1 00:00:01,050 --> 00:00:03,420 The following content is provided under a Creative 2 00:00:03,420 --> 00:00:04,810 Commons license. 3 00:00:04,810 --> 00:00:07,020 Your support will help MIT OpenCourseWare 4 00:00:07,020 --> 00:00:11,110 continue to offer high quality educational resources for free. 5 00:00:11,110 --> 00:00:13,680 To make a donation or to view additional materials 6 00:00:13,680 --> 00:00:17,640 from hundreds of MIT courses, visit MIT OpenCourseWare 7 00:00:17,640 --> 00:00:18,526 at ocw.mit.edu. 8 00:00:23,050 --> 00:00:24,050 PROFESSOR: Welcome back. 9 00:00:26,954 --> 00:00:30,970 I know we're sort of getting into the middle of the week. 10 00:00:30,970 --> 00:00:33,610 Some of you maybe want to keep your laptops open to work 11 00:00:33,610 --> 00:00:35,890 on other courses' midterms, but I 12 00:00:35,890 --> 00:00:38,940 will ask you to try to focus on Blockchain and Money today. 13 00:00:43,550 --> 00:00:46,610 Today, we're going to chat about permissioned 14 00:00:46,610 --> 00:00:48,762 versus permissionless systems. 15 00:00:48,762 --> 00:00:49,470 I want to thank-- 16 00:00:49,470 --> 00:00:51,060 I have a guest here. 17 00:00:51,060 --> 00:00:53,740 I call out guest when they're here and I know them. 18 00:00:53,740 --> 00:00:59,060 Mark Snyderman, who runs a fund at Fidelity-- 19 00:00:59,060 --> 00:01:02,110 Mark runs a 6-- about $6 billion? 20 00:01:02,110 --> 00:01:05,725 Called the Real Estate Income Fund at Fidelity. 21 00:01:05,725 --> 00:01:07,100 And you might say, well, why does 22 00:01:07,100 --> 00:01:09,725 he want to come to a Blockchain and Money course? 23 00:01:09,725 --> 00:01:11,600 It's because we went to high school together. 24 00:01:11,600 --> 00:01:13,320 [LAUGHTER] 25 00:01:13,320 --> 00:01:19,490 So-- and I was at his wedding. 26 00:01:19,490 --> 00:01:23,120 And he's getting married again next week. 27 00:01:23,120 --> 00:01:23,620 Yeah. 28 00:01:23,620 --> 00:01:27,050 [APPLAUSE] 29 00:01:29,510 --> 00:01:32,750 But you can all inundate Mark later about Fidelity, 30 00:01:32,750 --> 00:01:35,560 because he's been there for a lot of years. 31 00:01:35,560 --> 00:01:37,510 See what I'm going to do? 32 00:01:37,510 --> 00:01:42,650 I'm shamelessly for MIT and MIT students. 33 00:01:42,650 --> 00:01:45,790 So let's get going. 34 00:01:45,790 --> 00:01:48,380 I was letting a little bit of time for a few more people 35 00:01:48,380 --> 00:01:50,820 to wander in. 36 00:01:50,820 --> 00:01:52,120 Today, we're going to talk-- 37 00:01:52,120 --> 00:01:54,960 of course, we're going to chat a little bit about the readings 38 00:01:54,960 --> 00:01:56,610 and study questions. 39 00:01:56,610 --> 00:01:58,200 Going to go back again to, what are 40 00:01:58,200 --> 00:02:00,430 the technical and commercial challenges? 41 00:02:00,430 --> 00:02:02,100 And this is relative-- 42 00:02:02,100 --> 00:02:04,770 this is relevant for all of our lectures, 43 00:02:04,770 --> 00:02:06,780 but it's particularly relevant today, 44 00:02:06,780 --> 00:02:09,479 because we're going to be talking about two 45 00:02:09,479 --> 00:02:11,820 different database structures. 46 00:02:11,820 --> 00:02:14,490 One's the permissionless blockchain of Bitcoin. 47 00:02:14,490 --> 00:02:17,370 But now, today, we're going to introduce and go deeper 48 00:02:17,370 --> 00:02:21,330 into the permissioned, or private, set of blockchains, 49 00:02:21,330 --> 00:02:25,920 like the IBM Hyperledger and Corda. 50 00:02:25,920 --> 00:02:30,163 But the technical issues relate to all of that. 51 00:02:30,163 --> 00:02:31,830 And then, of course, we're going to talk 52 00:02:31,830 --> 00:02:36,090 about that against a third type of database-- 53 00:02:36,090 --> 00:02:37,680 traditional databases. 54 00:02:37,680 --> 00:02:43,520 So aligning permissionless, like Bitcoin, permissioned-- 55 00:02:43,520 --> 00:02:46,130 IBM Hyperledger type of-- 56 00:02:46,130 --> 00:02:49,760 and then traditional databases, and why, in a business setting, 57 00:02:49,760 --> 00:02:53,300 you might think of one versus another versus another. 58 00:02:53,300 --> 00:02:56,060 And I'm sure if I run off the rails anywhere 59 00:02:56,060 --> 00:02:59,370 on traditional databases, which I have not personally studied, 60 00:02:59,370 --> 00:03:02,480 Alon will help bail me out somewhere. 61 00:03:02,480 --> 00:03:03,950 Is that a deal? 62 00:03:03,950 --> 00:03:04,450 Maybe? 63 00:03:04,450 --> 00:03:04,755 AUDIENCE: OK. 64 00:03:04,755 --> 00:03:05,630 PROFESSOR: All right. 65 00:03:05,630 --> 00:03:06,653 [LAUGHTER] 66 00:03:06,653 --> 00:03:08,070 AUDIENCE: You asked for it. 67 00:03:08,070 --> 00:03:10,070 PROFESSOR: Yeah, I asked for it. 68 00:03:10,070 --> 00:03:13,640 Mark, Alon is a computer scientists 69 00:03:13,640 --> 00:03:17,720 from other parts of MIT, so he helps me out in subjects 70 00:03:17,720 --> 00:03:20,900 I don't know, and even in subjects I think I know. 71 00:03:25,250 --> 00:03:28,700 So let's just start a little bit with what did you 72 00:03:28,700 --> 00:03:29,717 take from the readings? 73 00:03:29,717 --> 00:03:31,550 There were four or five readings, of course. 74 00:03:31,550 --> 00:03:35,810 But what is a permissioned or private distributed ledger? 75 00:03:35,810 --> 00:03:37,080 Now, I can do this. 76 00:03:37,080 --> 00:03:39,590 I can light them up, as some people said, because Toledo's 77 00:03:39,590 --> 00:03:40,940 given me my list. 78 00:03:43,490 --> 00:03:47,540 But again, class participation-- anybody who hasn't spoken yet 79 00:03:47,540 --> 00:03:51,930 might want to sort of chime in and give it a shot. 80 00:03:51,930 --> 00:03:53,195 I'm not seeing any volunteers. 81 00:03:53,195 --> 00:03:55,560 If I can-- ooh, yes. 82 00:03:55,560 --> 00:03:57,790 Do you want to tell me your first name again? 83 00:03:57,790 --> 00:03:58,880 AUDIENCE: Jayati 84 00:03:58,880 --> 00:03:59,710 PROFESSOR: Jayati. 85 00:03:59,710 --> 00:04:00,520 AUDIENCE: Jayati. 86 00:04:00,520 --> 00:04:02,980 PROFESSOR: OK. 87 00:04:02,980 --> 00:04:06,480 AUDIENCE: So the permissioned DODs, 88 00:04:06,480 --> 00:04:11,520 unlike the permissioned [INAUDIBLE] Bitcoin, 89 00:04:11,520 --> 00:04:13,880 they limit the number of participants. 90 00:04:13,880 --> 00:04:15,770 Essentially, they require the participants 91 00:04:15,770 --> 00:04:18,079 to be authorized before they can enter 92 00:04:18,079 --> 00:04:21,110 into this sort of technology. 93 00:04:21,110 --> 00:04:25,640 And in addition to that, they are said to be stakeholders. 94 00:04:25,640 --> 00:04:27,350 And since they are the only ones involved 95 00:04:27,350 --> 00:04:30,860 in verifying the transaction-- 96 00:04:30,860 --> 00:04:32,540 unlike the permissionless ones, where 97 00:04:32,540 --> 00:04:35,270 it has to be verified at all the nodes, 98 00:04:35,270 --> 00:04:39,680 this limits the permissions required 99 00:04:39,680 --> 00:04:42,230 [INAUDIBLE] stakeholders, which increases, 100 00:04:42,230 --> 00:04:44,630 also, the transaction speed. 101 00:04:44,630 --> 00:04:48,320 So it's like the triangular dilemma 102 00:04:48,320 --> 00:04:51,770 we learned about, about the security and decentralization 103 00:04:51,770 --> 00:04:53,200 and scalability. 104 00:04:53,200 --> 00:04:56,690 So it moves away from decentralization 105 00:04:56,690 --> 00:04:57,937 and towards scalability. 106 00:04:57,937 --> 00:04:58,520 PROFESSOR: OK. 107 00:04:58,520 --> 00:05:02,112 So Jayeta-- 108 00:05:02,112 --> 00:05:03,360 AUDIENCE: Jayehta. 109 00:05:03,360 --> 00:05:07,230 PROFESSOR: Jayeta went through a whole lot. 110 00:05:07,230 --> 00:05:10,950 It's about the number of nodes and permissioned 111 00:05:10,950 --> 00:05:13,950 into it, and that it addresses some 112 00:05:13,950 --> 00:05:17,820 of that Buterin's trilemma. 113 00:05:17,820 --> 00:05:19,720 So at its core, it sort of addresses 114 00:05:19,720 --> 00:05:22,890 some of the scalability issues. 115 00:05:22,890 --> 00:05:26,460 But it does that at a trade off that it's not truly open. 116 00:05:26,460 --> 00:05:29,370 It's not, anyone can write to the ledger. 117 00:05:29,370 --> 00:05:31,630 I mean, that's sort of the fundamental things. 118 00:05:31,630 --> 00:05:35,910 Anything else that folks took out of the core readings 119 00:05:35,910 --> 00:05:39,750 as to how it separates and how it's different? 120 00:05:39,750 --> 00:05:41,480 I mean-- all right. 121 00:05:41,480 --> 00:05:44,640 We'll get a chance to add a little bit. 122 00:05:44,640 --> 00:05:49,790 And then we're going to dive into Hyperledger and Corda 123 00:05:49,790 --> 00:05:51,670 a bit. 124 00:05:51,670 --> 00:05:53,460 We'll talk about Digital Asset Holdings. 125 00:05:53,460 --> 00:05:54,750 What's Digital Asset Holdings? 126 00:05:54,750 --> 00:05:56,535 Anybody know this company? 127 00:05:59,145 --> 00:06:00,380 Alon knows it. 128 00:06:00,380 --> 00:06:02,320 Brotish knows it. 129 00:06:02,320 --> 00:06:03,592 Eilon. 130 00:06:03,592 --> 00:06:05,364 So Eilon. 131 00:06:05,364 --> 00:06:08,207 AUDIENCE: Yeah. 132 00:06:08,207 --> 00:06:10,040 It's basically a competitor of [INAUDIBLE].. 133 00:06:10,040 --> 00:06:12,890 They're trying to build a DLT protocol that 134 00:06:12,890 --> 00:06:15,590 will allow financial institutions to exchange 135 00:06:15,590 --> 00:06:17,180 information and value. 136 00:06:17,180 --> 00:06:19,280 And they started, to my understanding, 137 00:06:19,280 --> 00:06:23,480 in 2016, which is two years after R3 Corda funded 138 00:06:23,480 --> 00:06:27,310 by [INAUDIBLE] JPMorgan, Goldman Sachs. 139 00:06:27,310 --> 00:06:30,170 They're basically the same amount of money-- $110 million. 140 00:06:30,170 --> 00:06:33,860 PROFESSOR: And who, other than Alon 141 00:06:33,860 --> 00:06:36,350 can tell me who runs Digital Asset Holdings? 142 00:06:36,350 --> 00:06:39,262 Just, it shows that you did the reading. 143 00:06:39,262 --> 00:06:40,820 Somebody out there. 144 00:06:40,820 --> 00:06:41,320 Yes. 145 00:06:41,320 --> 00:06:41,820 Priya? 146 00:06:41,820 --> 00:06:42,600 AUDIENCE: It was-- 147 00:06:42,600 --> 00:06:42,840 AUDIENCE: Priya. 148 00:06:42,840 --> 00:06:43,710 AUDIENCE: Yes. 149 00:06:43,710 --> 00:06:45,920 Blythe Mas-- Blythe Masters, or Blythe-- 150 00:06:45,920 --> 00:06:47,090 PROFESSOR: Blythe Masters. 151 00:06:47,090 --> 00:06:47,360 AUDIENCE: Yes. 152 00:06:47,360 --> 00:06:48,235 PROFESSOR: All right. 153 00:06:48,235 --> 00:06:50,930 Does anybody know who Blythe Masters is 154 00:06:50,930 --> 00:06:54,250 and what she's famous for beyond Digital Asset Holdings? 155 00:06:54,250 --> 00:07:00,740 AUDIENCE: She worked at JPMorgan, and she [INAUDIBLE] 156 00:07:00,740 --> 00:07:02,360 credit default swaps. 157 00:07:02,360 --> 00:07:04,310 PROFESSOR: So she worked at JPMorgan, 158 00:07:04,310 --> 00:07:08,310 and she's known around the world of credit default swaps. 159 00:07:08,310 --> 00:07:10,900 Gillian Tett, who some of you-- 160 00:07:10,900 --> 00:07:12,590 the Sloan Fellows will get a chance 161 00:07:12,590 --> 00:07:16,010 to chat with Gillian Tett in New York in a few weeks. 162 00:07:16,010 --> 00:07:20,240 Gillian Tett wrote a whole book about credit default swaps 163 00:07:20,240 --> 00:07:21,340 and JPMorgan. 164 00:07:21,340 --> 00:07:24,080 And Blythe Masters is a central character 165 00:07:24,080 --> 00:07:27,770 in that whole narrative art. 166 00:07:27,770 --> 00:07:29,600 And this is what she's doing now. 167 00:07:29,600 --> 00:07:34,290 She's brilliant, and she's a very good businesswoman. 168 00:07:34,290 --> 00:07:36,510 I ran the Commodity Futures Trading Commission 169 00:07:36,510 --> 00:07:39,480 and had an opportunity to meet with her a lot 170 00:07:39,480 --> 00:07:45,300 because she ran the swap dealer association ISDA at the time, 171 00:07:45,300 --> 00:07:50,513 or was the outside chair of ISDA, if I recall. 172 00:07:50,513 --> 00:07:52,930 And then we're going to talk about the business trade-offs 173 00:07:52,930 --> 00:07:55,110 as well. 174 00:07:55,110 --> 00:07:57,340 But what do you think some of the central business 175 00:07:57,340 --> 00:07:59,470 trade-offs, if I could get two or three of you 176 00:07:59,470 --> 00:08:01,870 to help me out on what are the central business 177 00:08:01,870 --> 00:08:09,560 trade-offs between permissioned and permissionless blockchains? 178 00:08:09,560 --> 00:08:11,960 Let me see if I can get somebody other than Pria. 179 00:08:11,960 --> 00:08:12,860 Yes. 180 00:08:12,860 --> 00:08:14,240 Remind me of your first name. 181 00:08:14,240 --> 00:08:15,740 AUDIENCE: Misha. 182 00:08:15,740 --> 00:08:20,370 So permissioned blockchain have better privacy and protection, 183 00:08:20,370 --> 00:08:22,990 and better scalability. 184 00:08:22,990 --> 00:08:25,528 And no mining [INAUDIBLE]. 185 00:08:25,528 --> 00:08:27,320 PROFESSOR: Misha, let's pause for a minute. 186 00:08:27,320 --> 00:08:32,510 So better scalability, better privacy, and the third point 187 00:08:32,510 --> 00:08:33,679 you're saying? 188 00:08:33,679 --> 00:08:34,520 AUDIENCE: No mining. 189 00:08:34,520 --> 00:08:35,419 PROFESSOR: No mining. 190 00:08:35,419 --> 00:08:39,200 Or, mining is associated with this proof of work. 191 00:08:39,200 --> 00:08:41,826 Any other business trade-off? 192 00:08:41,826 --> 00:08:43,909 AUDIENCE: So for permissioned, you 193 00:08:43,909 --> 00:08:45,897 have some more flexibility with governance. 194 00:08:45,897 --> 00:08:48,605 If you need to make changes, you don't need to rely [INAUDIBLE].. 195 00:08:48,605 --> 00:08:51,560 PROFESSOR: So more flexibility of governance, in essence, 196 00:08:51,560 --> 00:08:54,830 because if you have a club deal or group deal, maybe 197 00:08:54,830 --> 00:08:57,140 you can do that governance changes 198 00:08:57,140 --> 00:09:01,690 amongst the group instead of having thousands of people 199 00:09:01,690 --> 00:09:03,020 participating. 200 00:09:03,020 --> 00:09:03,780 Kelly? 201 00:09:03,780 --> 00:09:07,030 AUDIENCE: There's also a key technological differentiator, 202 00:09:07,030 --> 00:09:10,780 which is the coding language sort of that the inputs can 203 00:09:10,780 --> 00:09:11,350 be made in. 204 00:09:11,350 --> 00:09:15,355 So with Hyperledger, they can use the smart contacts 205 00:09:15,355 --> 00:09:20,680 in any variety of them, versus the alternative, which is 206 00:09:20,680 --> 00:09:22,552 the domain specific language. 207 00:09:22,552 --> 00:09:23,260 PROFESSOR: Right. 208 00:09:23,260 --> 00:09:27,550 So Hyperledger at least promotes themselves in their own 209 00:09:27,550 --> 00:09:31,600 write-ups-- because the reading was really from them-- 210 00:09:31,600 --> 00:09:36,910 that their embedded language is much more flexible, 211 00:09:36,910 --> 00:09:40,030 and that you can use Java, you can use Go, and so forth-- 212 00:09:40,030 --> 00:09:42,320 at least they say they can. 213 00:09:42,320 --> 00:09:46,580 I don't know if it's really as limited in Ethereum. 214 00:09:46,580 --> 00:09:49,390 But they think and they promote that they're 215 00:09:49,390 --> 00:09:51,550 more flexible than Ethereum. 216 00:09:51,550 --> 00:09:53,050 So Kelly's right in that, too. 217 00:09:56,110 --> 00:09:57,730 So these were the readings. 218 00:09:57,730 --> 00:10:00,850 And now, so we're back to basics again. 219 00:10:00,850 --> 00:10:03,470 What is a blockchain? 220 00:10:03,470 --> 00:10:05,690 I know we've spent four or five lectures on it, 221 00:10:05,690 --> 00:10:08,740 but these key components-- 222 00:10:08,740 --> 00:10:09,610 let's start. 223 00:10:09,610 --> 00:10:13,780 Append-only timestamped logs. 224 00:10:13,780 --> 00:10:15,990 Are they both-- by show of hands, 225 00:10:15,990 --> 00:10:20,500 are they in permissionless blockchains? 226 00:10:20,500 --> 00:10:24,760 I hope every single hand goes up. 227 00:10:24,760 --> 00:10:28,900 Are they in permissioned or closed blockchains? 228 00:10:33,000 --> 00:10:33,860 I'm watching. 229 00:10:33,860 --> 00:10:35,560 I'm not-- I'm just going to watch. 230 00:10:35,560 --> 00:10:36,060 All right. 231 00:10:36,060 --> 00:10:38,130 Every hand should go up. 232 00:10:38,130 --> 00:10:42,670 So both sets of blockchains have this concept 233 00:10:42,670 --> 00:10:45,700 that goes back nearly 30 years to our friend 234 00:10:45,700 --> 00:10:51,370 Haber, who started the whole blockchain that's 235 00:10:51,370 --> 00:10:53,200 in The New York Times. 236 00:10:53,200 --> 00:10:56,300 This whole concept is in both sets of blockchains. 237 00:10:56,300 --> 00:10:57,069 Brotish? 238 00:10:57,069 --> 00:10:58,986 AUDIENCE: So I have a question about basically 239 00:10:58,986 --> 00:11:04,858 kind of [INAUDIBLE] append-only [INAUDIBLE] of the ledger. 240 00:11:04,858 --> 00:11:06,400 I think I read in one of the readings 241 00:11:06,400 --> 00:11:07,930 that Corda has a feature where you 242 00:11:07,930 --> 00:11:11,380 can make some changes to the history, which is not 243 00:11:11,380 --> 00:11:14,680 equivalent to a hard fault. So that will probably 244 00:11:14,680 --> 00:11:16,870 be one example where it is not purely 245 00:11:16,870 --> 00:11:18,805 append-only in a permission [INAUDIBLE].. 246 00:11:18,805 --> 00:11:20,680 PROFESSOR: So Brotish is raising that there's 247 00:11:20,680 --> 00:11:22,840 certain features that are promoted 248 00:11:22,840 --> 00:11:26,380 in one of the private blockchains, Corda, 249 00:11:26,380 --> 00:11:31,450 that may allow you not to append-only, but in essence, 250 00:11:31,450 --> 00:11:36,080 delete data, or replace data, possibly. 251 00:11:38,990 --> 00:11:42,305 And I'm not-- because there's also, in private blockchains, 252 00:11:42,305 --> 00:11:46,600 an inability to partition, and in essence, sh-- 253 00:11:46,600 --> 00:11:48,760 it's kind of a form of shorting-- 254 00:11:48,760 --> 00:11:51,380 but to partition the data. 255 00:11:51,380 --> 00:11:54,350 I don't know enough about Corda, whether they literally 256 00:11:54,350 --> 00:11:56,030 allow you to delete-- 257 00:11:56,030 --> 00:11:58,940 or is it something in this partitioning? 258 00:11:58,940 --> 00:12:00,120 But maybe Hugo? 259 00:12:00,120 --> 00:12:03,500 AUDIENCE: I was going to raise a similar point that if-- 260 00:12:03,500 --> 00:12:07,250 I mean, to my understanding, if you have a permissioned system 261 00:12:07,250 --> 00:12:10,595 and there are only a small number of-- we'll still 262 00:12:10,595 --> 00:12:14,660 call them nodes, then if you realize you did something 263 00:12:14,660 --> 00:12:16,750 wrong, or something needs to be changed, 264 00:12:16,750 --> 00:12:19,926 can't they all go in and change that? 265 00:12:19,926 --> 00:12:21,920 PROFESSOR: So Hugo's raising the point, 266 00:12:21,920 --> 00:12:24,710 if you're down to, like, three nodes-- 267 00:12:24,710 --> 00:12:29,180 as I've shared with you all, the Australian Stock Exchange 268 00:12:29,180 --> 00:12:31,400 is putting a permissioned blockchain 269 00:12:31,400 --> 00:12:35,600 up using Digital Asset Holdings, which I think is backed by also 270 00:12:35,600 --> 00:12:38,660 the Hyperledger Fabric technology. 271 00:12:38,660 --> 00:12:41,030 But if it's just the Australian Stock Exchange, 272 00:12:41,030 --> 00:12:43,700 and it's one node, or three nodes, 273 00:12:43,700 --> 00:12:46,330 couldn't they just change it on all three? 274 00:12:46,330 --> 00:12:48,860 And I think, in essence, yes. 275 00:12:48,860 --> 00:12:52,820 I mean, that may not be what the database structure is, 276 00:12:52,820 --> 00:12:57,740 but I still think it's potentially yes. 277 00:12:57,740 --> 00:12:59,850 So as you get more concentration, 278 00:12:59,850 --> 00:13:01,640 you have more chance. 279 00:13:01,640 --> 00:13:03,800 But Brotish, I will try to research the Corda 280 00:13:03,800 --> 00:13:06,882 thing more specifically. 281 00:13:06,882 --> 00:13:09,480 Other thoughts? 282 00:13:09,480 --> 00:13:12,860 So then you create an auditable database-- some database 283 00:13:12,860 --> 00:13:15,960 with cryptography. 284 00:13:15,960 --> 00:13:18,840 Hash functions, which half the class said they understood 285 00:13:18,840 --> 00:13:21,096 and the other half was sort of, you 286 00:13:21,096 --> 00:13:29,450 know, a little rough on, all is in permissioned systems, 287 00:13:29,450 --> 00:13:33,470 using cryptographic schemes to ensure 288 00:13:33,470 --> 00:13:36,560 for the validity of the data and, so to speak, 289 00:13:36,560 --> 00:13:39,710 the immutable nature of the data. 290 00:13:39,710 --> 00:13:43,760 What's different is consensus protocols. 291 00:13:43,760 --> 00:13:46,640 In essence, it all goes down, except for Brotish's point, 292 00:13:46,640 --> 00:13:50,630 maybe, about Corda, as to who gets the chance 293 00:13:50,630 --> 00:13:53,490 to add the additional data. 294 00:13:53,490 --> 00:13:59,000 Is it a small club deal, or is it broad, wide open? 295 00:13:59,000 --> 00:14:02,300 Everybody together, roughly? 296 00:14:02,300 --> 00:14:04,200 So back to the technical features. 297 00:14:04,200 --> 00:14:07,910 Remember, it's a bunch of cryptography. 298 00:14:07,910 --> 00:14:11,850 Love it or hate it, it actually allows us all 299 00:14:11,850 --> 00:14:14,100 to use the internet-- well past, obviously, 300 00:14:14,100 --> 00:14:18,000 the blockchain that we're discussing in this class. 301 00:14:18,000 --> 00:14:22,720 Network consensus, and then all the ledgers. 302 00:14:22,720 --> 00:14:25,870 So both permissioned and permissionless have ledgers, 303 00:14:25,870 --> 00:14:28,240 have cryptography. 304 00:14:28,240 --> 00:14:31,170 It's that middle bucket that's different between the two. 305 00:14:33,732 --> 00:14:35,190 So what were some of the challenges 306 00:14:35,190 --> 00:14:40,390 that we talked about in blockchain? 307 00:14:40,390 --> 00:14:42,490 We just talked about them 10 minutes ago, which 308 00:14:42,490 --> 00:14:44,410 is, what's the list again? 309 00:14:44,410 --> 00:14:45,700 See, this is the easy part. 310 00:14:45,700 --> 00:14:50,210 If you haven't spoken yet, this is, like, the easy questions. 311 00:14:50,210 --> 00:14:53,200 Oh, I'm hopelessly shameless. 312 00:14:53,200 --> 00:14:55,450 Yes. 313 00:14:55,450 --> 00:14:57,920 Do you want to say your first name so Toledo takes you 314 00:14:57,920 --> 00:14:59,530 off the list? 315 00:14:59,530 --> 00:15:01,000 AUDIENCE: [INAUDIBLE] 316 00:15:01,000 --> 00:15:02,310 PROFESSOR: Did you get it? 317 00:15:02,310 --> 00:15:02,810 No. 318 00:15:02,810 --> 00:15:04,160 Do you want to say it louder? 319 00:15:04,160 --> 00:15:05,460 Because Thalita [INAUDIBLE]. 320 00:15:05,460 --> 00:15:09,986 [INTERPOSING VOICES] 321 00:15:09,986 --> 00:15:11,954 [LAUGHTER] 322 00:15:12,950 --> 00:15:18,160 AUDIENCE: [INAUDIBLE] issue about the scalability. 323 00:15:18,160 --> 00:15:19,125 PROFESSOR: Scalability. 324 00:15:19,125 --> 00:15:21,800 AUDIENCE: [INAUDIBLE] takes the time 325 00:15:21,800 --> 00:15:24,195 to create the next block about [INAUDIBLE].. 326 00:15:24,195 --> 00:15:25,070 PROFESSOR: All right. 327 00:15:25,070 --> 00:15:27,470 So scalability and some time. 328 00:15:27,470 --> 00:15:29,420 So efficiency and scalability. 329 00:15:29,420 --> 00:15:30,970 Anything else? 330 00:15:30,970 --> 00:15:31,790 AUDIENCE: Privacy? 331 00:15:31,790 --> 00:15:32,582 PROFESSOR: Privacy. 332 00:15:35,610 --> 00:15:37,710 So it's basically the scalability, 333 00:15:37,710 --> 00:15:43,140 the privacy are two big things that permissioned systems 334 00:15:43,140 --> 00:15:44,580 address. 335 00:15:44,580 --> 00:15:48,870 Interoperability-- basically, how does this blockchain 336 00:15:48,870 --> 00:15:51,600 talk to other blockchains, or how does this blockchain 337 00:15:51,600 --> 00:15:55,500 talk to other legacy systems? 338 00:15:55,500 --> 00:15:57,540 Permissioned and permissionless systems 339 00:15:57,540 --> 00:16:01,710 both have issues of interoperability. 340 00:16:01,710 --> 00:16:03,840 However, the smaller the club deal, 341 00:16:03,840 --> 00:16:08,070 the more likely that a bank or the Australian Stock Exchange 342 00:16:08,070 --> 00:16:10,920 might address its interoperability right within 343 00:16:10,920 --> 00:16:14,790 the system, whereas if it's a big open-sourced, 344 00:16:14,790 --> 00:16:16,800 open project-- 345 00:16:16,800 --> 00:16:21,480 so IBM would say, we can even help you with interoperability. 346 00:16:21,480 --> 00:16:24,450 IBM would say, we can help you with all four lines. 347 00:16:24,450 --> 00:16:27,060 We can help you with scalability, efficiency. 348 00:16:27,060 --> 00:16:29,298 We can help you with privacy. 349 00:16:29,298 --> 00:16:30,840 We can help you with interoperability 350 00:16:30,840 --> 00:16:31,500 and governance. 351 00:16:31,500 --> 00:16:35,460 I think it's less clear they can help with all four, 352 00:16:35,460 --> 00:16:39,390 but they can certainly help with the first two. 353 00:16:39,390 --> 00:16:43,140 And Alon, you've switched from using a permissionless 354 00:16:43,140 --> 00:16:44,750 to a permissioned system-- 355 00:16:44,750 --> 00:16:45,500 AUDIENCE: Correct. 356 00:16:45,500 --> 00:16:48,150 PROFESSOR: --in your startup, right? 357 00:16:48,150 --> 00:16:51,390 Which of these four is the reason why you switched? 358 00:16:51,390 --> 00:16:53,012 AUDIENCE: So actually, I switched-- 359 00:16:53,012 --> 00:16:54,220 PROFESSOR: Or something else? 360 00:16:54,220 --> 00:16:55,262 AUDIENCE: Something else. 361 00:16:55,262 --> 00:16:57,110 So for me, it was a business use case. 362 00:16:57,110 --> 00:16:58,080 So I was trying-- 363 00:16:58,080 --> 00:16:59,615 my business is selling to banks. 364 00:16:59,615 --> 00:17:01,990 PROFESSOR: So that's the next line-- commercial use case. 365 00:17:01,990 --> 00:17:04,310 AUDIENCE: There it is. 366 00:17:04,310 --> 00:17:06,420 PROFESSOR: There's a setup. 367 00:17:06,420 --> 00:17:08,214 So what was the reason that you switched? 368 00:17:08,214 --> 00:17:10,589 AUDIENCE: So I was building on Ethereum, which is public. 369 00:17:10,589 --> 00:17:14,579 And I thought that it's unlikely that, in the near future, 370 00:17:14,579 --> 00:17:16,170 banks are going to adopt Ethereum 371 00:17:16,170 --> 00:17:17,730 as their underlying technology. 372 00:17:17,730 --> 00:17:20,280 And then I learned about R3 and Corda, 373 00:17:20,280 --> 00:17:23,040 and learned that banks actually funded that project. 374 00:17:23,040 --> 00:17:26,845 So I switched to where the banks put their money. 375 00:17:26,845 --> 00:17:28,470 PROFESSOR: So I would contend that it's 376 00:17:28,470 --> 00:17:30,300 a bit about interoperability. 377 00:17:30,300 --> 00:17:33,240 You felt your users would be more 378 00:17:33,240 --> 00:17:37,470 likely to be able to work with your system 379 00:17:37,470 --> 00:17:41,640 if you were using Corda, with which they were familiar. 380 00:17:41,640 --> 00:17:42,570 AUDIENCE: Correct. 381 00:17:42,570 --> 00:17:45,745 So there's a business reason and technological reason. 382 00:17:45,745 --> 00:17:47,370 PROFESSOR: And then, of course, there's 383 00:17:47,370 --> 00:17:48,930 the public policy reasons. 384 00:17:48,930 --> 00:17:51,480 And IBM would even say that they're 385 00:17:51,480 --> 00:17:55,350 going to score higher points on the public policy 386 00:17:55,350 --> 00:17:57,000 if, for no other reason, that there's 387 00:17:57,000 --> 00:17:59,470 better privacy and security. 388 00:17:59,470 --> 00:18:02,710 Now, I'm not trying to shill for IBM. 389 00:18:02,710 --> 00:18:05,800 I'm just saying these would be their selling points-- 390 00:18:05,800 --> 00:18:07,816 or Corda, or elsewhere. 391 00:18:11,470 --> 00:18:13,810 We talked about Buterin's trilemma. 392 00:18:13,810 --> 00:18:16,720 But in another way, many people would 393 00:18:16,720 --> 00:18:21,760 say decentralization competes with scalability and security. 394 00:18:21,760 --> 00:18:23,630 If you want scalability and security, 395 00:18:23,630 --> 00:18:26,200 you can't have decentralization. 396 00:18:26,200 --> 00:18:29,320 I'm not a big believer in this trilemma, 397 00:18:29,320 --> 00:18:31,510 even though I've now raised it twice. 398 00:18:31,510 --> 00:18:36,250 But it's often talked about at blockchain conferences, 399 00:18:36,250 --> 00:18:39,490 and blockchain papers, and business discussions. 400 00:18:39,490 --> 00:18:43,690 So I've always told you I want to be a fair representation 401 00:18:43,690 --> 00:18:47,230 of the debate that's going on. 402 00:18:47,230 --> 00:18:52,390 I think it's more possible to solve these three over time 403 00:18:52,390 --> 00:18:55,150 together than some others. 404 00:18:55,150 --> 00:18:58,660 But maybe I'm just a cockeyed optimist about technology. 405 00:19:02,080 --> 00:19:04,300 So public policy framework. 406 00:19:04,300 --> 00:19:05,800 What were the big three slipstreams? 407 00:19:05,800 --> 00:19:07,670 So I went fast last time-- 408 00:19:07,670 --> 00:19:08,530 last class. 409 00:19:08,530 --> 00:19:11,846 Leonardo, what are you going to tell me about the-- 410 00:19:11,846 --> 00:19:14,090 it's because you were adjusting your glasses. 411 00:19:14,090 --> 00:19:15,880 AUDIENCE: Yeah. 412 00:19:15,880 --> 00:19:17,400 No, we spoke about the difficulty 413 00:19:17,400 --> 00:19:21,810 to implement framework. 414 00:19:21,810 --> 00:19:26,430 And you were talking more about the need [INAUDIBLE] 415 00:19:26,430 --> 00:19:29,617 offer security to protect people. 416 00:19:29,617 --> 00:19:31,200 PROFESSOR: To protect people-- so it's 417 00:19:31,200 --> 00:19:32,670 sort of a privacy thing. 418 00:19:32,670 --> 00:19:37,860 I think I have a hand up here from Catalina. 419 00:19:37,860 --> 00:19:41,260 AUDIENCE: There are three big things that [INAUDIBLE] public 420 00:19:41,260 --> 00:19:45,090 policy working against illicit activity-- 421 00:19:45,090 --> 00:19:48,030 PROFESSOR: Illicit activity. 422 00:19:48,030 --> 00:19:51,360 AUDIENCE: --financial instability, and protecting 423 00:19:51,360 --> 00:19:52,751 the investors. 424 00:19:52,751 --> 00:19:53,751 PROFESSOR: There you go. 425 00:19:56,200 --> 00:19:56,700 All right. 426 00:19:56,700 --> 00:19:57,960 So you know my style. 427 00:19:57,960 --> 00:20:00,390 I try to drop things into three buckets. 428 00:20:00,390 --> 00:20:03,810 But it's the only way I can remember anything. 429 00:20:03,810 --> 00:20:07,920 But it is the three big buckets, of course. 430 00:20:07,920 --> 00:20:11,028 And since I went fast last time-- 431 00:20:11,028 --> 00:20:13,070 and we're going to be coming back to a lot of it, 432 00:20:13,070 --> 00:20:15,047 but were there any questions that you have? 433 00:20:15,047 --> 00:20:16,380 And this is just an opportunity. 434 00:20:16,380 --> 00:20:17,348 Brotish? 435 00:20:17,348 --> 00:20:19,390 AUDIENCE: So I have a question on the [INAUDIBLE] 436 00:20:19,390 --> 00:20:21,320 stability point. 437 00:20:21,320 --> 00:20:24,600 Like, you mentioned that because of the world 438 00:20:24,600 --> 00:20:28,800 value of the digital currency is very minimal compared to, 439 00:20:28,800 --> 00:20:31,500 let's say, [INAUDIBLE]. 440 00:20:31,500 --> 00:20:35,700 PROFESSOR: The value of crypto finance 441 00:20:35,700 --> 00:20:39,750 is about $220 billion right now, compared to worldwide capital 442 00:20:39,750 --> 00:20:43,215 markets that, in aggregate, are over $300 trillion 443 00:20:43,215 --> 00:20:46,780 of debt, bonds, and equities. 444 00:20:46,780 --> 00:20:49,410 AUDIENCE: So my question was more like [INAUDIBLE] opinion 445 00:20:49,410 --> 00:20:53,880 on this, that this value is also concentrated within a limited 446 00:20:53,880 --> 00:20:56,970 number of people compared to the other assets that we are 447 00:20:56,970 --> 00:20:59,490 talking about, and hence-- 448 00:20:59,490 --> 00:21:02,610 I mean, it can still be important to regularize 449 00:21:02,610 --> 00:21:04,455 on the financial stability side, because it 450 00:21:04,455 --> 00:21:08,520 can have a effect which is not in proportion to just the size. 451 00:21:08,520 --> 00:21:10,920 PROFESSOR: So Brotish is asking the question that even 452 00:21:10,920 --> 00:21:14,460 though it's only $200 billion versus $300 453 00:21:14,460 --> 00:21:19,900 plus trillion of worldwide financial assets, 454 00:21:19,900 --> 00:21:23,860 couldn't it still be relevant to financial stability? 455 00:21:23,860 --> 00:21:26,360 And the answer is yes. 456 00:21:26,360 --> 00:21:29,830 But it's still relatively-- it's less than one one-thousandths 457 00:21:29,830 --> 00:21:31,960 of the broad size. 458 00:21:31,960 --> 00:21:33,260 AUDIENCE: [INAUDIBLE] 459 00:21:33,260 --> 00:21:34,520 PROFESSOR: Yes, please. 460 00:21:34,520 --> 00:21:36,560 Better you than me. 461 00:21:36,560 --> 00:21:39,100 AUDIENCE: So basically, we did an analysis between 462 00:21:39,100 --> 00:21:45,200 the correlation [INAUDIBLE] public market [INAUDIBLE] 463 00:21:45,200 --> 00:21:48,320 and basically there's a correlation of roughly 80% 464 00:21:48,320 --> 00:21:52,160 to 85% when the public-- 465 00:21:52,160 --> 00:21:55,820 when the Bitcoin market's down, within the past year, 466 00:21:55,820 --> 00:21:57,620 the public market's likely to lift up. 467 00:21:57,620 --> 00:22:00,150 Because there's a lot of leverage built into the system. 468 00:22:00,150 --> 00:22:04,070 And once the volatility kicks in, 469 00:22:04,070 --> 00:22:06,410 a lot of the investors [INAUDIBLE] 470 00:22:06,410 --> 00:22:08,590 they have to get money out from the public market, 471 00:22:08,590 --> 00:22:10,990 and that kind of feeds into a loop. 472 00:22:10,990 --> 00:22:12,990 PROFESSOR: So you're saying there's correlation, 473 00:22:12,990 --> 00:22:15,540 and there may be feedback loops. 474 00:22:15,540 --> 00:22:16,308 Tom? 475 00:22:16,308 --> 00:22:17,850 AUDIENCE: Sort of a related question. 476 00:22:17,850 --> 00:22:20,940 So Mervyn King was in this room a couple hours ago, 477 00:22:20,940 --> 00:22:23,430 and he was talking about, in the '08 crisis, 478 00:22:23,430 --> 00:22:27,270 two issues of trust-- one where there was a counterparty trust 479 00:22:27,270 --> 00:22:28,650 issue. 480 00:22:28,650 --> 00:22:30,750 Even though the overall derivatives market 481 00:22:30,750 --> 00:22:34,230 was [INAUDIBLE],, banks didn't individually 482 00:22:34,230 --> 00:22:36,640 know which other bank was most at risk. 483 00:22:36,640 --> 00:22:39,120 And so they wouldn't trade with each other 484 00:22:39,120 --> 00:22:40,590 or lend to one another. 485 00:22:40,590 --> 00:22:43,680 And then on the opposite side, the solution-- 486 00:22:43,680 --> 00:22:47,040 the salvation-- was two people in a room trusting 487 00:22:47,040 --> 00:22:49,290 each other that the central bank-- that the government 488 00:22:49,290 --> 00:22:53,100 would fund tens of billions of dollars [INAUDIBLE] capital. 489 00:22:53,100 --> 00:22:58,170 And so I wonder your thoughts on blockchain's role 490 00:22:58,170 --> 00:23:00,690 in addressing that first problem of knowing 491 00:23:00,690 --> 00:23:02,190 your counterparty, or at least being 492 00:23:02,190 --> 00:23:05,880 able to trust their position, and then the risk 493 00:23:05,880 --> 00:23:09,780 of eliminating the second chance of injection of capital 494 00:23:09,780 --> 00:23:13,730 into the financial system in a blockchain. 495 00:23:13,730 --> 00:23:14,938 PROFESSOR: Russ, [INAUDIBLE]. 496 00:23:14,938 --> 00:23:16,063 AUDIENCE: I had a question. 497 00:23:16,063 --> 00:23:17,010 It's related to this-- 498 00:23:17,010 --> 00:23:17,100 PROFESSOR: All right. 499 00:23:17,100 --> 00:23:18,175 AUDIENCE: --which is-- 500 00:23:18,175 --> 00:23:19,300 PROFESSOR: Thanks, Brotish. 501 00:23:19,300 --> 00:23:20,650 [INAUDIBLE] 502 00:23:20,650 --> 00:23:22,910 AUDIENCE: But the question is this. 503 00:23:22,910 --> 00:23:25,410 It does strike me that you only have a financial stability 504 00:23:25,410 --> 00:23:32,400 problem if people are relying on the value of people's 505 00:23:32,400 --> 00:23:35,310 Bitcoin holdings, for example, which is the bulk of crypto 506 00:23:35,310 --> 00:23:37,230 assets, right? 507 00:23:37,230 --> 00:23:39,480 The reason you have the problem during the crisis 508 00:23:39,480 --> 00:23:41,472 is you have all these banks. 509 00:23:41,472 --> 00:23:43,680 They're carrying these assets on their balance sheet. 510 00:23:43,680 --> 00:23:45,138 And all of the sudden, people think 511 00:23:45,138 --> 00:23:47,910 they don't know what they're worth, right? 512 00:23:47,910 --> 00:23:51,750 Who, if anyone, is actually carrying a Bitcoin assets 513 00:23:51,750 --> 00:23:54,607 on a balance sheet that someone else is relying on? 514 00:23:54,607 --> 00:23:56,940 Or, to put it differently, who's extending the leverage? 515 00:23:56,940 --> 00:23:57,370 PROFESSOR: All right. 516 00:23:57,370 --> 00:23:58,060 So let me-- 517 00:23:58,060 --> 00:23:59,970 AUDIENCE: [INAUDIBLE] 518 00:23:59,970 --> 00:24:01,950 PROFESSOR: Let me wrap these together. 519 00:24:01,950 --> 00:24:05,640 I think that what many central bankers and the Financial 520 00:24:05,640 --> 00:24:08,190 Stability Board would say, at 200 billion 521 00:24:08,190 --> 00:24:14,170 versus 300 trillion, it's probably not that financially 522 00:24:14,170 --> 00:24:16,540 relevant at this point. 523 00:24:16,540 --> 00:24:21,310 Where it could get relevant is, as Russ has sort of suggested, 524 00:24:21,310 --> 00:24:23,410 is if there's leverage behind it. 525 00:24:23,410 --> 00:24:26,975 If it's in a balance sheet, it's an asset on a balance sheet, 526 00:24:26,975 --> 00:24:28,810 and there's a liability on the other side. 527 00:24:28,810 --> 00:24:31,880 And thus it could bring down that entity, 528 00:24:31,880 --> 00:24:35,950 whether it's a hedge fund, whether it's a bank. 529 00:24:35,950 --> 00:24:40,090 But something that's relying [INAUDIBLE] to tie it 530 00:24:40,090 --> 00:24:43,020 back to the questions that Tom raised 531 00:24:43,020 --> 00:24:48,760 is, could blockchain lower systemic risk? 532 00:24:48,760 --> 00:24:53,320 Possibly, if it's a better database solution. 533 00:24:53,320 --> 00:24:56,590 If it answers-- what Mervyn King was speaking about, 534 00:24:56,590 --> 00:24:59,000 we also addressed in the late 90s. 535 00:24:59,000 --> 00:25:02,680 I remember quite well when Secretary Rubin 536 00:25:02,680 --> 00:25:04,640 called a number of us into his office and said, 537 00:25:04,640 --> 00:25:05,560 I don't understand. 538 00:25:05,560 --> 00:25:08,562 The banks can't tell us their exposure to Korea. 539 00:25:08,562 --> 00:25:09,520 A number of countries-- 540 00:25:09,520 --> 00:25:12,610 South Korea was about to have-- but they 541 00:25:12,610 --> 00:25:14,230 didn't default on their debt, but they 542 00:25:14,230 --> 00:25:17,200 were coming close to defaulting on their debt. 543 00:25:17,200 --> 00:25:19,480 And of course, it was only a short while 544 00:25:19,480 --> 00:25:22,210 before other countries, like Indonesia and Thailand, 545 00:25:22,210 --> 00:25:27,310 were getting into that same debt challenge. 546 00:25:27,310 --> 00:25:31,080 But why couldn't banks tell the US Department of Treasury 547 00:25:31,080 --> 00:25:31,705 their exposure? 548 00:25:31,705 --> 00:25:35,080 It's usually through derivatives. 549 00:25:35,080 --> 00:25:38,640 And derivatives, both in the late '90s 550 00:25:38,640 --> 00:25:41,230 and the financial crisis, were often 551 00:25:41,230 --> 00:25:45,340 something that propagated risk through the system. 552 00:25:45,340 --> 00:25:47,590 Now, the numbers were larger-- 553 00:25:47,590 --> 00:25:49,510 I mean, depending upon the time. 554 00:25:49,510 --> 00:25:51,280 The '90s was less than this. 555 00:25:51,280 --> 00:25:54,910 But it was $300, $400 trillion of notional amount 556 00:25:54,910 --> 00:25:56,390 of derivatives. 557 00:25:56,390 --> 00:26:00,040 So just the sheer notional size, even 558 00:26:00,040 --> 00:26:01,900 though the capital at risk in derivatives 559 00:26:01,900 --> 00:26:05,670 was much smaller because the leverage 560 00:26:05,670 --> 00:26:08,220 of this notional-- big, notional size. 561 00:26:08,220 --> 00:26:11,980 And the transparency was pretty low. 562 00:26:11,980 --> 00:26:13,180 It wasn't zero, actually. 563 00:26:13,180 --> 00:26:14,560 It wasn't zero. 564 00:26:14,560 --> 00:26:17,180 But it was really low. 565 00:26:17,180 --> 00:26:21,490 And so I think that's what Mervyn King would 566 00:26:21,490 --> 00:26:22,540 have been mentioning. 567 00:26:22,540 --> 00:26:26,800 And I do think blockchain can help in that transparency. 568 00:26:26,800 --> 00:26:28,720 But it takes a big collective action. 569 00:26:28,720 --> 00:26:31,990 And so Europe, the US, and Asia-- a lot of laws 570 00:26:31,990 --> 00:26:34,930 were passed to get more transparency in the derivatives 571 00:26:34,930 --> 00:26:36,370 space. 572 00:26:36,370 --> 00:26:39,220 Brotish, I would say that there could be problems. 573 00:26:39,220 --> 00:26:42,640 And I've had conversations with elected leaders that 574 00:26:42,640 --> 00:26:45,460 say, don't get lulled into the sense 575 00:26:45,460 --> 00:26:49,100 that it won't have financial stability issues. 576 00:26:49,100 --> 00:26:52,830 Because that was what happened in the 1980s and '90s 577 00:26:52,830 --> 00:26:55,940 when people said, well, derivatives will not 578 00:26:55,940 --> 00:27:01,210 create instability, because it's only 579 00:27:01,210 --> 00:27:05,080 the institutional investors, the sophisticated investors-- 580 00:27:05,080 --> 00:27:10,030 the so to speak big boys or big girls, you know? 581 00:27:10,030 --> 00:27:16,990 And I was part of those debates, that consensus that formed that 582 00:27:16,990 --> 00:27:19,690 it was only institutional and it wouldn't-- 583 00:27:19,690 --> 00:27:23,170 but there was a lot of leverage and a lack of transparency when 584 00:27:23,170 --> 00:27:24,490 big leverage-- 585 00:27:24,490 --> 00:27:26,890 multiple hundreds of trillions of dollars-- 586 00:27:26,890 --> 00:27:31,060 associated with it, of course, was part of the crisis-- 587 00:27:31,060 --> 00:27:32,500 not the only part of the crisis. 588 00:27:32,500 --> 00:27:35,430 But any other questions before Hugo? 589 00:27:35,430 --> 00:27:36,160 AUDIENCE: Yeah. 590 00:27:36,160 --> 00:27:38,752 On the protecting the investing public front, 591 00:27:38,752 --> 00:27:42,310 I was just wondering, like, a lot of people 592 00:27:42,310 --> 00:27:45,940 are now using Robinhood and zero fee things like that 593 00:27:45,940 --> 00:27:48,130 to invest in the stock market. 594 00:27:48,130 --> 00:27:51,400 And companies like Robinhood and Vanguard 595 00:27:51,400 --> 00:27:55,420 often sell order book data to high frequency traders 596 00:27:55,420 --> 00:27:57,880 so that they can kind of know what's going on 597 00:27:57,880 --> 00:28:00,250 and maybe front run. 598 00:28:00,250 --> 00:28:05,320 So I was wondering how that fits into what 599 00:28:05,320 --> 00:28:07,000 we were talking about last time where 600 00:28:07,000 --> 00:28:11,440 there's no transparency on many of the cryptocurrency 601 00:28:11,440 --> 00:28:12,400 exchanges. 602 00:28:12,400 --> 00:28:17,380 And then in addition to that, big institutional investors 603 00:28:17,380 --> 00:28:22,690 can, I think, buy upwards of 5% of any stock 604 00:28:22,690 --> 00:28:28,100 and then wait a few days before they have to report on that. 605 00:28:28,100 --> 00:28:31,540 So that can also have a huge effect on stock price. 606 00:28:31,540 --> 00:28:35,380 And then they can, afterwards, dump on people. 607 00:28:35,380 --> 00:28:38,830 So how is that different from what's going on 608 00:28:38,830 --> 00:28:40,120 in cryptocurrency exchanges? 609 00:28:40,120 --> 00:28:40,390 PROFESSOR: All right. 610 00:28:40,390 --> 00:28:41,980 So that's an investor protection one. 611 00:28:41,980 --> 00:28:43,630 Was there another one in the-- 612 00:28:43,630 --> 00:28:45,040 is this investor protection, or-- 613 00:28:45,040 --> 00:28:46,790 AUDIENCE: No, it's a little bit different. 614 00:28:46,790 --> 00:28:47,880 PROFESSOR: A little different. 615 00:28:47,880 --> 00:28:48,210 All right. 616 00:28:48,210 --> 00:28:49,350 Any other investor protection? 617 00:28:49,350 --> 00:28:50,610 Because I'm going to collect them and then [INAUDIBLE].. 618 00:28:50,610 --> 00:28:52,170 AUDIENCE: It's about the former. 619 00:28:52,170 --> 00:28:53,460 I was a little bit-- 620 00:28:53,460 --> 00:28:58,120 it was kind of [INAUDIBLE] to me when you say the blockchain can 621 00:28:58,120 --> 00:29:00,100 help to stabilize the financial markets-- 622 00:29:00,100 --> 00:29:00,210 PROFESSOR: All right. 623 00:29:00,210 --> 00:29:01,540 So that's back to financial stability. 624 00:29:01,540 --> 00:29:02,100 AUDIENCE: Yeah. 625 00:29:02,100 --> 00:29:02,460 If it can-- 626 00:29:02,460 --> 00:29:03,335 PROFESSOR: All right. 627 00:29:03,335 --> 00:29:03,990 Can I hold it? 628 00:29:03,990 --> 00:29:07,300 Let me just answer Hugo's investor protection one. 629 00:29:07,300 --> 00:29:13,010 So all markets-- not just crypto markets-- 630 00:29:13,010 --> 00:29:17,810 all markets are susceptible to some forms of front running. 631 00:29:17,810 --> 00:29:21,380 In essence, if you have a client relationship 632 00:29:21,380 --> 00:29:24,470 and you get information from your client that 633 00:29:24,470 --> 00:29:28,100 might affect the market pricing, you 634 00:29:28,100 --> 00:29:32,480 might jump ahead of that client with their information. 635 00:29:32,480 --> 00:29:36,200 Their information might be a buy order, a sell order. 636 00:29:36,200 --> 00:29:38,850 Or frankly, it might be some other information. 637 00:29:38,850 --> 00:29:42,170 But traditionally, if they have a buy order or a sell order, 638 00:29:42,170 --> 00:29:43,790 you have information. 639 00:29:43,790 --> 00:29:49,300 And then you're sort of jumping ahead and saying, all right. 640 00:29:49,300 --> 00:29:51,190 I'll buy or sell in front of them. 641 00:29:51,190 --> 00:29:53,710 That's the classic sort of front running, 642 00:29:53,710 --> 00:29:57,550 even though there's other methods. 643 00:29:57,550 --> 00:29:59,050 It's not supposed to happen. 644 00:29:59,050 --> 00:30:04,870 On regulated exchanges, you have some policing of it. 645 00:30:04,870 --> 00:30:06,640 Even before governments stood in, 646 00:30:06,640 --> 00:30:08,110 you could have some policing of it 647 00:30:08,110 --> 00:30:11,590 in self-regulatory organizations. 648 00:30:11,590 --> 00:30:16,600 In the crypto world, there is Katy bar the door. 649 00:30:16,600 --> 00:30:18,670 Anything can really happen. 650 00:30:18,670 --> 00:30:21,820 And it's my belief and understanding 651 00:30:21,820 --> 00:30:24,940 that many crypto exchanges-- not all of them. 652 00:30:24,940 --> 00:30:25,670 Not all of them. 653 00:30:25,670 --> 00:30:27,940 But many crypto exchanges are basically 654 00:30:27,940 --> 00:30:32,200 trading in front of their customers. 655 00:30:32,200 --> 00:30:35,590 And in fact, most crypto exchanges 656 00:30:35,590 --> 00:30:39,190 make markets, as they are both market makers-- 657 00:30:39,190 --> 00:30:42,430 meaning they are buying when you're selling and selling 658 00:30:42,430 --> 00:30:43,390 when you're buying. 659 00:30:43,390 --> 00:30:45,300 That's the nature of a market maker. 660 00:30:45,300 --> 00:30:48,490 It's very typical, very legal, very important 661 00:30:48,490 --> 00:30:49,750 function of market. 662 00:30:49,750 --> 00:30:52,720 But the exchanges are both market makers, 663 00:30:52,720 --> 00:30:54,790 and then they show order books. 664 00:30:54,790 --> 00:31:00,120 And so it sort of helps them do front running. 665 00:31:00,120 --> 00:31:03,030 Not that a lot of people necessarily agree with me, 666 00:31:03,030 --> 00:31:05,250 I think these markets would be better off 667 00:31:05,250 --> 00:31:08,850 if there were some forms of rules about front running 668 00:31:08,850 --> 00:31:11,070 and manipulation in markets and the like. 669 00:31:11,070 --> 00:31:13,320 On your 5% question, can we take it up 670 00:31:13,320 --> 00:31:14,760 in office hours or something? 671 00:31:14,760 --> 00:31:16,650 Because it's sort of outside of crypto. 672 00:31:16,650 --> 00:31:19,320 But I would be glad to talk about the SEC rules 673 00:31:19,320 --> 00:31:21,168 about the 5%. 674 00:31:21,168 --> 00:31:22,335 Is this investor protection? 675 00:31:22,335 --> 00:31:23,270 And that-- 676 00:31:23,270 --> 00:31:24,210 AUDIENCE: I think so. 677 00:31:24,210 --> 00:31:27,390 The reading that we had that talked about the Fabric 678 00:31:27,390 --> 00:31:29,010 technology, it mentioned something 679 00:31:29,010 --> 00:31:31,950 about execute order validate. 680 00:31:31,950 --> 00:31:34,740 Does that mechanism tie into the way 681 00:31:34,740 --> 00:31:36,750 that front running would work at all? 682 00:31:36,750 --> 00:31:38,650 PROFESSOR: All blockchain technology, 683 00:31:38,650 --> 00:31:42,810 whether it's permissioned, like Fabric, or permissionless, 684 00:31:42,810 --> 00:31:44,690 can-- 685 00:31:44,690 --> 00:31:47,280 if it was efficient and scalable-- 686 00:31:47,280 --> 00:31:51,480 could help out, because you could actually timestamp 687 00:31:51,480 --> 00:31:53,340 when did the order come in. 688 00:31:53,340 --> 00:31:56,310 When did the client's information 689 00:31:56,310 --> 00:32:00,180 get to the exchange, and is somebody standing in front? 690 00:32:00,180 --> 00:32:01,910 AUDIENCE: So it's sort of like a cascade 691 00:32:01,910 --> 00:32:05,450 that you can't necessarily pause once it starts? 692 00:32:05,450 --> 00:32:06,450 PROFESSOR: Right, right. 693 00:32:06,450 --> 00:32:10,470 So if you look at the algorithms at the New York Stock Exchange 694 00:32:10,470 --> 00:32:13,382 right now, which are not blockchain-- 695 00:32:13,382 --> 00:32:15,090 they're not permissioned, and they're not 696 00:32:15,090 --> 00:32:16,390 permissionless blockchain. 697 00:32:16,390 --> 00:32:19,740 But if you look at both the algorithms and the data flows, 698 00:32:19,740 --> 00:32:22,023 they have very good timestamping. 699 00:32:22,023 --> 00:32:23,440 I'm not going to say it's perfect, 700 00:32:23,440 --> 00:32:24,940 but they have very good timestamping 701 00:32:24,940 --> 00:32:28,440 to ensure when are orders-- 702 00:32:28,440 --> 00:32:32,630 basically message timestamping every message that comes in. 703 00:32:32,630 --> 00:32:36,300 And some order books, like the New York Stock Exchange-- 704 00:32:36,300 --> 00:32:41,280 some order books are price and time priority, 705 00:32:41,280 --> 00:32:45,960 price priority meaning high bid gets hit before the next bid, 706 00:32:45,960 --> 00:32:49,800 or lowest offer gets lifted first. 707 00:32:49,800 --> 00:32:51,720 That's price priority. 708 00:32:51,720 --> 00:32:53,670 But they might also have time priority 709 00:32:53,670 --> 00:32:57,570 for any bid that has the same price and any offer that's 710 00:32:57,570 --> 00:32:59,040 the same price. 711 00:32:59,040 --> 00:33:01,140 So they have to have very good timestamping. 712 00:33:05,730 --> 00:33:08,040 I truly believe you could not use a blockchain 713 00:33:08,040 --> 00:33:12,540 solution for the New York Stock Exchange order book right now. 714 00:33:12,540 --> 00:33:15,020 Whether you can in 10 years, I'm not sure. 715 00:33:15,020 --> 00:33:19,740 But time latency is so relevant in the middle of those order 716 00:33:19,740 --> 00:33:20,830 books-- 717 00:33:20,830 --> 00:33:22,860 you know, the nanoseconds that matter. 718 00:33:22,860 --> 00:33:24,715 I'm going to take this stability question. 719 00:33:24,715 --> 00:33:26,280 There were two stability-- and then 720 00:33:26,280 --> 00:33:30,610 we're going to move on to the rest of today's lecture. 721 00:33:30,610 --> 00:33:31,760 You're having fun now. 722 00:33:31,760 --> 00:33:33,580 AUDIENCE: Can you explain a little bit more 723 00:33:33,580 --> 00:33:37,060 about, like, well, say if blockchain 724 00:33:37,060 --> 00:33:40,630 can be a better database so it can help 725 00:33:40,630 --> 00:33:42,340 stabilize the financial market? 726 00:33:42,340 --> 00:33:44,110 Can you give some examples? 727 00:33:44,110 --> 00:33:46,750 PROFESSOR: So how blockchain could possibly 728 00:33:46,750 --> 00:33:49,030 help-- be a stabilizer. 729 00:33:49,030 --> 00:33:52,840 A lot of financial instability-- 730 00:33:52,840 --> 00:33:54,640 or, beyond the financial markets, 731 00:33:54,640 --> 00:33:57,970 instability is a question of resilience. 732 00:33:57,970 --> 00:34:00,550 And when things are centralized, you 733 00:34:00,550 --> 00:34:04,490 create single points of failure in any system-- 734 00:34:04,490 --> 00:34:08,020 really, in military or financial. 735 00:34:08,020 --> 00:34:12,940 You centralize something, you then, thus-- 736 00:34:12,940 --> 00:34:14,830 you know what to attack. 737 00:34:14,830 --> 00:34:17,860 You can also maybe have higher walls or better moats, 738 00:34:17,860 --> 00:34:20,199 but you still know of something to attack. 739 00:34:20,199 --> 00:34:23,080 So blockchain, in its decentralized nature, 740 00:34:23,080 --> 00:34:26,409 might be a more resilient database, 741 00:34:26,409 --> 00:34:29,239 because even if half or two thirds of it goes down, 742 00:34:29,239 --> 00:34:30,580 you still have the other third. 743 00:34:30,580 --> 00:34:33,080 I'm going to say something about the New York Stock Exchange 744 00:34:33,080 --> 00:34:33,969 again. 745 00:34:33,969 --> 00:34:35,920 Whether it's the New York Stock Exchange, the London Stock 746 00:34:35,920 --> 00:34:37,670 Exchange, the Chicago Mercantile Exchange, 747 00:34:37,670 --> 00:34:43,449 they all are required by their various local laws 748 00:34:43,449 --> 00:34:47,449 to have backup data centers. 749 00:34:47,449 --> 00:34:52,380 And those backup data centers even have to be lots of miles 750 00:34:52,380 --> 00:34:57,936 away if, God forbid, a bomb comes and takes out the center. 751 00:34:57,936 --> 00:34:59,250 AUDIENCE: [INAUDIBLE] 752 00:34:59,250 --> 00:35:00,210 PROFESSOR: What's that? 753 00:35:00,210 --> 00:35:02,830 It's called-- Its basically disaster recovery. 754 00:35:02,830 --> 00:35:05,550 So maybe blockchain will be more resilient. 755 00:35:05,550 --> 00:35:07,050 I'm going to take two more questions 756 00:35:07,050 --> 00:35:08,425 and then get back to permissioned 757 00:35:08,425 --> 00:35:09,900 versus permissionless. 758 00:35:09,900 --> 00:35:12,195 Way back in the corner-- your first name, again, is? 759 00:35:12,195 --> 00:35:12,820 AUDIENCE: Matt. 760 00:35:12,820 --> 00:35:13,290 PROFESSOR: Matt. 761 00:35:13,290 --> 00:35:14,010 Thank you, Matt. 762 00:35:14,010 --> 00:35:15,585 I should know that. 763 00:35:15,585 --> 00:35:17,790 AUDIENCE: So I'm kind of just curious how, 764 00:35:17,790 --> 00:35:20,175 essentially, the nasency, for lack of a better word-- 765 00:35:20,175 --> 00:35:21,300 PROFESSOR: The [INAUDIBLE]? 766 00:35:21,300 --> 00:35:22,920 AUDIENCE: Like, the nasency-- 767 00:35:22,920 --> 00:35:25,680 how kind of, like, new this market 768 00:35:25,680 --> 00:35:29,700 is and how that affects policy, because when you're shaping 769 00:35:29,700 --> 00:35:32,340 policy, and you don't necessarily 770 00:35:32,340 --> 00:35:35,150 have a bunch of years of knowing how people are going 771 00:35:35,150 --> 00:35:38,078 to react to that policy, I feel like it could almost 772 00:35:38,078 --> 00:35:39,370 be a chicken and egg situation. 773 00:35:39,370 --> 00:35:40,920 PROFESSOR: OK, I understand. 774 00:35:40,920 --> 00:35:43,020 Anybody else with the same theme? 775 00:35:43,020 --> 00:35:43,710 No? 776 00:35:43,710 --> 00:35:47,820 So the question is, how does it affect policy makers 777 00:35:47,820 --> 00:35:50,880 when a whole technology is new? 778 00:35:50,880 --> 00:35:55,510 And I would say, we have a lot of history with this. 779 00:35:55,510 --> 00:35:58,080 Whether it was railroads, or the telegraph, 780 00:35:58,080 --> 00:36:01,665 or the telephone, or television, we have a lot of history. 781 00:36:04,180 --> 00:36:07,150 Technology and commercial applications 782 00:36:07,150 --> 00:36:11,600 move ahead of the public sector. 783 00:36:11,600 --> 00:36:14,560 I mean, it's just inevitable. 784 00:36:14,560 --> 00:36:16,720 The official sector, the public sector-- 785 00:36:16,720 --> 00:36:19,420 unless it just basically does what Mark Carney says, 786 00:36:19,420 --> 00:36:22,720 and the choice is to isolate something, to sort of put up 787 00:36:22,720 --> 00:36:24,580 the moats of a society and says, we 788 00:36:24,580 --> 00:36:27,220 can't use that technology here. 789 00:36:27,220 --> 00:36:32,570 Unless you have that, technology usually supersedes the markets, 790 00:36:32,570 --> 00:36:36,080 and the markets and technology come before official sector. 791 00:36:36,080 --> 00:36:39,990 Depending upon the area, it can take 792 00:36:39,990 --> 00:36:43,710 single digit years and sometimes decades for public sector 793 00:36:43,710 --> 00:36:46,280 to catch up, literally. 794 00:36:46,280 --> 00:36:48,220 But let's take the internet. 795 00:36:48,220 --> 00:36:53,340 The internet was being worked on for 15ish years. 796 00:36:53,340 --> 00:36:57,030 But the protocol layer that helped take off 797 00:36:57,030 --> 00:37:01,290 was the worldwide web in 1991 or '92, and then 798 00:37:01,290 --> 00:37:05,210 the security protocols in '96. 799 00:37:05,210 --> 00:37:07,880 Just taking financial regulation, 800 00:37:07,880 --> 00:37:16,075 the SEC was asked in 1995, could bulletin boards-- 801 00:37:16,075 --> 00:37:18,540 there were electronic bulletin boards 802 00:37:18,540 --> 00:37:20,850 listing stocks and bonds-- 803 00:37:20,850 --> 00:37:25,890 be exempted from being considered stock exchanges? 804 00:37:25,890 --> 00:37:27,950 Well, the people that were asking that 805 00:37:27,950 --> 00:37:29,930 were the people putting up the bulletin boards. 806 00:37:29,930 --> 00:37:32,280 They didn't want to be regulated. 807 00:37:32,280 --> 00:37:35,670 The New York Stock Exchange that was fully regulated, 808 00:37:35,670 --> 00:37:37,500 and NASDAQ that was fully regulated, 809 00:37:37,500 --> 00:37:39,180 was coming at the other side and asking 810 00:37:39,180 --> 00:37:41,640 the SEC to shut them down. 811 00:37:41,640 --> 00:37:44,610 They didn't want the competition from the disruptors. 812 00:37:44,610 --> 00:37:49,480 It took three years for the SEC to answer that question. 813 00:37:49,480 --> 00:37:52,680 I don't mean answer it, like, with a letter. 814 00:37:52,680 --> 00:37:55,870 They had to propose a rule, and they had to do a final rule, 815 00:37:55,870 --> 00:37:59,840 and it was three years to do that. 816 00:37:59,840 --> 00:38:04,310 And blockchain, I think, we are through some 817 00:38:04,310 --> 00:38:06,050 of the big questions. 818 00:38:06,050 --> 00:38:09,005 We know how most jurisdictions tax-- 819 00:38:09,005 --> 00:38:12,770 T-A-X-- tax it as property and not as currency, 820 00:38:12,770 --> 00:38:14,660 and some of the tax issues. 821 00:38:14,660 --> 00:38:17,150 We know, in most jurisdictions, how 822 00:38:17,150 --> 00:38:19,970 it fits into, broadly, Bank Secrecy 823 00:38:19,970 --> 00:38:23,440 Act and illicit activity. 824 00:38:23,440 --> 00:38:25,570 There's very choppy implementation, 825 00:38:25,570 --> 00:38:27,520 very spotty implementation. 826 00:38:27,520 --> 00:38:30,160 I would say the investor protection side, 827 00:38:30,160 --> 00:38:31,720 we're at the early stages. 828 00:38:31,720 --> 00:38:34,330 And it's going to take, in most jurisdictions, another three 829 00:38:34,330 --> 00:38:38,570 years, maybe, to sort through some of how-- 830 00:38:38,570 --> 00:38:40,308 and this is just crypto finance. 831 00:38:40,308 --> 00:38:40,850 I don't know. 832 00:38:40,850 --> 00:38:42,530 Matt, does that-- 833 00:38:42,530 --> 00:38:44,470 I'm giving you some predictions. 834 00:38:44,470 --> 00:38:46,230 Some of it could be multiple decades. 835 00:38:46,230 --> 00:38:48,160 We're still, today, trying to figure out-- 836 00:38:48,160 --> 00:38:51,670 the public sector's figuring out how to regulate Facebook 837 00:38:51,670 --> 00:38:52,210 and Google. 838 00:38:52,210 --> 00:38:54,730 I'm going to take one last question, because I've 839 00:38:54,730 --> 00:38:55,720 got to do permissioned. 840 00:38:55,720 --> 00:38:58,560 AUDIENCE: [INAUDIBLE] because I was kind of, like-- 841 00:38:58,560 --> 00:39:01,300 it kind of answers the side following the technology. 842 00:39:01,300 --> 00:39:02,840 But I think, for me, what I was kind 843 00:39:02,840 --> 00:39:05,950 of wondering about was this, the way we're like-- as soon as you 844 00:39:05,950 --> 00:39:08,770 apply regulations to something that at least a good portion 845 00:39:08,770 --> 00:39:12,160 of the market seems to value being deregulated, 846 00:39:12,160 --> 00:39:14,620 it seems like a lot of the activity will change, 847 00:39:14,620 --> 00:39:15,780 either in scope or scale. 848 00:39:15,780 --> 00:39:18,730 PROFESSOR: So Matt's raising there's 849 00:39:18,730 --> 00:39:20,980 a trade-off about bringing something 850 00:39:20,980 --> 00:39:27,850 under regulation or into the public policy sphere. 851 00:39:27,850 --> 00:39:29,410 I'm probably just one voice of this, 852 00:39:29,410 --> 00:39:31,570 but I think it's probably true. 853 00:39:31,570 --> 00:39:33,970 There's very few economic activities 854 00:39:33,970 --> 00:39:38,860 that grow large that stay fully outside of the public policy 855 00:39:38,860 --> 00:39:41,560 framework of a society. 856 00:39:41,560 --> 00:39:44,230 It doesn't mean that public policy frameworks don't change, 857 00:39:44,230 --> 00:39:46,420 get adopted, get adapted. 858 00:39:46,420 --> 00:39:48,430 The internet comes along, and at first, it's 859 00:39:48,430 --> 00:39:50,650 a question of-- you know, with Amazon, 860 00:39:50,650 --> 00:39:52,450 is it going to be taxed or not? 861 00:39:52,450 --> 00:39:53,830 Is there sales tax? 862 00:39:53,830 --> 00:39:56,290 And then later-- you know, at first, it's not. 863 00:39:56,290 --> 00:39:57,570 Later, it is. 864 00:39:57,570 --> 00:40:00,490 And in some jurisdictions, questions on the internet 865 00:40:00,490 --> 00:40:02,760 was liability. 866 00:40:02,760 --> 00:40:06,040 And this was a very critical question of early internet was, 867 00:40:06,040 --> 00:40:09,370 was there liability for any of the information 868 00:40:09,370 --> 00:40:12,610 or the flow of information? 869 00:40:12,610 --> 00:40:16,180 I'm talking about libel laws and all the liability issues 870 00:40:16,180 --> 00:40:17,800 and so forth. 871 00:40:17,800 --> 00:40:21,700 And so in the US, there's a section of the law in 1996 872 00:40:21,700 --> 00:40:23,350 was passed. 873 00:40:23,350 --> 00:40:25,810 And now we're coming back and saying, wait a minute. 874 00:40:25,810 --> 00:40:29,560 That gave too broad-- it basically exempted the internet 875 00:40:29,560 --> 00:40:32,530 from liability as carriers. 876 00:40:32,530 --> 00:40:36,073 But now, that was modified in 2018-- 877 00:40:36,073 --> 00:40:40,100 22 years later-- to say, well, if it has to do, I think, 878 00:40:40,100 --> 00:40:44,080 with basically children trafficking and slavery 879 00:40:44,080 --> 00:40:47,860 and everything, maybe we should tighten that up a little bit. 880 00:40:50,560 --> 00:40:57,000 So I don't believe any broad economic activity 881 00:40:57,000 --> 00:41:04,155 is going to remain fully outside public policy frameworks. 882 00:41:04,155 --> 00:41:08,160 And I'm glad to be challenged on that. 883 00:41:08,160 --> 00:41:11,085 And I think blockchain, and particularly crypto finance, 884 00:41:11,085 --> 00:41:13,740 is at this grappling stage to get in. 885 00:41:13,740 --> 00:41:15,240 So let me move forward, because this 886 00:41:15,240 --> 00:41:17,160 is more about permissioned and permissionless. 887 00:41:17,160 --> 00:41:19,410 I just wanted to cover some of these. 888 00:41:19,410 --> 00:41:22,470 We will cover initial coin offerings and the public policy 889 00:41:22,470 --> 00:41:24,990 issues around initial coin offerings in the Howey Test. 890 00:41:24,990 --> 00:41:26,780 We will cover crypto exchanges, and we 891 00:41:26,780 --> 00:41:30,210 will cover central bank digital currencies a lot 892 00:41:30,210 --> 00:41:31,710 in the second half of this semester. 893 00:41:34,590 --> 00:41:38,810 So back to the trade-offs that we talked about already-- 894 00:41:38,810 --> 00:41:41,960 the trade-offs of centralization and decentralizations 895 00:41:41,960 --> 00:41:46,430 and Coase's work from the 1930s about the firm. 896 00:41:46,430 --> 00:41:47,960 This is the cost, remember. 897 00:41:47,960 --> 00:41:48,830 This is the cost. 898 00:41:48,830 --> 00:41:52,160 As we get centralized, there's more cost of economic rents, 899 00:41:52,160 --> 00:41:56,600 single points of failure, and capture. 900 00:41:56,600 --> 00:41:59,780 Some fragility, in a sense, in the system. 901 00:41:59,780 --> 00:42:03,990 But decentralization has its cost as well. 902 00:42:03,990 --> 00:42:07,340 You'll notice that these two lines are crossing somewhere. 903 00:42:07,340 --> 00:42:10,850 It was my failed attempt to say, you know, 904 00:42:10,850 --> 00:42:12,260 maybe there's a balance. 905 00:42:12,260 --> 00:42:16,130 Maybe overall, while some organizations 906 00:42:16,130 --> 00:42:18,530 will find all the way to decentralization, 907 00:42:18,530 --> 00:42:20,780 and some to centralization, I think 908 00:42:20,780 --> 00:42:24,350 economic systems tend more towards the centralized side 909 00:42:24,350 --> 00:42:26,090 of things. 910 00:42:26,090 --> 00:42:30,310 But you can change the slope of these two lines if you want. 911 00:42:30,310 --> 00:42:36,240 I'm just trying to give you a framework of thinking about it. 912 00:42:36,240 --> 00:42:39,390 So as we've talked about, the financial sector 913 00:42:39,390 --> 00:42:42,840 pretty much favors permissioned systems. 914 00:42:42,840 --> 00:42:47,590 The financial sector is saying, going back, no. 915 00:42:47,590 --> 00:42:52,170 Too many costs of coordination, governance, security, privacy, 916 00:42:52,170 --> 00:42:54,262 and scalability. 917 00:42:54,262 --> 00:42:55,470 We're over in the other side. 918 00:42:58,250 --> 00:43:02,390 And so that's where they are today. 919 00:43:02,390 --> 00:43:05,040 I'm not sure that's where they'll be in five or 10 years, 920 00:43:05,040 --> 00:43:06,870 but it is definitely where they are today. 921 00:43:09,800 --> 00:43:13,983 So let's go back down our list, in a sense. 922 00:43:13,983 --> 00:43:15,650 I'm going to have a bunch of check marks 923 00:43:15,650 --> 00:43:16,840 and x's on the right. 924 00:43:16,840 --> 00:43:20,580 Where do you think I'm going to have check marks and x's? 925 00:43:20,580 --> 00:43:22,600 This is the same three big buckets, 926 00:43:22,600 --> 00:43:24,040 because I can't think of anything. 927 00:43:24,040 --> 00:43:27,823 Emily, do you have a point of view? 928 00:43:27,823 --> 00:43:29,240 AUDIENCE: I mean, I think in terms 929 00:43:29,240 --> 00:43:32,480 of the permissioned technical features, one of them 930 00:43:32,480 --> 00:43:36,213 is obviously that it's not using that proof of work. 931 00:43:36,213 --> 00:43:37,130 PROFESSOR: Not using-- 932 00:43:37,130 --> 00:43:38,540 AUDIENCE: It's not using the same proof of work. 933 00:43:38,540 --> 00:43:39,998 PROFESSOR: Not using proof of work. 934 00:43:39,998 --> 00:43:45,040 So I had to give away-- you know, reveal the [INAUDIBLE].. 935 00:43:45,040 --> 00:43:51,430 All the cryptography that's used in permissionless systems 936 00:43:51,430 --> 00:43:53,310 is used in the permissioned systems. 937 00:43:53,310 --> 00:43:55,430 It might be used a little differently. 938 00:43:55,430 --> 00:43:59,080 I'm not going to say the Merkle trees are exactly the same. 939 00:43:59,080 --> 00:44:02,880 But all that broccoli we were eating a few lectures ago 940 00:44:02,880 --> 00:44:04,465 are relevant on both sides. 941 00:44:07,070 --> 00:44:09,415 Sorry, Alin. 942 00:44:09,415 --> 00:44:10,290 AUDIENCE: [INAUDIBLE] 943 00:44:10,290 --> 00:44:10,860 PROFESSOR: What's that? 944 00:44:10,860 --> 00:44:11,090 You-- 945 00:44:11,090 --> 00:44:11,796 AUDIENCE: I like broccoli. 946 00:44:11,796 --> 00:44:13,004 PROFESSOR: You like broccoli. 947 00:44:13,004 --> 00:44:15,630 A computer scientist speaking. 948 00:44:15,630 --> 00:44:18,300 Everything, though, about digital signatures and hash 949 00:44:18,300 --> 00:44:22,480 functions and so forth are relevant to both of these. 950 00:44:22,480 --> 00:44:24,870 But they're not necessarily relevant 951 00:44:24,870 --> 00:44:27,600 to every traditional database, which we'll 952 00:44:27,600 --> 00:44:31,320 talk about in 10 minutes or so. 953 00:44:31,320 --> 00:44:34,460 But as Emily said, there's really not-- 954 00:44:34,460 --> 00:44:35,150 I said, no. 955 00:44:35,150 --> 00:44:38,180 There's not decentralized network consensus. 956 00:44:38,180 --> 00:44:41,150 I'm stretching it a little bit, because permissioned systems 957 00:44:41,150 --> 00:44:42,600 can be decentralized. 958 00:44:42,600 --> 00:44:46,930 There could be five or 10 or 20 nodes. 959 00:44:46,930 --> 00:44:48,730 And so that is decentralized. 960 00:44:48,730 --> 00:44:51,160 So I shouldn't have really said no. 961 00:44:51,160 --> 00:44:53,560 I should have said maybe, or hybrid. 962 00:44:56,530 --> 00:44:59,250 And instead of proof of work, the permissioned systems 963 00:44:59,250 --> 00:45:02,160 use a bunch of consensus mechanisms. 964 00:45:02,160 --> 00:45:04,290 And I just listed a couple notary 965 00:45:04,290 --> 00:45:13,700 nodes or PBFT for Byzantine Fault Tolerance again. 966 00:45:13,700 --> 00:45:16,380 But they don't have native currencies. 967 00:45:16,380 --> 00:45:20,060 So that is a very big difference-- 968 00:45:20,060 --> 00:45:21,560 no native currency. 969 00:45:21,560 --> 00:45:24,860 If you have a solution for your final projects, 970 00:45:24,860 --> 00:45:27,878 or you have a solution for some entrepreneurial business 971 00:45:27,878 --> 00:45:29,420 that you're going to do in the future 972 00:45:29,420 --> 00:45:32,420 that you want a native currency, you're probably more over 973 00:45:32,420 --> 00:45:35,777 into the permissionless world. 974 00:45:35,777 --> 00:45:37,860 You want to have an incentive structure, something 975 00:45:37,860 --> 00:45:43,080 to motivate customers or users, or create token economics. 976 00:45:43,080 --> 00:45:47,290 Token economics is more in the permissionless. 977 00:45:47,290 --> 00:45:51,460 I say more because you can create a token even 978 00:45:51,460 --> 00:45:53,680 in the permissioned space. 979 00:45:53,680 --> 00:45:56,670 We'll talk about it in a moment. 980 00:45:56,670 --> 00:45:59,880 And then transaction script or UTXOs. 981 00:45:59,880 --> 00:46:02,700 They're not technically UTXOs in a bunch 982 00:46:02,700 --> 00:46:06,180 of permissioned systems, but you need some ledger. 983 00:46:06,180 --> 00:46:09,540 This last box, if I just called it ledgers, 984 00:46:09,540 --> 00:46:12,360 you'd still have ledgers in. 985 00:46:12,360 --> 00:46:15,570 So it's really the consensus mechanism in the middle, 986 00:46:15,570 --> 00:46:18,610 as Emily said. 987 00:46:18,610 --> 00:46:22,030 A couple of key design features. 988 00:46:22,030 --> 00:46:25,900 First, membership's limited to basically an authorized set 989 00:46:25,900 --> 00:46:27,552 of nodes. 990 00:46:27,552 --> 00:46:31,640 So does anybody want to say, if you were a bank, 991 00:46:31,640 --> 00:46:35,150 who you might-- you know, what might you do if you were 992 00:46:35,150 --> 00:46:37,940 a bank, and who might you authorize if you're doing 993 00:46:37,940 --> 00:46:38,560 a bank-- 994 00:46:38,560 --> 00:46:41,410 I don't know. 995 00:46:41,410 --> 00:46:46,030 Let's say that it was loans, or if it's 996 00:46:46,030 --> 00:46:47,640 Mark Snyderman's real estate. 997 00:46:47,640 --> 00:46:50,340 It's real estate loans. 998 00:46:50,340 --> 00:46:51,930 Could you ever see real estate loans 999 00:46:51,930 --> 00:46:55,672 going on a blockchain, Mark? 1000 00:46:55,672 --> 00:46:56,880 MARK SNYDERMAN: Loans, maybe. 1001 00:46:56,880 --> 00:46:57,900 Loans could. 1002 00:46:57,900 --> 00:46:58,140 PROFESSOR: All right. 1003 00:46:58,140 --> 00:46:58,650 So-- 1004 00:46:58,650 --> 00:47:01,463 MARK SNYDERMAN: Loans aren't traded as much. 1005 00:47:01,463 --> 00:47:02,880 PROFESSOR: So if real estate loans 1006 00:47:02,880 --> 00:47:06,080 were on a permissioned blockchain, 1007 00:47:06,080 --> 00:47:08,870 who might be that membership, the limited-- 1008 00:47:08,870 --> 00:47:14,040 who would-- it's not just a rhetorical question. 1009 00:47:14,040 --> 00:47:17,742 Who would actually care about the database of the loans? 1010 00:47:17,742 --> 00:47:19,200 MARK SNYDERMAN: Broker dealers that 1011 00:47:19,200 --> 00:47:23,280 are active in loans, institutional investors that 1012 00:47:23,280 --> 00:47:24,380 are active in loans. 1013 00:47:27,730 --> 00:47:29,230 PROFESSOR: So the broker dealers who 1014 00:47:29,230 --> 00:47:31,563 are actually trading the loans, and maybe the investors. 1015 00:47:31,563 --> 00:47:32,075 Alon? 1016 00:47:32,075 --> 00:47:32,700 AUDIENCE: Yeah. 1017 00:47:32,700 --> 00:47:36,338 Just rating agencies when you securitize those loans-- 1018 00:47:36,338 --> 00:47:37,380 AUDIENCE: Loan servicers. 1019 00:47:37,380 --> 00:47:37,760 AUDIENCE: Loan servicers. 1020 00:47:37,760 --> 00:47:39,440 PROFESSOR: Amanda, loan servicers. 1021 00:47:39,440 --> 00:47:42,620 So maybe loan servicers, maybe the rating agencies. 1022 00:47:42,620 --> 00:47:47,870 So it's basically, who has a need and needs that data? 1023 00:47:47,870 --> 00:47:50,450 I'm not sure that this is the right community, 1024 00:47:50,450 --> 00:47:52,670 but it's the discussion I'm thinking about. 1025 00:47:52,670 --> 00:47:54,425 And again, as you're thinking about-- 1026 00:47:54,425 --> 00:47:57,290 I'm jumping again to a little bit final projects, but when 1027 00:47:57,290 --> 00:48:01,700 you're thinking about who actually needs data 1028 00:48:01,700 --> 00:48:06,140 and who has a reason to amend the data, 1029 00:48:06,140 --> 00:48:09,060 or write to the ledger-- 1030 00:48:09,060 --> 00:48:10,980 because if you don't have a desire and a need 1031 00:48:10,980 --> 00:48:18,300 to actually amend or change the data, move value in the system, 1032 00:48:18,300 --> 00:48:22,190 you might not need an open database. 1033 00:48:22,190 --> 00:48:22,690 Please. 1034 00:48:22,690 --> 00:48:24,190 AUDIENCE: When we were talking about 1035 00:48:24,190 --> 00:48:27,910 that additional layer of blocks, is that layer also-- 1036 00:48:27,910 --> 00:48:33,152 can you change the membership with that layer? 1037 00:48:33,152 --> 00:48:35,110 PROFESSOR: I'm not sure I follow your question. 1038 00:48:35,110 --> 00:48:37,726 When you say an additional layer of blocks-- 1039 00:48:37,726 --> 00:48:39,670 [INTERPOSING VOICES] 1040 00:48:39,670 --> 00:48:40,860 PROFESSOR: Oh, I'm sorry. 1041 00:48:40,860 --> 00:48:41,505 Layer 2. 1042 00:48:41,505 --> 00:48:42,260 AUDIENCE: Yeah. 1043 00:48:42,260 --> 00:48:44,220 Can you alter the-- 1044 00:48:44,220 --> 00:48:48,750 because it's a different level of refinement of information 1045 00:48:48,750 --> 00:48:50,310 that's stored. 1046 00:48:50,310 --> 00:48:53,940 PROFESSOR: So what Kelly is asking is, 1047 00:48:53,940 --> 00:48:57,270 we've talked about how to make permissionless 1048 00:48:57,270 --> 00:49:00,240 blockchains more scalable by having a layer 1049 00:49:00,240 --> 00:49:02,400 2, like the Lightning Network. 1050 00:49:02,400 --> 00:49:04,800 That's what you're referencing. 1051 00:49:04,800 --> 00:49:09,430 So in permissionless systems, it's open to everyone, 1052 00:49:09,430 --> 00:49:13,860 even though if you're opening up an individual payment 1053 00:49:13,860 --> 00:49:15,730 channel in the Lightning Network, 1054 00:49:15,730 --> 00:49:17,550 it's really just two counterparties 1055 00:49:17,550 --> 00:49:19,680 opening up a channel. 1056 00:49:19,680 --> 00:49:22,620 So in a sense, you've already narrowed the scope, 1057 00:49:22,620 --> 00:49:24,390 because you might just have a payment 1058 00:49:24,390 --> 00:49:26,790 channel between two parties-- 1059 00:49:26,790 --> 00:49:29,460 some side chains or multiple parties. 1060 00:49:29,460 --> 00:49:31,650 Lightning Network and payment channels 1061 00:49:31,650 --> 00:49:34,627 tend to be two parties. 1062 00:49:34,627 --> 00:49:35,460 So I don't know if-- 1063 00:49:35,460 --> 00:49:36,660 was that what you're asking? 1064 00:49:36,660 --> 00:49:38,610 AUDIENCE: That network in and of itself 1065 00:49:38,610 --> 00:49:42,308 is a mechanism of membership. 1066 00:49:45,060 --> 00:49:47,940 PROFESSOR: The layer 2 can be a membership. 1067 00:49:47,940 --> 00:49:49,200 That is correct. 1068 00:49:49,200 --> 00:49:52,800 But it's a technology that's available to everybody. 1069 00:49:52,800 --> 00:49:58,010 So it has attributes where you can open up bilateral channels. 1070 00:49:58,010 --> 00:49:58,510 James. 1071 00:49:58,510 --> 00:49:59,050 AUDIENCE: I was just going to say, 1072 00:49:59,050 --> 00:50:00,800 for permissioned blockchains, couldn't you 1073 00:50:00,800 --> 00:50:04,330 imagine that the layer 1 would be, say, the Fed 1074 00:50:04,330 --> 00:50:07,125 with all of the big commercial banks? 1075 00:50:07,125 --> 00:50:08,500 But in effect, if you're thinking 1076 00:50:08,500 --> 00:50:09,917 about a different layer, you could 1077 00:50:09,917 --> 00:50:12,180 have another layer on top of that, which 1078 00:50:12,180 --> 00:50:13,310 [INAUDIBLE] central banks. 1079 00:50:13,310 --> 00:50:13,890 So you could naturally-- 1080 00:50:13,890 --> 00:50:14,280 PROFESSOR: All right. 1081 00:50:14,280 --> 00:50:16,072 AUDIENCE: --have two layers of deals with-- 1082 00:50:16,072 --> 00:50:17,770 PROFESSOR: So where we're headed is, 1083 00:50:17,770 --> 00:50:21,970 is do permissioned systems have the same need 1084 00:50:21,970 --> 00:50:26,080 to have a second layer as permissionless? 1085 00:50:26,080 --> 00:50:28,000 And even if they don't have the same needs, 1086 00:50:28,000 --> 00:50:30,550 might it actually provide something? 1087 00:50:30,550 --> 00:50:33,220 So I would say, I don't think they have the same need 1088 00:50:33,220 --> 00:50:36,460 to have a second layer, because they're more scalable, 1089 00:50:36,460 --> 00:50:39,430 they're more efficient right now. 1090 00:50:39,430 --> 00:50:43,230 And they're already closed clubs. 1091 00:50:43,230 --> 00:50:46,260 But even if they don't have the same need, 1092 00:50:46,260 --> 00:50:49,230 they might actually want to put a second layer on top. 1093 00:50:49,230 --> 00:50:53,350 And some of them actually do this right in their technology. 1094 00:50:53,350 --> 00:50:57,510 And it's what I put up here as the second and third bullet 1095 00:50:57,510 --> 00:50:58,390 point. 1096 00:50:58,390 --> 00:51:01,620 The transactions can be limited to only authorized 1097 00:51:01,620 --> 00:51:03,240 known participants. 1098 00:51:03,240 --> 00:51:08,932 So in many permissioned systems, it's a one broad technology. 1099 00:51:08,932 --> 00:51:14,450 It might only be 20 people that can authorize transactions. 1100 00:51:14,450 --> 00:51:18,380 But now in some transactions, it will only be Anton and May. 1101 00:51:18,380 --> 00:51:22,940 And in other transactions, it will be Alpha and Amanda. 1102 00:51:22,940 --> 00:51:25,550 So I might not be able to-- 1103 00:51:25,550 --> 00:51:27,590 by the way, Alpha might be Goldman Sachs, 1104 00:51:27,590 --> 00:51:30,110 and Amanda might be Barclays. 1105 00:51:30,110 --> 00:51:32,150 And then it might make more sense. 1106 00:51:32,150 --> 00:51:35,230 And Mark might be Fidelity. 1107 00:51:35,230 --> 00:51:38,530 And I don't know, am I the US Department of Treasury 1108 00:51:38,530 --> 00:51:39,030 in this one? 1109 00:51:42,800 --> 00:51:44,780 Mark and I both started out in finance. 1110 00:51:44,780 --> 00:51:48,740 I just went off to public service later. 1111 00:51:48,740 --> 00:51:54,830 But so Corda and Hyperledger and many of the private blockchains 1112 00:51:54,830 --> 00:51:59,210 literally allow for partitioning right in the blockchain. 1113 00:51:59,210 --> 00:52:00,980 I don't know if you'd need to put a layer 1114 00:52:00,980 --> 00:52:05,630 2 on top, because they allow for partitioning and segregated 1115 00:52:05,630 --> 00:52:09,060 pools of authority inside of the blockchain. 1116 00:52:09,060 --> 00:52:09,560 I'm sorry? 1117 00:52:09,560 --> 00:52:11,120 There was a question, I thought. 1118 00:52:11,120 --> 00:52:12,374 Oh. 1119 00:52:12,374 --> 00:52:13,210 Oh, yeah. 1120 00:52:13,210 --> 00:52:15,440 AUDIENCE: [INAUDIBLE] you see insurance companies, 1121 00:52:15,440 --> 00:52:17,773 like title insurance, as part of this. 1122 00:52:17,773 --> 00:52:18,940 PROFESSOR: Can you speak up? 1123 00:52:18,940 --> 00:52:19,457 [INAUDIBLE] 1124 00:52:19,457 --> 00:52:21,790 AUDIENCE: Insurance companies for title for the houses-- 1125 00:52:21,790 --> 00:52:22,670 PROFESSOR: Right. 1126 00:52:22,670 --> 00:52:25,010 AUDIENCE: --I think blockchain could be really helpful, 1127 00:52:25,010 --> 00:52:28,250 because sometimes you have the trust issue around who 1128 00:52:28,250 --> 00:52:31,160 owned the house before you, or the title where 1129 00:52:31,160 --> 00:52:33,420 it goes backwards in history. 1130 00:52:33,420 --> 00:52:37,700 So I think title insurance companies 1131 00:52:37,700 --> 00:52:40,220 would be part of this blockchain in terms of [INAUDIBLE].. 1132 00:52:40,220 --> 00:52:43,910 PROFESSOR: So your point is, is that titling of real estate 1133 00:52:43,910 --> 00:52:49,065 could be an important part of blockchains, and as [AUDIO OUT] 1134 00:52:49,065 --> 00:52:50,440 MARK SNYDERMAN: Probably someday, 1135 00:52:50,440 --> 00:52:54,850 but every little town has its own system for keeping property 1136 00:52:54,850 --> 00:52:56,200 records. 1137 00:52:56,200 --> 00:53:00,850 And getting them all to cooperate in a blockchain 1138 00:53:00,850 --> 00:53:05,233 kind of solution seems remote. 1139 00:53:05,233 --> 00:53:07,150 They don't think they have a real problem that 1140 00:53:07,150 --> 00:53:07,990 needs solving. 1141 00:53:07,990 --> 00:53:10,640 So yeah, maybe someday. 1142 00:53:10,640 --> 00:53:13,540 But I doubt it will happen anytime soon. 1143 00:53:13,540 --> 00:53:17,240 But it makes some logical sense. 1144 00:53:17,240 --> 00:53:20,460 And that would probably be a permissionless system, 1145 00:53:20,460 --> 00:53:23,880 because there's not a huge amount of times a piece 1146 00:53:23,880 --> 00:53:27,480 of property transacts, right? 1147 00:53:27,480 --> 00:53:31,260 So it's not like moving money around the world. 1148 00:53:31,260 --> 00:53:35,063 It's a much simpler, less frequent event. 1149 00:53:35,063 --> 00:53:35,730 AUDIENCE: Yeah-- 1150 00:53:35,730 --> 00:53:38,010 MARK SNYDERMAN: And land records are public. 1151 00:53:38,010 --> 00:53:42,710 People often have to go to the town offices to look at them. 1152 00:53:42,710 --> 00:53:45,900 AUDIENCE: Yeah, you have to pay a lot of transaction fees, 1153 00:53:45,900 --> 00:53:49,595 whether you're transacting on a plant or anything that involves 1154 00:53:49,595 --> 00:53:51,660 real estate here in the US. 1155 00:53:51,660 --> 00:53:54,420 You have to pay considerable amount of fees 1156 00:53:54,420 --> 00:53:56,910 to the title insurance companies. 1157 00:53:56,910 --> 00:53:59,550 And in my mind, that could eliminate that role 1158 00:53:59,550 --> 00:54:05,720 if you have blockchain and [INAUDIBLE] where the system. 1159 00:54:05,720 --> 00:54:08,580 MARK SNYDERMAN: Maybe, but you're paying money 1160 00:54:08,580 --> 00:54:10,742 because you're asking-- 1161 00:54:10,742 --> 00:54:11,700 AUDIENCE: Somebody to-- 1162 00:54:11,700 --> 00:54:14,820 MARK SNYDERMAN: --them and lawyers 1163 00:54:14,820 --> 00:54:16,665 to make sure the title's free and clear. 1164 00:54:16,665 --> 00:54:17,550 AUDIENCE: [INAUDIBLE] 1165 00:54:17,550 --> 00:54:19,967 MARK SNYDERMAN: And there's all sorts of claims on a title 1166 00:54:19,967 --> 00:54:20,580 to property-- 1167 00:54:20,580 --> 00:54:23,900 somebody that fixed the roof, or the utility company 1168 00:54:23,900 --> 00:54:28,050 that has an easement through the middle to put lines 1169 00:54:28,050 --> 00:54:30,280 and so forth. 1170 00:54:30,280 --> 00:54:33,450 So there are some complications that 1171 00:54:33,450 --> 00:54:37,090 are different than just transferring money. 1172 00:54:37,090 --> 00:54:42,572 PROFESSOR: So to bring it back, to broaden it out, 1173 00:54:42,572 --> 00:54:44,060 the que-- it's Rahim? 1174 00:54:44,060 --> 00:54:44,927 AUDIENCE: Rahem. 1175 00:54:44,927 --> 00:54:45,635 PROFESSOR: Rahem. 1176 00:54:45,635 --> 00:54:47,760 Rahem's question is, is, well, what 1177 00:54:47,760 --> 00:54:49,680 about real estate and title? 1178 00:54:49,680 --> 00:54:50,880 Might that be appropriate? 1179 00:54:50,880 --> 00:54:55,350 And as Mark has given us a sense of, yes, it might be. 1180 00:54:55,350 --> 00:54:58,620 But we're back to that sort of thing of the collective action 1181 00:54:58,620 --> 00:55:01,680 issues that we've talked about in previous classes-- 1182 00:55:01,680 --> 00:55:07,140 the collective action of many municipal authorities 1183 00:55:07,140 --> 00:55:10,485 that keep the land records, that keep the real estate records. 1184 00:55:10,485 --> 00:55:15,120 Yeah, it might help, but why do I need to do this? 1185 00:55:15,120 --> 00:55:16,320 And it's also a [INAUDIBLE]. 1186 00:55:16,320 --> 00:55:20,740 It's a low volume, low transaction. 1187 00:55:20,740 --> 00:55:23,350 And yet, it could be an enormous benefit, 1188 00:55:23,350 --> 00:55:25,630 because there's something called title insurance. 1189 00:55:25,630 --> 00:55:28,090 And title insurance trying to clean up the title 1190 00:55:28,090 --> 00:55:32,030 and making sure that something is free and clear 1191 00:55:32,030 --> 00:55:35,520 could be recorded in the future. 1192 00:55:35,520 --> 00:55:37,620 So I'll be a long-term optimist. 1193 00:55:37,620 --> 00:55:40,410 I don't see this one being solved in the next handful 1194 00:55:40,410 --> 00:55:42,480 of years-- single digit years-- 1195 00:55:42,480 --> 00:55:45,000 particularly because of the [INAUDIBLE] collective action 1196 00:55:45,000 --> 00:55:45,940 issues. 1197 00:55:45,940 --> 00:55:46,440 I'm sorry. 1198 00:55:46,440 --> 00:55:47,815 We have somebody here who's going 1199 00:55:47,815 --> 00:55:52,990 to take the other side before I give my conclusionary estimate. 1200 00:55:52,990 --> 00:55:53,490 Yes? 1201 00:55:53,490 --> 00:55:54,030 AUDIENCE: Yes. 1202 00:55:54,030 --> 00:55:55,113 That's exactly my startup. 1203 00:55:55,113 --> 00:55:59,055 The first phase is collecting all the data 1204 00:55:59,055 --> 00:56:02,583 from all those munis departmentals [INAUDIBLE].. 1205 00:56:02,583 --> 00:56:04,000 The information is there publicly. 1206 00:56:04,000 --> 00:56:08,820 So yes, they will not probably immediately adopt a Corda node 1207 00:56:08,820 --> 00:56:10,050 and record their stuff. 1208 00:56:10,050 --> 00:56:13,860 But you can publicly pool that information 1209 00:56:13,860 --> 00:56:16,650 and then use that information and record that information 1210 00:56:16,650 --> 00:56:17,770 on the blockchain. 1211 00:56:17,770 --> 00:56:21,720 And then once I become stronger and bigger, 1212 00:56:21,720 --> 00:56:23,815 then they will probably have to adopt it. 1213 00:56:23,815 --> 00:56:24,690 PROFESSOR: All right. 1214 00:56:24,690 --> 00:56:25,850 [LAUGHTER] 1215 00:56:25,850 --> 00:56:27,900 So maybe I'll move up my estimate 1216 00:56:27,900 --> 00:56:30,660 from double digit years to single digit years. 1217 00:56:30,660 --> 00:56:35,527 But I think it's going to be, with all respect, some time. 1218 00:56:35,527 --> 00:56:36,860 Pria, then I'm going to move on. 1219 00:56:36,860 --> 00:56:39,200 AUDIENCE: So I used to work at Habitat 1220 00:56:39,200 --> 00:56:40,530 for Humanity International. 1221 00:56:40,530 --> 00:56:43,190 So access to proper land rights or land 1222 00:56:43,190 --> 00:56:45,590 title, that was a big issue for us overseas, 1223 00:56:45,590 --> 00:56:47,480 not so much in the US. 1224 00:56:47,480 --> 00:56:49,400 And so I could see a real application there, 1225 00:56:49,400 --> 00:56:52,430 where in several countries where there is no land record, 1226 00:56:52,430 --> 00:56:55,370 or there are no title systems, or they're buried 1227 00:56:55,370 --> 00:56:57,630 in layers of obfuscation. 1228 00:56:57,630 --> 00:56:59,990 This is a great solution to get something 1229 00:56:59,990 --> 00:57:05,000 like that started in those contexts, where there is-- 1230 00:57:05,000 --> 00:57:08,330 establishing the property right of record 1231 00:57:08,330 --> 00:57:10,563 could be of immeasurable social value. 1232 00:57:10,563 --> 00:57:12,350 PROFESSOR: You know, I agree, and I 1233 00:57:12,350 --> 00:57:15,080 think there's been a tremendous amount of literature 1234 00:57:15,080 --> 00:57:18,840 in the last 24 months around whether blockchain 1235 00:57:18,840 --> 00:57:25,100 can help free up a lot of illiquid capital and assets. 1236 00:57:25,100 --> 00:57:30,350 I'm just saying I'm probably closer to Mark's thinking 1237 00:57:30,350 --> 00:57:31,928 than to Alon's thinking. 1238 00:57:31,928 --> 00:57:33,470 I think it's going to take some time. 1239 00:57:33,470 --> 00:57:35,330 AUDIENCE: But depending on the context, right? 1240 00:57:35,330 --> 00:57:35,997 In the context-- 1241 00:57:35,997 --> 00:57:40,580 PROFESSOR: Context, the country, the system, the legal system, 1242 00:57:40,580 --> 00:57:44,150 how decentralized, and how tough the collective action 1243 00:57:44,150 --> 00:57:45,380 issues are. 1244 00:57:45,380 --> 00:57:48,560 So just going back to private blockchains, 1245 00:57:48,560 --> 00:57:51,290 technical features-- let's go back. 1246 00:57:51,290 --> 00:57:55,790 Membership limited to authorized nodes. 1247 00:57:55,790 --> 00:57:58,910 Transactions can be partitioned as well. 1248 00:57:58,910 --> 00:58:00,260 It's kind of Kelly's point. 1249 00:58:00,260 --> 00:58:04,040 Right in the technology, you can partition and segregate 1250 00:58:04,040 --> 00:58:05,300 information. 1251 00:58:05,300 --> 00:58:07,760 So data and ledgers can be partitioned 1252 00:58:07,760 --> 00:58:11,090 because transactions can be partitioned. 1253 00:58:11,090 --> 00:58:13,160 That doesn't mean you can't have a layer 2. 1254 00:58:13,160 --> 00:58:15,830 I'm just saying there's a lot less need for a layer 2, 1255 00:58:15,830 --> 00:58:19,590 because they do it in the data structures itself. 1256 00:58:19,590 --> 00:58:24,900 The consensus is permissioned private protocols. 1257 00:58:24,900 --> 00:58:26,310 They do have a consensus. 1258 00:58:26,310 --> 00:58:28,320 Somebody has to agree on the next block. 1259 00:58:31,230 --> 00:58:35,080 But it's a tight, limited group. 1260 00:58:35,080 --> 00:58:38,260 It does use cryptography, but often, there's 1261 00:58:38,260 --> 00:58:40,990 something called a registration authority. 1262 00:58:40,990 --> 00:58:46,690 The registration authority is helping mask data. 1263 00:58:46,690 --> 00:58:50,560 So they address privacy two ways. 1264 00:58:50,560 --> 00:58:53,830 They address privacy because it's a smaller group. 1265 00:58:53,830 --> 00:58:56,710 They can actually see the whole network. 1266 00:58:56,710 --> 00:59:00,090 But even within the network, even on the network, 1267 00:59:00,090 --> 00:59:03,890 they're further addressing privacy 1268 00:59:03,890 --> 00:59:09,590 that Alpha's and my transaction, Amanda maybe 1269 00:59:09,590 --> 00:59:12,350 can't even see, even though she's on the network, 1270 00:59:12,350 --> 00:59:14,330 because it's encrypted. 1271 00:59:14,330 --> 00:59:16,670 And then there's sort of what's called a registration 1272 00:59:16,670 --> 00:59:23,910 authority, or authority within the system, that can unmask it. 1273 00:59:23,910 --> 00:59:24,410 Yes. 1274 00:59:24,410 --> 00:59:26,243 AUDIENCE: I just have a question about the-- 1275 00:59:26,243 --> 00:59:32,330 how is a segregated system a subgroup of two [INAUDIBLE] 1276 00:59:32,330 --> 00:59:33,810 different from a layer? 1277 00:59:33,810 --> 00:59:37,520 If [INAUDIBLE] is already doing subnetting off of calculations, 1278 00:59:37,520 --> 00:59:39,500 or whatever it may be-- transactions-- 1279 00:59:39,500 --> 00:59:42,015 and then the net [INAUDIBLE]-- 1280 00:59:42,015 --> 00:59:42,890 PROFESSOR: All right. 1281 00:59:42,890 --> 00:59:45,057 So I've got another question I have to follow up on. 1282 00:59:45,057 --> 00:59:47,810 I have to follow up on Brotish's Corda question, 1283 00:59:47,810 --> 00:59:50,030 and now James' question about, well, wait a minute. 1284 00:59:50,030 --> 00:59:53,368 How is this partitioning different than a layer 2? 1285 00:59:53,368 --> 00:59:55,160 And it's beyond my knowledge, but I'm going 1286 00:59:55,160 --> 00:59:58,460 to see if I can get it for you. 1287 00:59:58,460 --> 01:00:03,560 That's a good-- they're both-- they're good questions. 1288 01:00:03,560 --> 01:00:07,460 And then smart contracts-- yes, smart contracts 1289 01:00:07,460 --> 01:00:14,600 can happen on these systems as well as permissionless systems. 1290 01:00:14,600 --> 01:00:16,820 IBM-- they say they use chaincode. 1291 01:00:16,820 --> 01:00:20,570 But they say chaincode really could be any language, at least 1292 01:00:20,570 --> 01:00:24,745 they advertise, and no native currency. 1293 01:00:24,745 --> 01:00:26,120 So that's kind of the technical-- 1294 01:00:26,120 --> 01:00:27,440 not deeply technical. 1295 01:00:27,440 --> 01:00:29,120 It's not like learning hash functions. 1296 01:00:29,120 --> 01:00:32,360 But yes to cryptography. 1297 01:00:32,360 --> 01:00:35,150 Well, they partition the ledgers. 1298 01:00:35,150 --> 01:00:37,100 The consensus is closed loop. 1299 01:00:37,100 --> 01:00:40,950 Yes, they have smart contracts. 1300 01:00:40,950 --> 01:00:42,920 But they can even have additional privacy. 1301 01:00:42,920 --> 01:00:45,030 But it comes at a cost. 1302 01:00:45,030 --> 01:00:47,840 There's an authority that has to protect 1303 01:00:47,840 --> 01:00:51,900 something-- a registration authority inside of it. 1304 01:00:51,900 --> 01:00:53,310 Oh, I forgot. 1305 01:00:53,310 --> 01:00:54,870 The code is generally open source. 1306 01:00:54,870 --> 01:00:56,700 It does not have to be open source. 1307 01:00:56,700 --> 01:00:59,430 Hyperledger is open source, but it 1308 01:00:59,430 --> 01:01:00,850 doesn't have to be open source. 1309 01:01:03,570 --> 01:01:06,030 So this was from one of the readings. 1310 01:01:06,030 --> 01:01:08,530 I'm just putting up that chart that was in the reading. 1311 01:01:08,530 --> 01:01:09,960 And I just found it helpful. 1312 01:01:09,960 --> 01:01:10,860 It's a different way. 1313 01:01:10,860 --> 01:01:12,090 It's not Gensler's way. 1314 01:01:12,090 --> 01:01:14,660 It's somebody else's way of thinking about Ethereum, 1315 01:01:14,660 --> 01:01:16,870 Hyperledger, and Corda. 1316 01:01:16,870 --> 01:01:18,480 I'm only using Hyperledger and Corda 1317 01:01:18,480 --> 01:01:22,170 because they're two of the biggest private platforms. 1318 01:01:22,170 --> 01:01:23,220 There are many others. 1319 01:01:26,860 --> 01:01:28,600 Different programming language-- you 1320 01:01:28,600 --> 01:01:32,120 might say, from a business point of view, who cares? 1321 01:01:32,120 --> 01:01:35,810 And to some level, you're probably right. 1322 01:01:35,810 --> 01:01:37,940 But you're not completely right. 1323 01:01:37,940 --> 01:01:41,120 I mean, there's probably more that people can write 1324 01:01:41,120 --> 01:01:43,190 on top of Hyperledger Fabric. 1325 01:01:43,190 --> 01:01:49,470 They say more developers could work on that than Solidity. 1326 01:01:49,470 --> 01:01:52,760 Governance-- that's the governance 1327 01:01:52,760 --> 01:01:55,250 of the system itself. 1328 01:01:55,250 --> 01:01:57,320 They all can do smart contracts. 1329 01:01:57,320 --> 01:02:00,110 We've talked about consensus. 1330 01:02:00,110 --> 01:02:02,740 And the scalability is much harder. 1331 01:02:02,740 --> 01:02:05,870 Ethereum, remember, almost crashed on CryptoKitties 1332 01:02:05,870 --> 01:02:08,220 last December. 1333 01:02:08,220 --> 01:02:11,820 And I shared that story about one initial coin offering 1334 01:02:11,820 --> 01:02:15,900 was 30% of the Ethereum network on the one day 1335 01:02:15,900 --> 01:02:19,995 it was closing [INAUDIBLE]. 1336 01:02:19,995 --> 01:02:25,190 DTCC, in a given day, can do as much as 100 1337 01:02:25,190 --> 01:02:28,960 million transactions in a day. 1338 01:02:28,960 --> 01:02:33,820 That's the Depository Trust Corporation. 1339 01:02:33,820 --> 01:02:37,460 Ethereum does about a million and a half transactions a day, 1340 01:02:37,460 --> 01:02:42,040 and Bitcoin, about 400,000 to 500,000 transactions a day. 1341 01:02:42,040 --> 01:02:49,120 So you know, we've still got a ways to go on this scalability. 1342 01:02:49,120 --> 01:02:51,340 So what about traditional databases? 1343 01:02:51,340 --> 01:02:53,800 I think that to talk about permissioned 1344 01:02:53,800 --> 01:02:56,780 versus permissionless, a lot of people, 1345 01:02:56,780 --> 01:02:58,540 even some of my colleagues here at MIT, 1346 01:02:58,540 --> 01:03:01,430 says, well, if you're talking about Hyperledger, Corda, 1347 01:03:01,430 --> 01:03:06,220 and Fabric and Corda, that's like Oracle. 1348 01:03:06,220 --> 01:03:08,350 That's just a traditional database. 1349 01:03:08,350 --> 01:03:11,280 That's not really blockchain. 1350 01:03:11,280 --> 01:03:13,390 The Bitcoin and blockchain purist 1351 01:03:13,390 --> 01:03:20,160 would say, if it doesn't have Nakamoto consensus, forget it. 1352 01:03:20,160 --> 01:03:22,490 It's not a member of the club. 1353 01:03:22,490 --> 01:03:26,240 That's not how I've chosen to teach that this class, 1354 01:03:26,240 --> 01:03:27,710 as you know. 1355 01:03:27,710 --> 01:03:29,960 I think both are relevant, possibly, 1356 01:03:29,960 --> 01:03:32,750 to business solutions, and it's worthwhile to think of them. 1357 01:03:32,750 --> 01:03:35,950 But so what separates-- 1358 01:03:35,950 --> 01:03:38,390 and this is not in the reading, so I'm just going to, 1359 01:03:38,390 --> 01:03:39,800 you know, kind of-- 1360 01:03:39,800 --> 01:03:43,640 but what separates it versus a traditional database? 1361 01:03:43,640 --> 01:03:46,590 Well, back to the basics. 1362 01:03:46,590 --> 01:03:50,160 Permissioned private blockchains have append-only logs. 1363 01:03:50,160 --> 01:03:52,320 You're adding-- I know. 1364 01:03:52,320 --> 01:03:54,390 Thank you, Brotish. 1365 01:03:54,390 --> 01:03:59,190 But basically, they have append-only logs. 1366 01:03:59,190 --> 01:04:02,880 Traditional databases-- and I'm at the edge of my knowledge-- 1367 01:04:02,880 --> 01:04:03,870 you can create. 1368 01:04:03,870 --> 01:04:04,710 You can read. 1369 01:04:04,710 --> 01:04:05,550 You can update. 1370 01:04:05,550 --> 01:04:08,263 You can delete. 1371 01:04:08,263 --> 01:04:12,490 Again, I'm taking the mainstream traditional databases, 1372 01:04:12,490 --> 01:04:15,430 or what's called CRUD, if you wish-- 1373 01:04:15,430 --> 01:04:22,610 C-R-U-D. Whereas these, you always are adding 1374 01:04:22,610 --> 01:04:25,010 to the database. 1375 01:04:25,010 --> 01:04:28,790 Two, these databases use cryptography 1376 01:04:28,790 --> 01:04:30,530 to have commitment schemes. 1377 01:04:30,530 --> 01:04:33,590 You may remember, what is a hash function? 1378 01:04:33,590 --> 01:04:36,350 It not only compresses data, but it means 1379 01:04:36,350 --> 01:04:39,430 you're committing to the data. 1380 01:04:39,430 --> 01:04:42,040 And when you finish The New York Times crossword puzzle, 1381 01:04:42,040 --> 01:04:43,960 you can actually send it in. 1382 01:04:43,960 --> 01:04:45,880 And if it's identical to-- 1383 01:04:45,880 --> 01:04:47,140 and hash it. 1384 01:04:47,140 --> 01:04:49,420 And if you remember our discussion about The New York 1385 01:04:49,420 --> 01:04:53,110 Times crossword puzzle, if it's hashed, and the hashes match, 1386 01:04:53,110 --> 01:04:54,250 voilá. 1387 01:04:54,250 --> 01:04:57,610 You solved The New York Times crossword puzzle. 1388 01:04:57,610 --> 01:05:02,480 So you can have commitment schemes for the data. 1389 01:05:02,480 --> 01:05:04,190 And it can be distributed. 1390 01:05:04,190 --> 01:05:06,220 Yes, Joequinn. 1391 01:05:06,220 --> 01:05:09,040 AUDIENCE: So even if it's append-only, 1392 01:05:09,040 --> 01:05:11,700 if you take away the proof of work, 1393 01:05:11,700 --> 01:05:16,930 you can rewrite it in a very quick manner. 1394 01:05:16,930 --> 01:05:19,700 PROFESSOR: So Joequinn's raising a very important point. 1395 01:05:19,700 --> 01:05:23,940 So even if it's append-only, you get rid of proof of work, 1396 01:05:23,940 --> 01:05:26,700 can't somebody go in and rewrite it? 1397 01:05:26,700 --> 01:05:33,480 Permissioned blockchains make a different trade-off on trust. 1398 01:05:33,480 --> 01:05:36,630 Permissionless system say, we're going to address trust 1399 01:05:36,630 --> 01:05:38,850 through this proof of work. 1400 01:05:38,850 --> 01:05:42,480 And at least 51% of all of the network 1401 01:05:42,480 --> 01:05:46,370 has to, in essence, agree and validate, and the like. 1402 01:05:46,370 --> 01:05:47,970 Permissioned basically are saying, 1403 01:05:47,970 --> 01:05:53,780 I'm going to trust the 10 or 15 or 20 1404 01:05:53,780 --> 01:05:58,750 nodes that are in the system can validate against each other. 1405 01:05:58,750 --> 01:06:03,980 And so there's something about the club that's 1406 01:06:03,980 --> 01:06:06,260 authorized in the network. 1407 01:06:06,260 --> 01:06:08,390 But they're still using append-only. 1408 01:06:08,390 --> 01:06:10,610 You're right that within those 20, 1409 01:06:10,610 --> 01:06:13,250 somebody could try to rush through and do 1410 01:06:13,250 --> 01:06:16,190 a whole entire new append-only. 1411 01:06:16,190 --> 01:06:22,345 But then within the club, it's exposed. 1412 01:06:22,345 --> 01:06:24,640 AUDIENCE: If the club agrees that the last two 1413 01:06:24,640 --> 01:06:29,110 days have to be rewritten, they can do it in a very easy way. 1414 01:06:29,110 --> 01:06:30,100 PROFESSOR: Correct. 1415 01:06:30,100 --> 01:06:32,650 But I would contend that even in Bitcoin, 1416 01:06:32,650 --> 01:06:36,490 if 10,000 nodes decided to rewrite the last two days, 1417 01:06:36,490 --> 01:06:38,220 they could. 1418 01:06:38,220 --> 01:06:38,720 It's a-- 1419 01:06:38,720 --> 01:06:41,265 AUDIENCE: Taken a lot of time, because they have-- 1420 01:06:41,265 --> 01:06:41,890 PROFESSOR: Yes. 1421 01:06:41,890 --> 01:06:42,837 AUDIENCE: --to find-- 1422 01:06:42,837 --> 01:06:43,920 PROFESSOR: That's correct. 1423 01:06:43,920 --> 01:06:44,580 AUDIENCE: [INAUDIBLE] 1424 01:06:44,580 --> 01:06:45,663 PROFESSOR: That's correct. 1425 01:06:45,663 --> 01:06:47,940 So what Joequinn's saying is in Bitcoin, 1426 01:06:47,940 --> 01:06:51,270 what protects against that is partly the proof of work, 1427 01:06:51,270 --> 01:06:55,320 and partly that it takes 10,000, or at least 51% of them, 1428 01:06:55,320 --> 01:06:58,740 probably, to rewrite the last two days. 1429 01:06:58,740 --> 01:07:00,690 The economic incentives in Bitcoin 1430 01:07:00,690 --> 01:07:04,980 is, is why waste all your computer power and electricity 1431 01:07:04,980 --> 01:07:09,880 and so forth doing that when you may as well just go forward? 1432 01:07:09,880 --> 01:07:14,050 I think Bitcoin could be susceptible to state actors. 1433 01:07:14,050 --> 01:07:17,010 I think, you know, if North Korea wanted-- 1434 01:07:17,010 --> 01:07:21,210 or Russia wanted-- to spend probably single digit billions, 1435 01:07:21,210 --> 01:07:24,870 but at most maybe 10 billion, they could overwhelm the whole 1436 01:07:24,870 --> 01:07:27,270 system and bring-- 1437 01:07:27,270 --> 01:07:29,790 they could do a 51% attack by buying up 1438 01:07:29,790 --> 01:07:34,700 enough mining equipment, the ASICs, 1439 01:07:34,700 --> 01:07:37,850 and just bigfooting the whole system. 1440 01:07:37,850 --> 01:07:40,800 But it's multiple billio-- and there's academic papers on this 1441 01:07:40,800 --> 01:07:41,300 now. 1442 01:07:41,300 --> 01:07:42,920 It's single digit billions. 1443 01:07:42,920 --> 01:07:45,740 A state actor could-- 1444 01:07:45,740 --> 01:07:47,330 an individual wouldn't want to do it, 1445 01:07:47,330 --> 01:07:51,680 because you'd crush $110 billion market cap down to [INAUDIBLE].. 1446 01:07:51,680 --> 01:07:53,885 So that would be kind of unwise. 1447 01:07:57,860 --> 01:08:02,300 This is a very critical part of what a private blockchain-- 1448 01:08:02,300 --> 01:08:04,850 to me, economically, what a private blockchain 1449 01:08:04,850 --> 01:08:09,200 can do and facilitate better than a traditional database. 1450 01:08:09,200 --> 01:08:11,630 I'm not saying this list is the only list. 1451 01:08:11,630 --> 01:08:14,990 I'm not even saying this list is completely the right list. 1452 01:08:14,990 --> 01:08:17,600 This is Gary Gensler's list trying 1453 01:08:17,600 --> 01:08:20,600 to say, what's the business that separates 1454 01:08:20,600 --> 01:08:22,340 traditional blockchains? 1455 01:08:22,340 --> 01:08:25,540 Append-only timestamping. 1456 01:08:25,540 --> 01:08:29,020 Some cryptography and schemes that 1457 01:08:29,020 --> 01:08:35,080 makes the data immutable so you can put it in a ledger-- 1458 01:08:35,080 --> 01:08:37,630 so it's back to ledgers and so forth-- 1459 01:08:37,630 --> 01:08:40,069 and thus, finality of settlement. 1460 01:08:40,069 --> 01:08:43,660 Does anybody want to remind the class what settlement means? 1461 01:08:46,859 --> 01:08:47,510 Anybody? 1462 01:08:47,510 --> 01:08:49,990 Take a shot. 1463 01:08:49,990 --> 01:08:52,147 Where's my accountant? 1464 01:08:52,147 --> 01:08:55,270 Aviva's not here. 1465 01:08:55,270 --> 01:08:56,150 Oh, please. 1466 01:08:58,705 --> 01:09:02,633 AUDIENCE: [INAUDIBLE] do with the UTXO. 1467 01:09:02,633 --> 01:09:04,970 PROFESSOR: No, forget about UTXO. 1468 01:09:04,970 --> 01:09:06,859 Just tell me what settlement means. 1469 01:09:06,859 --> 01:09:09,865 AUDIENCE: Oh, like, making off your debit and credit. 1470 01:09:09,865 --> 01:09:12,229 PROFESSOR: Yeah, debits and credits. 1471 01:09:12,229 --> 01:09:18,420 Settlement means changing the data in a ledger, 1472 01:09:18,420 --> 01:09:21,720 and doing it with finality. 1473 01:09:21,720 --> 01:09:26,120 In essence, do I have $10 or $11? 1474 01:09:26,120 --> 01:09:30,569 A payment settlement system is when you finally say, 1475 01:09:30,569 --> 01:09:33,390 I've got $11, no longer $10. 1476 01:09:33,390 --> 01:09:38,930 And Amanda's got $9 instead of $10. 1477 01:09:38,930 --> 01:09:39,649 You went down. 1478 01:09:39,649 --> 01:09:40,409 I went up. 1479 01:09:40,409 --> 01:09:41,170 Is that all right? 1480 01:09:41,170 --> 01:09:42,050 AUDIENCE: Yeah. 1481 01:09:42,050 --> 01:09:43,870 PROFESSOR: Thank you. 1482 01:09:43,870 --> 01:09:46,210 You'll hear a lot about clearing and settlement. 1483 01:09:46,210 --> 01:09:48,880 The word settlement means to finally change 1484 01:09:48,880 --> 01:09:52,029 the data in a ledger, and that we're not 1485 01:09:52,029 --> 01:09:53,830 going to come back to it. 1486 01:09:53,830 --> 01:09:54,750 It's done. 1487 01:09:54,750 --> 01:09:56,680 It's final. 1488 01:09:56,680 --> 01:09:58,600 If we're moving something of value-- 1489 01:09:58,600 --> 01:10:01,580 and the value could be grapes. 1490 01:10:01,580 --> 01:10:03,080 The value could be diamonds. 1491 01:10:03,080 --> 01:10:04,490 The value can be real estate. 1492 01:10:04,490 --> 01:10:05,780 The value can be money. 1493 01:10:05,780 --> 01:10:08,000 But if we're moving something of value 1494 01:10:08,000 --> 01:10:16,440 and storing it on a ledger, it's more adaptable to blockchain. 1495 01:10:16,440 --> 01:10:18,960 If you're doing some database that 1496 01:10:18,960 --> 01:10:21,840 has nothing to do with things of value and ledger, 1497 01:10:21,840 --> 01:10:26,250 you don't need final immutable settlement. 1498 01:10:26,250 --> 01:10:28,530 I would contend you probably have less reason 1499 01:10:28,530 --> 01:10:30,420 to use any of this blockchain stuff. 1500 01:10:32,940 --> 01:10:33,610 Zan. 1501 01:10:33,610 --> 01:10:35,620 AUDIENCE: Can you maybe talk a little bit 1502 01:10:35,620 --> 01:10:38,050 about a real-world implementation of this? 1503 01:10:38,050 --> 01:10:41,470 So I saw recently that Walmart is forcing their suppliers 1504 01:10:41,470 --> 01:10:45,840 to basically add themselves into this private blockchain 1505 01:10:45,840 --> 01:10:48,198 to basically track their supply chain. 1506 01:10:48,198 --> 01:10:49,990 I'm having a really hard time understanding 1507 01:10:49,990 --> 01:10:52,240 why that needs to be on a blockchain, 1508 01:10:52,240 --> 01:10:54,010 whether permissioned or nonpermissioned, 1509 01:10:54,010 --> 01:10:56,920 and why that needs to be. 1510 01:10:56,920 --> 01:10:59,500 PROFESSOR: So the question is, why is Walmart 1511 01:10:59,500 --> 01:11:02,290 putting supply chain with a bunch of agricultural products 1512 01:11:02,290 --> 01:11:03,910 on a blockchain? 1513 01:11:03,910 --> 01:11:08,255 Agricultural products are something of value, 1514 01:11:08,255 --> 01:11:13,020 whether it's corn, wheat, grapes. 1515 01:11:13,020 --> 01:11:14,490 It's something of value. 1516 01:11:14,490 --> 01:11:18,700 And when it moves from one owner to another owner, 1517 01:11:18,700 --> 01:11:20,200 if you keep that on a ledger, you 1518 01:11:20,200 --> 01:11:23,350 can record that Amanda has the grapes, 1519 01:11:23,350 --> 01:11:26,450 and Gary no longer has the grapes. 1520 01:11:26,450 --> 01:11:31,770 And another thing that blockchains help with 1521 01:11:31,770 --> 01:11:36,030 is they lower the cost of reconciling multiple databases. 1522 01:11:36,030 --> 01:11:39,780 We could keep-- as in agriculture-- 1523 01:11:39,780 --> 01:11:44,280 we can keep separate ledgers, separate databases. 1524 01:11:44,280 --> 01:11:48,720 The farmers keeping their database. 1525 01:11:48,720 --> 01:11:50,970 The wholesaler with the grain elevators 1526 01:11:50,970 --> 01:11:54,240 are keeping their database, and up the supply chain, 1527 01:11:54,240 --> 01:11:56,280 from the farm, to the grain elevator, 1528 01:11:56,280 --> 01:11:58,950 to the millers and merchants, all the way 1529 01:11:58,950 --> 01:12:04,560 to General Mills, all the way to Kroger and the store. 1530 01:12:04,560 --> 01:12:08,400 Right now, those are multiple databases that, by and large, 1531 01:12:08,400 --> 01:12:10,470 don't have to communicate. 1532 01:12:10,470 --> 01:12:14,070 So the question is, will Walmart get to the place 1533 01:12:14,070 --> 01:12:15,900 where that's a good idea? 1534 01:12:15,900 --> 01:12:22,030 But I'm just pointing out if you're moving things of value 1535 01:12:22,030 --> 01:12:25,870 where you want to keep final settlement that you know who 1536 01:12:25,870 --> 01:12:29,650 owns that thing of value in ledgers, 1537 01:12:29,650 --> 01:12:33,620 and lastly, if you want to lower any reconciliation-- 1538 01:12:37,840 --> 01:12:41,950 if you have multiple ledgers and things of value, 1539 01:12:41,950 --> 01:12:47,170 blockchain, I think, is more valuable. 1540 01:12:47,170 --> 01:12:49,448 Zan's asking, well, is that Walmart? 1541 01:12:49,448 --> 01:12:49,990 I don't know. 1542 01:12:49,990 --> 01:12:51,090 We're going to find out. 1543 01:12:51,090 --> 01:12:53,148 Tom, and I've got about seven minutes to finish. 1544 01:12:53,148 --> 01:12:54,690 AUDIENCE: There was a talk last year. 1545 01:12:54,690 --> 01:12:57,070 Somebody [INAUDIBLE] came in and talked about the-- 1546 01:12:57,070 --> 01:12:59,050 they were then piloting this program. 1547 01:12:59,050 --> 01:13:01,780 But the example they used was, in E. coli outbreak 1548 01:13:01,780 --> 01:13:05,830 in their spinach supply, rather than taking a whole supplier, 1549 01:13:05,830 --> 01:13:08,050 an entire wholesaler offline and recalling 1550 01:13:08,050 --> 01:13:10,300 all of that product, through the blockchain, 1551 01:13:10,300 --> 01:13:12,760 they could trace it to a particular farm 1552 01:13:12,760 --> 01:13:16,030 and a single point of origin almost instantaneously 1553 01:13:16,030 --> 01:13:18,352 and just take off a much smaller portion of it. 1554 01:13:18,352 --> 01:13:19,560 PROFESSOR: So let me do this. 1555 01:13:19,560 --> 01:13:20,320 Let me just-- hold on. 1556 01:13:20,320 --> 01:13:21,885 I know we've got two questions here. 1557 01:13:21,885 --> 01:13:24,940 Let me just hold on, just for-- is that all right? 1558 01:13:24,940 --> 01:13:25,440 All right. 1559 01:13:25,440 --> 01:13:26,660 We got permission. 1560 01:13:26,660 --> 01:13:30,140 So this was what we talked about earlier-- 1561 01:13:30,140 --> 01:13:34,430 Coase's trade-offs of costs. 1562 01:13:34,430 --> 01:13:36,943 Let me try another one, which is going 1563 01:13:36,943 --> 01:13:41,310 to be my shot at blockchains and traditional databases. 1564 01:13:41,310 --> 01:13:43,980 So one is access control. 1565 01:13:43,980 --> 01:13:45,870 Who has access to the ledger? 1566 01:13:45,870 --> 01:13:50,060 Who has access, if I might, to the data? 1567 01:13:50,060 --> 01:13:53,180 Three different types of approaches we talked about-- 1568 01:13:53,180 --> 01:13:56,420 open permissionless, multiple permissioned, 1569 01:13:56,420 --> 01:13:57,410 and client server. 1570 01:13:57,410 --> 01:13:59,090 I'm using the word client server just 1571 01:13:59,090 --> 01:14:01,190 to say traditional database. 1572 01:14:01,190 --> 01:14:03,200 You can use another words. 1573 01:14:03,200 --> 01:14:06,530 But I'm just trying to say three different types of data 1574 01:14:06,530 --> 01:14:07,340 structures. 1575 01:14:07,340 --> 01:14:09,410 And to Zan's question, why would you 1576 01:14:09,410 --> 01:14:11,420 be in one bucket versus another? 1577 01:14:11,420 --> 01:14:12,710 And I'm saying this. 1578 01:14:12,710 --> 01:14:15,170 If you're a venture capitalist one day, 1579 01:14:15,170 --> 01:14:18,890 and somebody comes in to you with a blockchain solution, 1580 01:14:18,890 --> 01:14:20,600 this is the same thing. 1581 01:14:20,600 --> 01:14:22,400 Hopefully, you'll make it better. 1582 01:14:22,400 --> 01:14:24,890 This is just my approach to it. 1583 01:14:24,890 --> 01:14:29,360 Public blockchain-- do you need, basically, 1584 01:14:29,360 --> 01:14:32,090 public write capability, that lots of people 1585 01:14:32,090 --> 01:14:34,550 can write to these ledgers? 1586 01:14:34,550 --> 01:14:36,230 Do you need that somewhere? 1587 01:14:36,230 --> 01:14:39,830 Secondly, do you want some peer to peer transaction capability 1588 01:14:39,830 --> 01:14:44,040 across the distributed ledger? 1589 01:14:44,040 --> 01:14:46,155 Do you kind of not want any central intermediary? 1590 01:14:46,155 --> 01:14:48,030 Maybe you don't want the central intermediary 1591 01:14:48,030 --> 01:14:49,650 because they're slow and sluggish. 1592 01:14:49,650 --> 01:14:51,525 Maybe you don't want the central intermediary 1593 01:14:51,525 --> 01:14:53,520 because they have a lot of economic rents. 1594 01:14:53,520 --> 01:14:55,590 It doesn't matter to me why you want to get rid 1595 01:14:55,590 --> 01:14:57,072 of the central intermediary. 1596 01:14:57,072 --> 01:14:59,280 But if you want to get rid of a central intermediary, 1597 01:14:59,280 --> 01:15:02,630 or lessen their control, or lower their control, 1598 01:15:02,630 --> 01:15:06,690 raise peer to peer transactions, have a lot of public writing-- 1599 01:15:06,690 --> 01:15:09,810 oh, and by the way, if you want some token economics, 1600 01:15:09,810 --> 01:15:11,700 you're probably somewhere over here. 1601 01:15:11,700 --> 01:15:13,620 I'm not saying you have to have all four. 1602 01:15:13,620 --> 01:15:16,230 But to me, those are the kind of three 1603 01:15:16,230 --> 01:15:18,360 or four things that are floating around 1604 01:15:18,360 --> 01:15:21,520 why you might be over there. 1605 01:15:21,520 --> 01:15:23,860 Private blockchains-- well, maybe you actually 1606 01:15:23,860 --> 01:15:25,960 don't want public write capability. 1607 01:15:25,960 --> 01:15:27,050 You need private. 1608 01:15:27,050 --> 01:15:32,110 Whether it's just the Australian Stock Exchange or 15 to 20 club 1609 01:15:32,110 --> 01:15:36,670 deal, you still want kind of multiple people to write, 1610 01:15:36,670 --> 01:15:39,550 but you want it to be private. 1611 01:15:39,550 --> 01:15:46,670 But if you still need finality of data and an append-only log, 1612 01:15:46,670 --> 01:15:48,990 that you need this appending-- 1613 01:15:48,990 --> 01:15:53,090 the log, and you need some public verifiability, 1614 01:15:53,090 --> 01:15:55,880 that you still need that data to be verified amongst-- 1615 01:15:55,880 --> 01:15:59,360 and this public means 15 or 20 players. 1616 01:15:59,360 --> 01:16:02,280 But you need it to be able to be verified-- 1617 01:16:02,280 --> 01:16:05,210 so the trust mechanism is not thousands and open. 1618 01:16:05,210 --> 01:16:08,330 The trust mechanism is 15 or 20 or 5 parties. 1619 01:16:08,330 --> 01:16:09,530 But you still need it-- 1620 01:16:09,530 --> 01:16:11,600 you might be there. 1621 01:16:11,600 --> 01:16:14,630 Where I come out is there's some cases that you just 1622 01:16:14,630 --> 01:16:17,470 don't need any of this. 1623 01:16:17,470 --> 01:16:21,190 There's a trusted party that hosts the data. 1624 01:16:21,190 --> 01:16:23,920 Your trust mechanism is one party. 1625 01:16:23,920 --> 01:16:25,570 The trusted party can do the-- 1626 01:16:25,570 --> 01:16:27,460 I use the letters CRUD. 1627 01:16:27,460 --> 01:16:36,390 That's the create, and update, and delete, and so forth. 1628 01:16:36,390 --> 01:16:40,400 And that's based on client service architecture, 1629 01:16:40,400 --> 01:16:42,170 this client server architecture. 1630 01:16:42,170 --> 01:16:45,260 There's always, in that circumstance, somebody right 1631 01:16:45,260 --> 01:16:49,250 in the middle that basically owns, 1632 01:16:49,250 --> 01:16:54,130 updates, governs that whole architecture. 1633 01:16:54,130 --> 01:16:59,050 My business judgment on this is if you're 1634 01:16:59,050 --> 01:17:03,620 moving things of value, you want those things to value 1635 01:17:03,620 --> 01:17:08,340 to have some final settlement and immutability. 1636 01:17:08,340 --> 01:17:10,710 And immutability append-only logs give you 1637 01:17:10,710 --> 01:17:13,550 some of that final settlement. 1638 01:17:13,550 --> 01:17:15,800 We talked very briefly about a lawsuit 1639 01:17:15,800 --> 01:17:19,690 that was 300 plus years ago-- the Crawford case in Scotland 1640 01:17:19,690 --> 01:17:22,340 that you won't have to remember. 1641 01:17:22,340 --> 01:17:25,820 But this-- watch this. 1642 01:17:25,820 --> 01:17:27,380 So steal this from me. 1643 01:17:27,380 --> 01:17:28,230 Just do me a favor. 1644 01:17:28,230 --> 01:17:29,070 Just steal it. 1645 01:17:29,070 --> 01:17:29,570 OK. 1646 01:17:29,570 --> 01:17:31,610 Right now, I have a right of action. 1647 01:17:31,610 --> 01:17:34,310 But if you give it to-- who are you going to give it to? 1648 01:17:34,310 --> 01:17:37,780 I don't have any rights against you. 1649 01:17:37,780 --> 01:17:38,870 No. 1650 01:17:38,870 --> 01:17:39,380 Well, wait. 1651 01:17:39,380 --> 01:17:42,020 It depends whether he knew you stole it from me. 1652 01:17:42,020 --> 01:17:45,270 But it's gone. 1653 01:17:45,270 --> 01:17:47,670 That's final settlement. 1654 01:17:47,670 --> 01:17:50,250 Those $2 are gone. 1655 01:17:50,250 --> 01:17:51,930 That's final settlement. 1656 01:17:51,930 --> 01:17:52,770 It's immutable. 1657 01:17:52,770 --> 01:17:54,270 I can't get them back. 1658 01:17:54,270 --> 01:17:55,050 What, Christopher? 1659 01:17:55,050 --> 01:17:55,820 You want them? 1660 01:17:55,820 --> 01:17:57,570 AUDIENCE: I just was asking for the money. 1661 01:17:57,570 --> 01:17:59,550 PROFESSOR: You were asking for the money. 1662 01:17:59,550 --> 01:18:02,580 I just use that as a visualization. 1663 01:18:02,580 --> 01:18:06,780 Money is a social consensus and a social construct. 1664 01:18:06,780 --> 01:18:08,910 But if you think of it in terms of ledgers-- 1665 01:18:08,910 --> 01:18:09,900 look, the money's gone. 1666 01:18:09,900 --> 01:18:10,890 I can't get it. 1667 01:18:10,890 --> 01:18:12,570 That's it, you know? 1668 01:18:12,570 --> 01:18:13,470 That's good. 1669 01:18:13,470 --> 01:18:16,340 You all go to Sloan, I can tell. 1670 01:18:16,340 --> 01:18:18,600 Yeah, yeah. 1671 01:18:18,600 --> 01:18:20,490 But it's an important concept. 1672 01:18:20,490 --> 01:18:22,750 I think of this whole that way. 1673 01:18:22,750 --> 01:18:24,870 So there were two questions, and then we'll-- 1674 01:18:24,870 --> 01:18:27,150 and then this is the decentralized end. 1675 01:18:27,150 --> 01:18:30,330 Bitcoin and Ethereum are kind of at the decentralized end. 1676 01:18:30,330 --> 01:18:34,160 I would contend Bitcoin's more decentralized than Ethereum. 1677 01:18:34,160 --> 01:18:37,610 And centralized databases-- initial coin offerings 1678 01:18:37,610 --> 01:18:39,170 that we'll talk about later, I think, 1679 01:18:39,170 --> 01:18:40,670 are actually maybe a little bit more 1680 01:18:40,670 --> 01:18:42,920 centralized, even, than permissioned blockchains. 1681 01:18:42,920 --> 01:18:45,140 But we'll get to that. 1682 01:18:45,140 --> 01:18:46,970 AUDIENCE: No, mine was already answered. 1683 01:18:46,970 --> 01:18:48,428 And then the other comment was more 1684 01:18:48,428 --> 01:18:49,640 about the supply chain link. 1685 01:18:49,640 --> 01:18:51,557 When you think about the millions and millions 1686 01:18:51,557 --> 01:18:54,160 of dollars, the element of transparency 1687 01:18:54,160 --> 01:18:56,930 that blockchain kind of brings in to all the parties 1688 01:18:56,930 --> 01:18:59,570 probably saves an insane amount of money 1689 01:18:59,570 --> 01:19:01,950 to what you're talking about kind of reconciling 1690 01:19:01,950 --> 01:19:04,470 all the databases. 1691 01:19:04,470 --> 01:19:08,650 PROFESSOR: So we're going to talk 1692 01:19:08,650 --> 01:19:12,400 about finance next Thursday and some of the strengths 1693 01:19:12,400 --> 01:19:14,902 and weaknesses in finance, and some of the attributes, 1694 01:19:14,902 --> 01:19:16,360 and you know, one of the readings-- 1695 01:19:16,360 --> 01:19:17,950 Sheila Bair-- and there's a bunch 1696 01:19:17,950 --> 01:19:20,050 of optional readings about what's 1697 01:19:20,050 --> 01:19:23,108 happened after the financial crisis and so forth. 1698 01:19:23,108 --> 01:19:25,150 And then we're going to talk about the economics. 1699 01:19:25,150 --> 01:19:28,840 But I've had three or four groups come in already talking 1700 01:19:28,840 --> 01:19:30,030 about the final projects. 1701 01:19:30,030 --> 01:19:32,680 So I'll just say, think about whatever you're working 1702 01:19:32,680 --> 01:19:35,950 on as this new technology-- blockchain-- 1703 01:19:35,950 --> 01:19:38,590 whether it's permissioned or permissionless-- 1704 01:19:38,590 --> 01:19:42,820 either one-- does this new technology really address some 1705 01:19:42,820 --> 01:19:46,450 pain point that you're trying to solve, 1706 01:19:46,450 --> 01:19:50,220 whether it's about payment systems or-- 1707 01:19:50,220 --> 01:19:52,500 actually, there's a group that's talking about doing 1708 01:19:52,500 --> 01:19:55,730 supply chain on wine. 1709 01:19:55,730 --> 01:19:56,930 And I figured, why not? 1710 01:19:56,930 --> 01:19:58,070 I'll probably approve it. 1711 01:20:01,310 --> 01:20:04,940 But it's, what is the pain point that you're actually 1712 01:20:04,940 --> 01:20:07,320 trying to solve? 1713 01:20:07,320 --> 01:20:11,070 Or will decentralized peer to peer networking 1714 01:20:11,070 --> 01:20:15,000 somehow create a business opportunity in a new model? 1715 01:20:15,000 --> 01:20:16,390 It might not be a pain point. 1716 01:20:16,390 --> 01:20:18,330 It might be a business opportunity 1717 01:20:18,330 --> 01:20:22,620 that decentralized systems solve. 1718 01:20:22,620 --> 01:20:25,380 Or it might be an opportunity that token economics 1719 01:20:25,380 --> 01:20:27,430 can help solve. 1720 01:20:27,430 --> 01:20:30,400 But if there's no pain point you're solving, 1721 01:20:30,400 --> 01:20:36,390 no token economics, no decentralization peer to peer, 1722 01:20:36,390 --> 01:20:38,280 I ask you to use your critical reasoning 1723 01:20:38,280 --> 01:20:41,940 and move to the next use case, because this is really 1724 01:20:41,940 --> 01:20:44,460 about-- this is a business class. 1725 01:20:44,460 --> 01:20:47,640 This is about finding places for this incredible set 1726 01:20:47,640 --> 01:20:52,745 of new technologies in the context of markets 1727 01:20:52,745 --> 01:20:57,720 and to have a really healthy sense of ground 1728 01:20:57,720 --> 01:21:00,600 truth about what's possible. 1729 01:21:00,600 --> 01:21:02,670 If not, traditional databases. 1730 01:21:02,670 --> 01:21:03,810 Move on. 1731 01:21:03,810 --> 01:21:05,880 I'm not looking for projects at the end 1732 01:21:05,880 --> 01:21:09,670 of the semester that are using traditional databases. 1733 01:21:09,670 --> 01:21:12,000 It's got to really use blockchain. 1734 01:21:12,000 --> 01:21:13,330 So I thank you. 1735 01:21:13,330 --> 01:21:15,380 I'll see you next Thursday.