cash on. But once 0:07:23.520 --> 0:07:26.080 we realized that, yeah, it might be that they may 0:07:26.080 --> 0:07:29.640 be close withdrawals for a bit, maybe maybe they just 0:07:29.720 --> 0:07:33.240 be super swamped with withdrawal requests, so it's just not 0:07:33.640 --> 0:07:36.040 great to have our capital on it. So we kind 0:07:36.040 --> 0:07:39.040 of stopped doing that and we decreased it. But the 0:07:39.040 --> 0:07:42.000 initial assessment was, yeah, we want to keep some capital 0:07:42.040 --> 0:07:43.840 on because it's going to be great, great trading with 0:07:43.920 --> 0:07:47.280 all this moilutility, Yeah, which turned out to be not 0:07:47.280 --> 0:07:49.880 not very great. I want to go back to something 0:07:49.960 --> 0:07:53.880 you said about maybe l meter or just bad market makers, 0:07:54.000 --> 0:07:57.040 or maybe they were bad market makers because they were 0:07:57.440 --> 0:08:00.280 the official or the unofficial in house, like did the 0:08:00.480 --> 0:08:04.240 provider at f t X. I believe and correct me 0:08:04.320 --> 0:08:07.920 if I'm wrong. I believe that prior to all of this, 0:08:08.120 --> 0:08:10.720 say like prior to this month or whatever, there was 0:08:10.760 --> 0:08:14.600 a perception that LAMA had an unfair advantage in its 0:08:14.640 --> 0:08:18.160 trading by dint of its relationship with f t X. 0:08:18.200 --> 0:08:21.720 In other words, that it was very profitable but only 0:08:21.760 --> 0:08:24.800 because it had some sort of relationship with f t X. 0:08:25.600 --> 0:08:28.960 Did you have that perception beforehand that it gave an advantage, 0:08:29.200 --> 0:08:32.040 and now do you sort of view this other side 0:08:32.080 --> 0:08:35.720 that essentially the relationship actually made it harder for ELMA 0:08:35.760 --> 0:08:39.240 to make money. I guess my perception was that Alameda 0:08:40.080 --> 0:08:42.880 was not really market making an f t X to 0:08:43.040 --> 0:08:46.360 make the liquid it's bigger simply because like, there's enough 0:08:46.400 --> 0:08:49.079 market makers and ft X like. It was probably the 0:08:49.080 --> 0:08:54.079 second biggest exchange in primitives at some point, and honestly, 0:08:54.200 --> 0:08:56.840 there was no need for somebody to well, some for 0:08:56.880 --> 0:08:59.839 some internal market making desk to provide this liquidity too. 0:08:59.840 --> 0:09:02.560 So I was always thinking, okay, if they would be 0:09:02.559 --> 0:09:04.480 losing money and that they would just pull off and 0:09:04.679 --> 0:09:07.240 let others do their job, because yes, there is there 0:09:07.280 --> 0:09:10.000 are enough crypto market makers to fill in, and so 0:09:10.240 --> 0:09:13.560 my assumption was barely the advantage that would be getting 0:09:13.600 --> 0:09:18.280 would be from basically realizing some synergies between like ft 0:09:18.520 --> 0:09:21.640 X listings for example, like especially I think it was 0:09:21.960 --> 0:09:25.680 especially apparent on Slana where alamido ft X with investing 0:09:25.720 --> 0:09:29.520 the protocols and then protocol would pretty much immediately get 0:09:29.600 --> 0:09:32.720 listed an f t X and basically, since Alameda would 0:09:32.760 --> 0:09:37.360 have well hising inventory of tokens or barely better knowledge 0:09:37.400 --> 0:09:39.480 about like how all this stuff works, they would be 0:09:39.640 --> 0:09:42.480 just better position to make money out of it. And 0:09:42.640 --> 0:09:46.080 that that to me sounded like a pretty good money 0:09:46.160 --> 0:09:48.600 making machine, and that that was one of the reasons 0:09:48.679 --> 0:09:51.439 I kind of expected them to print billions last year 0:09:51.520 --> 0:09:54.000 and maybe you know, just losing some of the venture 0:09:54.000 --> 0:10:13.480 positions this year, and that's a so one of the 0:10:13.600 --> 0:10:17.559 criticisms that's come out about this whole thing, and my gosh, 0:10:17.720 --> 0:10:20.160 there are a lot of things that you could say here, 0:10:20.320 --> 0:10:22.800 but one of them is that this is all about 0:10:22.960 --> 0:10:27.560 centralized finance because f t X was a centralized exchange, 0:10:27.720 --> 0:10:31.679 it wasn't actually decentralized like a lot of other ones are, 0:10:31.800 --> 0:10:34.959 like Newna Swap being the primary example. What do you 0:10:35.080 --> 0:10:38.760 say to that criticism? Is that fair? I think it's 0:10:38.840 --> 0:10:43.679 very fair. Basically, I I really disagree that this is 0:10:43.720 --> 0:10:46.760 a failure of Crypta. I think it's, if anything, it's 0:10:46.800 --> 0:10:51.040 a vindication of everything Crypta is standing for, which is 0:10:51.080 --> 0:10:56.680 basically being completely decentralized, being committed, transparent, and basically, yeah, 0:10:56.760 --> 0:11:00.400 people have an access to their own assets, self study, 0:11:00.720 --> 0:11:04.520 and basically not relying on anything really because like, if 0:11:04.559 --> 0:11:07.199 you look at all the failures of this year, like 0:11:07.320 --> 0:11:10.040 even Terror, even though it was a decentralized protocol, the 0:11:10.120 --> 0:11:13.439 only reason Terror blew up to such a massive proportions 0:11:13.640 --> 0:11:16.360 was well, I think, in my opinion at least, the 0:11:16.480 --> 0:11:19.920 main reason was that this crazy twenty percent yield was 0:11:20.080 --> 0:11:23.959 repackaged by centralized institutions in the very opaque way it 0:11:24.080 --> 0:11:27.400 was repackaged to the retail, and so retail would not 0:11:27.559 --> 0:11:31.840 get ten percent guild on their account. Those centralized entities 0:11:31.880 --> 0:11:34.400 would get like ten percent spread over the Terror, and 0:11:34.520 --> 0:11:37.920 everyone was happy unless until until basically everything blew up. 0:11:38.600 --> 0:11:41.800 The three arrow blob exactly the same thing, like everyone 0:11:42.000 --> 0:11:45.199 lent to three arrows, but nobody knew what the combined 0:11:45.600 --> 0:11:48.240 value of the loans was, and then when they blew up, Yeah, 0:11:48.760 --> 0:11:51.959 all the centralized landers got hurt simply because it was 0:11:52.040 --> 0:11:56.000 just centralized and transparent. And finally f t X exactly 0:11:56.000 --> 0:11:58.719 the same thing they lent Alamida. Nobody knew about it 0:11:58.800 --> 0:12:01.280 because it was yeah, it was not public. It was 0:12:01.480 --> 0:12:04.599 like somewhere on the internal ledger, not even most of 0:12:04.640 --> 0:12:06.679 the place knew about it, I think. And again it 0:12:06.760 --> 0:12:10.079 blew up in a very centralized fation. So to me, 0:12:10.840 --> 0:12:13.920 the crypto kind of solves it. And it's quite unfortunate 0:12:13.960 --> 0:12:16.640 toward a trip to exchange, but it's a to me, 0:12:16.760 --> 0:12:20.120 it's an indication of all the cryptoe ideals. So I 0:12:20.240 --> 0:12:22.880 want to go back to the experience of trading on 0:12:23.080 --> 0:12:26.640 f t X specifically, you mentioned that your perception prior 0:12:26.679 --> 0:12:29.080 to this was that f t X was a high 0:12:29.200 --> 0:12:33.280 quality exchange, a safe place to park capital when it 0:12:33.559 --> 0:12:36.880 wasn't in a in a trade. What was good about 0:12:36.960 --> 0:12:39.040 f t X. It does seem to me like when 0:12:39.080 --> 0:12:42.600 I've talked to professional crypto traders, they liked this site. 0:12:42.679 --> 0:12:44.840 And I think part of the reason maybe people are 0:12:44.880 --> 0:12:49.559 so blindsided by this collapse is the fact that by 0:12:49.600 --> 0:12:52.920 all accounts, the f t X platform seemed to be 0:12:53.000 --> 0:12:55.240 a good one that people really like the experience that 0:12:55.360 --> 0:12:57.679 had a lot of uptime, People like the way the 0:12:57.760 --> 0:13:01.400 liquidations worked on ft X, can you talk about your 0:13:01.520 --> 0:13:04.880 experience just being a traitor on it and in your 0:13:05.000 --> 0:13:08.760 view why f t X got the prominence that it did. 0:13:10.000 --> 0:13:13.920 I think it's very much linked to SBF to be honest, Like, yes, 0:13:14.000 --> 0:13:17.280 sure it was a good platform to trade, especially because 0:13:17.320 --> 0:13:22.760 they were pretty yeah there's a pretty quick about listen ussets, 0:13:22.760 --> 0:13:26.480 they're pretty quick about well. Basically I guess that perhaps 0:13:26.480 --> 0:13:28.240 were great and there were like a lot of different 0:13:28.280 --> 0:13:31.199 products products. One sense it wasn't great about f t 0:13:31.480 --> 0:13:34.520 X and everyone knows about it as well, is this 0:13:34.640 --> 0:13:37.400 throughput was pretty a bit small, like a number of 0:13:37.440 --> 0:13:39.800 others you can send per second. Like all those limits, 0:13:39.840 --> 0:13:42.599 they kept increasing them by I don't know. Sometimes it 0:13:42.600 --> 0:13:44.920 would come and say, oh, good, great news for increased 0:13:45.800 --> 0:13:49.760 but they needed like a two th increase. And one 0:13:49.840 --> 0:13:52.040 of the reasons I think, and yeah, I don't know, 0:13:52.200 --> 0:13:54.000 I have no idea, like I haven't looked through the 0:13:54.080 --> 0:13:56.599 code base, but I guess one of the reasons for 0:13:56.720 --> 0:14:02.400 this was basically what was promoting very much is that 0:14:02.600 --> 0:14:07.120 the matching engine was combined with their risk engine, because 0:14:07.200 --> 0:14:09.680 you could use all those assets on margin as collateral 0:14:09.800 --> 0:14:13.199 to trade all this perpetual products, like every time you 0:14:13.240 --> 0:14:15.240 send another and it gets matched. It would need to 0:14:15.280 --> 0:14:19.400 do both matching and their risk calculation, which I would 0:14:19.400 --> 0:14:22.720 imagine would slow it quite considerably. And I was always 0:14:22.800 --> 0:14:27.360 like very curious how much you can leverage that, like 0:14:27.440 --> 0:14:30.880 how much how much you can expand that, because honestly, 0:14:30.960 --> 0:14:33.720 like none of the centralized cryptic exchanges are close to 0:14:33.800 --> 0:14:38.360 strouput to even like the worst I know European Stock 0:14:38.400 --> 0:14:41.640 Exchange for example, they just all really really bad, like 0:14:41.960 --> 0:14:46.000 even violence, and ft X was quite worse than violence 0:14:46.200 --> 0:14:49.440 in the throughput twice even though it was smaller volume wis. 0:14:49.560 --> 0:14:52.080 So to me, it was always a question like how 0:14:52.160 --> 0:14:53.760 is it going to grow? Like how is it gonna 0:14:54.480 --> 0:15:00.040 continue making this throutput, especially because like if as a 0:15:00.120 --> 0:15:02.480 bitcoin in the studium, that's fine. But the idea I 0:15:02.640 --> 0:15:06.440 was I think was like expanded the stocks, expanded, the commodities, 0:15:06.520 --> 0:15:09.560 expanding everything. Also, I don't know, maybe maybe it was 0:15:09.640 --> 0:15:12.480 all the facade, but to me, it was always securious 0:15:12.480 --> 0:15:15.600 saying like, yeah, you mentioned engine clearing, it's not great 0:15:15.920 --> 0:15:19.960 that you actually should it. It's physically possible to maintain 0:15:20.040 --> 0:15:24.520 this model. So just to be clear, your perception was 0:15:24.640 --> 0:15:27.280 that there was this trade off that the throughput in 0:15:27.440 --> 0:15:30.120 terms of total volume at f t X was not 0:15:30.920 --> 0:15:33.680 say what you could get it binance, but that was 0:15:33.880 --> 0:15:38.920 because it had this superior matching engine and ability to 0:15:39.120 --> 0:15:42.120 cross margin assets. Can you just explain what that means 0:15:42.280 --> 0:15:45.520 for the listener, that the matching engine, the cross margining asset, 0:15:45.800 --> 0:15:50.160 and what is the advantage of that as a professional trader. Yeah, 0:15:50.200 --> 0:15:52.400 so basically if you go if you go on binance 0:15:52.440 --> 0:15:55.440 and trade binance features, you need to post b U, 0:15:55.560 --> 0:15:58.280 s D or some I think there's a cross margin 0:15:58.400 --> 0:16:00.560 was also stable quins as well now, and but that's 0:16:00.560 --> 0:16:02.880 that's just one instrument. If you if you would trade 0:16:02.920 --> 0:16:04.960 on ft X, you could use your bitcoin, you can 0:16:05.080 --> 0:16:06.960 use an e Syria, and you can use like a 0:16:07.080 --> 0:16:11.480 dozens of other different tokens for which the price would 0:16:11.480 --> 0:16:13.800 be quite volatile. Right, So let's say you have one 0:16:13.840 --> 0:16:16.040 million works of bitcoin and one million works of dollars. 0:16:16.560 --> 0:16:18.680 You could use both of those positions too as a 0:16:18.760 --> 0:16:23.000 margin to enter delative contracts, which is very handy because 0:16:23.080 --> 0:16:27.960 you like you trade sport, you trade perpet you trade perpetuals. 0:16:28.000 --> 0:16:30.360 You can do it all on one account. And yeah, 0:16:30.480 --> 0:16:34.040 it's it's very convenient from the capital perspective because you, 0:16:34.200 --> 0:16:36.440 for example, like what if you only have bitcoin then 0:16:36.520 --> 0:16:40.640 just parking bitcoin and still get margin on on FTX. 0:16:40.760 --> 0:16:43.560 That's great, but that also means that you're met that 0:16:43.680 --> 0:16:47.200 the matching engine continuously had to keep in its mind 0:16:47.520 --> 0:16:51.040 what the price of bitcoin is, because maybe if bitcoin fALS, 0:16:51.120 --> 0:16:53.480 you'll need to liquidate this guy, because really if his 0:16:53.560 --> 0:16:57.040 positions are you know, like his his unrealized profit as 0:16:57.360 --> 0:16:59.480 like THEO low, and so maybe we need to liquidate 0:16:59.560 --> 0:17:03.400 his bitcoin line and then initiate like the whole liquidation process, 0:17:03.480 --> 0:17:06.959 which which is basically way more complex than for binus futures, 0:17:07.000 --> 0:17:10.960 which is just yeah, we just take your perpetual exposure. 0:17:11.480 --> 0:17:14.600 It's it's generous dollars bn L and you only have 0:17:14.720 --> 0:17:17.600 dollars and your accounts. It's very easy. This actually leads 0:17:17.640 --> 0:17:19.560 into something I wanted to ask you, which is a 0:17:19.640 --> 0:17:23.280 sort of I guess fundamental criticism about defy, which is, 0:17:23.400 --> 0:17:27.840 even if you're dealing with decentralized exchanges or you know, 0:17:28.040 --> 0:17:31.040 you hold your own keys or whatever, because you don't 0:17:31.080 --> 0:17:36.440 actually know who you're lending to, DEFY often places a 0:17:36.600 --> 0:17:40.440 lot of emphasis on the underlying collateral. The quality of 0:17:40.480 --> 0:17:42.840 the collateral and the price of it, and so if 0:17:42.880 --> 0:17:45.480 it starts to go down, you do get these big 0:17:45.560 --> 0:17:49.359 margin calls because you don't really know who your counterparty is. 0:17:49.640 --> 0:17:55.520 That seems like a potential vulnerability for defy, particularly in 0:17:55.600 --> 0:17:58.440 an environment where we're seeing the collateral price, you know, 0:17:58.800 --> 0:18:04.200 be extremely volatile. Well, I think that like defy challenges 0:18:04.280 --> 0:18:06.639 potentially is they don't stem from the fact that you 0:18:06.720 --> 0:18:10.160 don't know who your counterpart is. They're primarily stemming from 0:18:10.200 --> 0:18:12.560 two things. One is just a quality of collectal like 0:18:12.640 --> 0:18:15.560 you mentioned, and basically the fact that in extreme market 0:18:15.640 --> 0:18:18.720 moves it might not be possible to liquidated quick enough 0:18:18.840 --> 0:18:22.200 on the access and that basic basically the quality of 0:18:22.240 --> 0:18:25.000 the collateral can drop so quickly, is that there will 0:18:25.040 --> 0:18:27.680 be no incentive for anyone to do liquidations, and that 0:18:27.920 --> 0:18:32.280 can create gaps and products like whole wholes and protocol 0:18:32.400 --> 0:18:35.200 treasuries for example, which which happened after Lunar with with 0:18:35.320 --> 0:18:39.200 some protocols. And the second second issue is it's also 0:18:39.280 --> 0:18:43.800 related to just potential manipulation vectors and that that has 0:18:43.840 --> 0:18:46.200 to do with oracles because you like what we what 0:18:46.280 --> 0:18:48.280 we've seen with malgo markets. So a few weeks ago, 0:18:48.359 --> 0:18:51.760 for example, was that yeah, if you manipulate the oracle price, 0:18:51.880 --> 0:18:54.320 you can potentially create sort of run the bank for 0:18:54.400 --> 0:18:57.400 those sounds of slending protocols as well. Honestly, I think 0:18:57.440 --> 0:19:00.440 it's all like both of those issues are solvable. Basically 0:19:00.600 --> 0:19:03.840 most most of the time issues with descentralized protocols were 0:19:03.880 --> 0:19:07.119 caused by just the quality of a collateral, which works 0:19:07.200 --> 0:19:09.880 both on like a float level, so if if it's 0:19:09.920 --> 0:19:12.040 just some old coin which has a very low float 0:19:12.240 --> 0:19:14.879 yet it just wouldn't be able to liok at it 0:19:15.359 --> 0:19:17.719 fast enough, but also on the oracle level as well, 0:19:18.119 --> 0:19:20.359 if it's like a very small talk and it's very 0:19:20.359 --> 0:19:24.040 easy to manipulate surprise, just like the service Mango. But overall, 0:19:24.160 --> 0:19:28.920 like the defied protocols over the last I know, nine 0:19:29.000 --> 0:19:32.760 twelve months worked perfectly well in terms of stress. Like 0:19:33.040 --> 0:19:35.600 you haven't seen any issues with other you haven't seen 0:19:35.600 --> 0:19:38.960 any issues with Compound. They worked just like they intended 0:19:39.000 --> 0:19:41.720 to be. People are doing liquidations at all. It was 0:19:41.760 --> 0:19:44.800 all very much elderly, so I think from that perspective, 0:19:44.880 --> 0:19:48.720 I think it's still pretticol system to look at. Are 0:19:48.840 --> 0:19:53.240 you going to change the way you think about counterparty 0:19:53.520 --> 0:19:57.520 risk and keeping money on exchanges, any exchanges in the 0:19:57.600 --> 0:20:00.439 wake of this, how are you changing your business up rations? 0:20:01.400 --> 0:20:03.520 We're basically playing it very safe. So there are like 0:20:03.800 --> 0:20:07.480 very small number of exchanges that we considered to be safe, 0:20:07.680 --> 0:20:09.320 and we have is drawing from a lot of like 0:20:09.440 --> 0:20:11.560 what we call t S three exchanges at the moment, 0:20:11.680 --> 0:20:15.360 and even some of the tear too exchanges. We basically 0:20:16.600 --> 0:20:19.440 and I cannot really name names because we actually have 0:20:19.520 --> 0:20:22.560 a pretty good relationship with most of them. And like 0:20:22.720 --> 0:20:25.600 if I say it Withdrew from Exchange, why it does 0:20:25.720 --> 0:20:27.840 mean that we think the bankrupt. It just means that 0:20:27.960 --> 0:20:30.240 we are We're just waiting for our due diligence. We're 0:20:30.280 --> 0:20:32.480 waiting for them to provide financials. We're just waiting for 0:20:32.600 --> 0:20:35.240 them to kind of prove to us that they're fine. 0:20:35.359 --> 0:20:38.640 And so the idea for us is just to get 0:20:38.720 --> 0:20:40.800 this due diligence done and in a week or two 0:20:40.880 --> 0:20:43.280 we just resume training and all the exchanges that remain. 0:20:43.480 --> 0:20:47.239 But my unfortunate yeah, understanding of the situation is there 0:20:47.320 --> 0:20:51.680 might be more exchanges failing, like big and small. So yeah, 0:20:51.760 --> 0:20:55.960 it's just very good security practice to be very safe 0:20:56.000 --> 0:20:59.240 in this environment. I think for individual users, I would 0:20:59.480 --> 0:21:02.639 generally advice to try self custody if you can, or 0:21:02.760 --> 0:21:06.040 withdraw to like very very high quality exchanges if you cannot. 0:21:06.800 --> 0:21:10.399 What exactly is the path for contagion here? Like what 0:21:10.600 --> 0:21:13.320 happens if ft X goes down? Like why should we 0:21:13.400 --> 0:21:17.800 worry about a Tier three exchange? What's the knock on effect? Um? 0:21:17.920 --> 0:21:21.280 The knock on effect can be I guess like multiple vectors. 0:21:21.640 --> 0:21:24.520 First of all, sounds of exchanges like the way they 0:21:24.600 --> 0:21:27.280 operate with market makers. Some of them they provide letters 0:21:27.320 --> 0:21:30.560 of credit, so they basically can extend what leverage effectively 0:21:30.640 --> 0:21:33.639 or just be your free capital to market makers so 0:21:33.800 --> 0:21:37.399 that the market makers would have better incentive to provide markets. 0:21:37.960 --> 0:21:41.040 So if any of those history exchanges gave lots of 0:21:41.119 --> 0:21:45.359 credit Dolometer and Telometer lost money, they could be facing 0:21:45.400 --> 0:21:48.840 some losses. Second way it can be can be happening 0:21:49.000 --> 0:21:51.160 is basically a lot of exchanges will be keeping there 0:21:51.560 --> 0:21:54.520 like some of their capital on FTX. And we've seen 0:21:54.600 --> 0:21:58.120 like some some exchanges coming forward and meeting that whether 0:21:58.200 --> 0:22:00.800 it can be for internal market maker in operations, So 0:22:00.920 --> 0:22:05.159 it can be simply because they just sometimes basically like 0:22:05.320 --> 0:22:08.840 broadcast liquidity on the exchange via ft X either way, 0:22:08.920 --> 0:22:13.200 they could potentially keep some of their money on ft X. 0:22:13.600 --> 0:22:16.440 You know, one of the interesting things that makes crypto 0:22:16.720 --> 0:22:20.440 distinct from say tread fire the stock market if you 0:22:20.560 --> 0:22:24.960 do have this separation of the broker and the market. 0:22:25.040 --> 0:22:27.159 And so if I log into a Schwab account to 0:22:27.240 --> 0:22:31.359 trade stocks, Schwab is the broker, but Schwab is not 0:22:31.840 --> 0:22:36.200 the stock market itself. It seems like that different in crypto, 0:22:36.280 --> 0:22:39.240 where it's like finance or f t X, that's you 0:22:39.320 --> 0:22:41.760 have your account there, but it also is the market 0:22:41.800 --> 0:22:45.160 and that has the order book. Would it make sense 0:22:45.320 --> 0:22:48.840 to separate those? Yeah, it's a very good question. Um, 0:22:50.640 --> 0:22:53.320 I guess it depends as the right answer. On one hand, 0:22:54.080 --> 0:22:57.520