1 00:00:01,235 --> 00:00:03,560 The following content is provided under a Creative 2 00:00:03,560 --> 00:00:04,950 Commons license. 3 00:00:04,950 --> 00:00:07,160 Your support will help MIT OpenCourseWare 4 00:00:07,160 --> 00:00:11,250 continue to offer high quality educational resources for free. 5 00:00:11,250 --> 00:00:13,790 To make a donation or to view additional materials 6 00:00:13,790 --> 00:00:17,750 from hundreds of MIT courses, visit MIT OpenCourseWare 7 00:00:17,750 --> 00:00:18,770 at ocw.mit.edu. 8 00:00:22,640 --> 00:00:26,570 GARY GENSLER: Thank you for everybody coming back. 9 00:00:26,570 --> 00:00:32,150 And I should tell you, when I start 10 00:00:32,150 --> 00:00:35,030 the session with that little shh, 11 00:00:35,030 --> 00:00:40,420 I learned from a congressman in Baltimore. 12 00:00:40,420 --> 00:00:44,270 As some of you know, I've spent a lot of time around politics. 13 00:00:44,270 --> 00:00:46,610 And one of my roles in politics was 14 00:00:46,610 --> 00:00:50,060 that I was the treasurer of the Maryland Democratic Party. 15 00:00:50,060 --> 00:00:51,800 Which, if you anybody ever asks you 16 00:00:51,800 --> 00:00:53,810 to be the treasurer of a state party, come. 17 00:00:53,810 --> 00:00:57,080 I can give you some advice about what 18 00:00:57,080 --> 00:00:58,910 that's all about when your home state 19 00:00:58,910 --> 00:01:00,510 senator asks you to do it. 20 00:01:00,510 --> 00:01:05,420 So I had to quiet down the annual Jefferson Jackson Day 21 00:01:05,420 --> 00:01:06,890 dinners. 22 00:01:06,890 --> 00:01:10,070 I'd organize these big dinners, and there's 400, 500, 600 23 00:01:10,070 --> 00:01:11,630 people at these dinners. 24 00:01:11,630 --> 00:01:12,450 I couldn't do it. 25 00:01:12,450 --> 00:01:13,955 I couldn't get their attention. 26 00:01:13,955 --> 00:01:16,338 And Congressman Elijah Cummings comes up, 27 00:01:16,338 --> 00:01:17,630 and he just leans into the mic. 28 00:01:17,630 --> 00:01:19,290 Shh. 29 00:01:19,290 --> 00:01:21,170 And it quieted the whole place down. 30 00:01:21,170 --> 00:01:23,530 I said, Congressman, what is this? 31 00:01:23,530 --> 00:01:25,580 Is this something you've learned in politics, 32 00:01:25,580 --> 00:01:32,060 something you learned from your minister, your priest? 33 00:01:32,060 --> 00:01:35,730 He said, it works every time, works every time. 34 00:01:35,730 --> 00:01:39,110 So Elijah Cummings gave me that little duty. 35 00:01:39,110 --> 00:01:41,510 Blockchain and money, we're here. 36 00:01:41,510 --> 00:01:43,490 I know it's a little bit like eating broccoli 37 00:01:43,490 --> 00:01:46,040 these last couple of classes, because we went 38 00:01:46,040 --> 00:01:49,610 through cryptography, and then we moved a little bit 39 00:01:49,610 --> 00:01:54,320 into consensus protocols. 40 00:01:54,320 --> 00:01:58,460 And today, we're going to pick that up and try to finish up 41 00:01:58,460 --> 00:02:02,150 the design bits of Bitcoin. 42 00:02:02,150 --> 00:02:06,263 I wanted to introduce, though, one walk-in. 43 00:02:06,263 --> 00:02:08,180 We're going to have walk-ins from time to time 44 00:02:08,180 --> 00:02:08,763 in this class. 45 00:02:08,763 --> 00:02:11,690 But Patrick Murck, he's hiding over here 46 00:02:11,690 --> 00:02:13,550 with the cap in the hoodie. 47 00:02:13,550 --> 00:02:16,190 He's got the look for a lawyer who's 48 00:02:16,190 --> 00:02:20,270 been spending his time around Bitcoin for seven years. 49 00:02:20,270 --> 00:02:23,732 Now, he's currently affiliated with Berkman Klein at Harvard. 50 00:02:23,732 --> 00:02:25,690 And I didn't know Patrick was going to be here, 51 00:02:25,690 --> 00:02:26,648 so I'm calling him out. 52 00:02:26,648 --> 00:02:28,910 He's also special counsel of the Cooley law firm. 53 00:02:28,910 --> 00:02:31,880 And he has a bunch of clients who 54 00:02:31,880 --> 00:02:33,620 try to do the right thing by the law. 55 00:02:33,620 --> 00:02:36,535 But sometimes they find themselves 56 00:02:36,535 --> 00:02:38,660 dealing with the Securities and Exchange Commission 57 00:02:38,660 --> 00:02:41,580 or other fine government institutions. 58 00:02:41,580 --> 00:02:43,890 But Patrick was also general counsel 59 00:02:43,890 --> 00:02:45,323 to the Bitcoin Foundation. 60 00:02:45,323 --> 00:02:47,240 And for a short while, you ran it, didn't you, 61 00:02:47,240 --> 00:02:50,358 before it kind of went puff? 62 00:02:50,358 --> 00:02:51,150 PATRICK MURCK: Yes. 63 00:02:51,150 --> 00:02:52,718 [INAUDIBLE] 64 00:02:52,718 --> 00:02:53,510 GARY GENSLER: Yeah. 65 00:02:53,510 --> 00:02:55,700 But Patrick's the type of lawyer that wears 66 00:02:55,700 --> 00:02:58,500 a hoodie and a baseball cap. 67 00:02:58,500 --> 00:03:01,440 He's a Bitcoin lawyer. 68 00:03:01,440 --> 00:03:04,860 If you ever need Patrick for one of your entrepreneurial 69 00:03:04,860 --> 00:03:07,970 efforts, I'm sure Cooley law firm will like advising you, 70 00:03:07,970 --> 00:03:09,626 too. 71 00:03:09,626 --> 00:03:11,087 AUDIENCE: Do you take bitcoin? 72 00:03:11,087 --> 00:03:12,170 GARY GENSLER: What's that? 73 00:03:12,170 --> 00:03:13,890 AUDIENCE: Does he take bitcoin? 74 00:03:16,990 --> 00:03:18,350 PATRICK MURCK: I have. 75 00:03:18,350 --> 00:03:22,000 [INAUDIBLE] In fact, I've never bought a bitcoin, ever. 76 00:03:22,000 --> 00:03:25,030 [INAUDIBLE] paid 100% in bitcoin, so it's all good. 77 00:03:25,030 --> 00:03:26,593 [LAUGHTER] 78 00:03:26,593 --> 00:03:28,010 GARY GENSLER: So Patrick, when you 79 00:03:28,010 --> 00:03:32,660 get paid 100% in bitcoin from a client, 80 00:03:32,660 --> 00:03:35,630 how long do you hold it? 81 00:03:35,630 --> 00:03:38,938 PATRICK MURCK: Well, it depends, right? 82 00:03:38,938 --> 00:03:40,922 AUDIENCE: [INAUDIBLE] 83 00:03:40,922 --> 00:03:42,680 PATRICK MURCK: Often, in the early days-- 84 00:03:42,680 --> 00:03:44,960 not very long, because I had to pay 85 00:03:44,960 --> 00:03:48,183 for my mortgage and my kid's daycare and things like that, 86 00:03:48,183 --> 00:03:49,350 and they don't take bitcoin. 87 00:03:49,350 --> 00:03:53,180 So I've sold a lot of bitcoin over the years. 88 00:03:53,180 --> 00:03:56,385 Something sadly, but I never bought them. 89 00:03:56,385 --> 00:03:57,265 I just earned them. 90 00:03:57,265 --> 00:03:59,390 GARY GENSLER: Earned them-- so you're like a miner, 91 00:03:59,390 --> 00:04:01,370 except for you're a lawyer who gets bitcoin. 92 00:04:01,370 --> 00:04:01,870 [LAUGHTER] 93 00:04:01,870 --> 00:04:03,328 PATRICK MURCK: I mine with my mind. 94 00:04:03,328 --> 00:04:05,490 [LAUGHTER] 95 00:04:07,940 --> 00:04:10,580 GARY GENSLER: The study questions for today 96 00:04:10,580 --> 00:04:12,740 was, we were going to turn to that last piece 97 00:04:12,740 --> 00:04:17,779 about transactions and something called the unspent transaction 98 00:04:17,779 --> 00:04:23,540 output, and script code that is a computer code that's 99 00:04:23,540 --> 00:04:26,968 used inside of Bitcoin. 100 00:04:26,968 --> 00:04:29,900 We're going to talk a little bit about the design 101 00:04:29,900 --> 00:04:32,360 features bringing it all together, and particularly 102 00:04:32,360 --> 00:04:36,230 around the reading, which was the academic pedigree 103 00:04:36,230 --> 00:04:38,240 of Bitcoin. 104 00:04:38,240 --> 00:04:40,130 Which is also a reading that Patrick 105 00:04:40,130 --> 00:04:46,310 assigns when he does his study group at Harvard Law, as well. 106 00:04:46,310 --> 00:04:48,690 And then, yes, all of you are going 107 00:04:48,690 --> 00:04:49,990 to be able to participate. 108 00:04:49,990 --> 00:04:53,540 And we're going to take a survey amongst all of you 109 00:04:53,540 --> 00:04:56,210 as to who's Satoshi Nakamoto. 110 00:04:56,210 --> 00:04:58,540 Ah, yeah, yeah, yeah. 111 00:04:58,540 --> 00:05:01,220 Or would you rather do more on transactions script? 112 00:05:01,220 --> 00:05:02,930 [LAUGHTER] 113 00:05:02,930 --> 00:05:06,240 All right, we'll do a little bit of both. 114 00:05:06,240 --> 00:05:10,015 The readings, of course, the Bitcoin Academic Pedigree, 115 00:05:10,015 --> 00:05:12,140 which we're going to talk about in the latter part. 116 00:05:12,140 --> 00:05:14,120 And I'll do a little cold-calling 117 00:05:14,120 --> 00:05:17,870 and get some feedback as to what you thought from the reading 118 00:05:17,870 --> 00:05:20,930 or if you're still skimming it. 119 00:05:20,930 --> 00:05:23,750 And then there was a CoinDesk article, just 120 00:05:23,750 --> 00:05:26,870 making sense of it all. 121 00:05:26,870 --> 00:05:31,840 And so as I said, we'll go through the transactions. 122 00:05:31,840 --> 00:05:34,350 We'll do a little bit of putting it all together 123 00:05:34,350 --> 00:05:36,160 and the Academic Pedigree. 124 00:05:36,160 --> 00:05:37,710 We'll have a little fun. 125 00:05:37,710 --> 00:05:40,800 You all are going to have a chance 126 00:05:40,800 --> 00:05:47,610 to tell us who Satoshi Nakamoto is, or was, or the committee. 127 00:05:47,610 --> 00:05:50,760 So transactions-- you've seen this graph before. 128 00:05:50,760 --> 00:05:54,750 But transactions, the format in a transaction 129 00:05:54,750 --> 00:05:57,780 ledger-- not only in Bitcoin, but really everything-- 130 00:05:57,780 --> 00:06:00,420 is somebody on one side of a transaction 131 00:06:00,420 --> 00:06:03,910 and somebody on another side of a transaction. 132 00:06:03,910 --> 00:06:07,370 So in Bitcoin, there's an input. 133 00:06:07,370 --> 00:06:15,690 And the input inside of Bitcoin is an ID of a previous output. 134 00:06:15,690 --> 00:06:21,450 So not only is Bitcoin a series of blocks of information-- 135 00:06:21,450 --> 00:06:27,860 each block that has 1,000 to 2,000 individual transactions-- 136 00:06:27,860 --> 00:06:31,760 but in a sense, there's a separate chain that's going on. 137 00:06:31,760 --> 00:06:33,500 I sometimes think of a blockchain 138 00:06:33,500 --> 00:06:38,700 as, yes, there's the chain of big chunks of data, 139 00:06:38,700 --> 00:06:41,660 but there's also a chain of individual transactions. 140 00:06:41,660 --> 00:06:44,852 Shimon, is that a hand raised or just scratching your head? 141 00:06:44,852 --> 00:06:46,148 AUDIENCE: I'll be more careful. 142 00:06:46,148 --> 00:06:47,606 GARY GENSLER: No, that's all right. 143 00:06:47,606 --> 00:06:50,257 Shimon is a faculty member of the Finance department 144 00:06:50,257 --> 00:06:50,840 here at Sloan. 145 00:06:50,840 --> 00:06:55,545 And some of you might take his Finance 1 course, I assume. 146 00:06:55,545 --> 00:06:56,795 Somebody in here is, probably. 147 00:06:59,990 --> 00:07:02,570 So the input is just really the idea 148 00:07:02,570 --> 00:07:04,780 of wherever the output was. 149 00:07:04,780 --> 00:07:09,590 You can think of a chain of a transaction. 150 00:07:09,590 --> 00:07:13,080 Can anybody tell me where all transactions come from? 151 00:07:13,080 --> 00:07:17,420 Where's sort of the genesis of the value, if you 152 00:07:17,420 --> 00:07:22,450 follow a transaction chain back to its origin? 153 00:07:22,450 --> 00:07:23,498 Hugo. 154 00:07:23,498 --> 00:07:24,290 AUDIENCE: Coinbase. 155 00:07:24,290 --> 00:07:26,510 GARY GENSLER: Coinbase-- so anybody 156 00:07:26,510 --> 00:07:30,305 want to say what a Coinbase is? 157 00:07:30,305 --> 00:07:36,140 AUDIENCE: It's a generation of a freshly minted bitcoin. 158 00:07:36,140 --> 00:07:37,730 GARY GENSLER: Right, so the generation 159 00:07:37,730 --> 00:07:40,580 of a freshly minted bitcoin-- not the way that Patrick Murck 160 00:07:40,580 --> 00:07:44,360 got it when he's selling his services for the law, 161 00:07:44,360 --> 00:07:49,010 but how a miner gets it, they get their transactions. 162 00:07:49,010 --> 00:07:52,850 Remember initially, 50 coins was what happened initially back 163 00:07:52,850 --> 00:07:54,350 in 2009. 164 00:07:54,350 --> 00:07:56,670 Now it's 12 and 1/2 coins. 165 00:07:56,670 --> 00:07:58,760 And in a short couple-- 166 00:07:58,760 --> 00:08:00,470 maybe it's about 18 or 20 months, 167 00:08:00,470 --> 00:08:02,450 it goes to 6 and 1/4 coins. 168 00:08:02,450 --> 00:08:07,370 And it keeps splitting half and half and half, eventually 169 00:08:07,370 --> 00:08:10,010 to no Coinbase. 170 00:08:10,010 --> 00:08:12,410 But every transaction, in essence, 171 00:08:12,410 --> 00:08:14,420 had to go back all the way to Coinbase-- 172 00:08:14,420 --> 00:08:20,270 some Coinbase, some 50 or 25 or 12 and 1/2 coins being issued. 173 00:08:20,270 --> 00:08:21,290 And then the output-- 174 00:08:21,290 --> 00:08:24,560 so a transaction format is pretty straightforward, 175 00:08:24,560 --> 00:08:25,190 in a sense. 176 00:08:25,190 --> 00:08:30,590 The input is a previous output and a digital signature, 177 00:08:30,590 --> 00:08:33,200 and then, to whom do you want to send it. 178 00:08:33,200 --> 00:08:35,510 And it's sent to Bitcoin addresses. 179 00:08:35,510 --> 00:08:37,309 That's why we spent just a moment-- 180 00:08:37,309 --> 00:08:40,429 we just glossed over what a Bitcoin address is. 181 00:08:40,429 --> 00:08:43,880 It's sort of a translation to a public key, 182 00:08:43,880 --> 00:08:46,910 but it's not identical to a public key. 183 00:08:46,910 --> 00:08:48,950 And of course, you need value. 184 00:08:48,950 --> 00:08:51,470 Value measured in bitcoins or satoshis, 185 00:08:51,470 --> 00:08:53,270 or if we're on the Ethereum network, 186 00:08:53,270 --> 00:08:56,750 it would be an ether and gas, et cetera, et cetera. 187 00:08:56,750 --> 00:08:58,760 On 1,600 different platforms, it could 188 00:08:58,760 --> 00:09:01,005 be a different native currency. 189 00:09:03,900 --> 00:09:06,085 Lock time, I don't think lock time 190 00:09:06,085 --> 00:09:07,210 was in any of the readings. 191 00:09:07,210 --> 00:09:10,910 But anybody want to tell me what a lock time is? 192 00:09:10,910 --> 00:09:14,600 It's relevant to some of how the technology goes. 193 00:09:14,600 --> 00:09:17,090 Anybody want to take a guess? 194 00:09:17,090 --> 00:09:18,020 Where's Alin? 195 00:09:18,020 --> 00:09:21,170 Not Alin the PhD, but Alin from the digital currency 196 00:09:21,170 --> 00:09:21,680 initiative. 197 00:09:21,680 --> 00:09:22,130 No? 198 00:09:22,130 --> 00:09:22,713 You're hiding. 199 00:09:22,713 --> 00:09:24,780 You don't want to say what a lock time is? 200 00:09:24,780 --> 00:09:29,640 AUDIENCE: I thought it's some sort of protection mechanism 201 00:09:29,640 --> 00:09:34,212 for, I guess, double spending. 202 00:09:34,212 --> 00:09:36,710 GARY GENSLER: It is a protection mechanism, but not so much 203 00:09:36,710 --> 00:09:38,260 about double spending. 204 00:09:38,260 --> 00:09:39,647 Tom, did I see your hand up? 205 00:09:39,647 --> 00:09:40,730 AUDIENCE: This is a guess. 206 00:09:40,730 --> 00:09:44,080 Is it when the transaction is hashed? 207 00:09:44,080 --> 00:09:45,080 GARY GENSLER: It's when. 208 00:09:45,080 --> 00:09:46,310 It's about time. 209 00:09:46,310 --> 00:09:48,570 It's when the transaction can happen. 210 00:09:48,570 --> 00:09:51,620 So right now, it's 2:45 on September 20. 211 00:09:51,620 --> 00:09:56,510 If you put a lock time in at 2:50 or 2:55, 212 00:09:56,510 --> 00:09:59,210 it couldn't happen until 2:55. 213 00:09:59,210 --> 00:10:01,790 So you can actually conditional the transaction. 214 00:10:01,790 --> 00:10:03,470 That's all lock time is. 215 00:10:03,470 --> 00:10:06,790 But you can say, it can't happen until. 216 00:10:06,790 --> 00:10:08,528 You could put tomorrow's date. 217 00:10:08,528 --> 00:10:10,070 AUDIENCE: The counterparty sets that, 218 00:10:10,070 --> 00:10:11,468 so it's like the date on a check? 219 00:10:11,468 --> 00:10:13,510 GARY GENSLER: It's a little different than a date 220 00:10:13,510 --> 00:10:14,980 on a check. 221 00:10:14,980 --> 00:10:18,940 Because one of the things that gets validated in Bitcoin 222 00:10:18,940 --> 00:10:23,310 is, is that it will not validate a transaction early. 223 00:10:23,310 --> 00:10:28,330 And if you put October 6 on a handwritten check, 224 00:10:28,330 --> 00:10:30,760 the bank might still take it, even though they probably 225 00:10:30,760 --> 00:10:32,530 shouldn't. 226 00:10:32,530 --> 00:10:35,880 So it is like the date on a check, 227 00:10:35,880 --> 00:10:42,020 except it's verified and validated and so forth. 228 00:10:42,020 --> 00:10:45,110 I'm going to take Emily and then Shimon. 229 00:10:45,110 --> 00:10:47,610 AUDIENCE: So if you were to set the lock time in the future, 230 00:10:47,610 --> 00:10:53,100 could that mess up the blocks in the chain? 231 00:10:53,100 --> 00:10:56,850 What is kind of the right chain of events 232 00:10:56,850 --> 00:10:59,430 that you're recognizing if you're choosing to set 233 00:10:59,430 --> 00:11:00,953 a lock time in the future? 234 00:11:00,953 --> 00:11:03,120 I guess the broader question would be, why would you 235 00:11:03,120 --> 00:11:05,643 set a lock time in the future? 236 00:11:05,643 --> 00:11:07,560 GARY GENSLER: Shimon, were you answering that? 237 00:11:07,560 --> 00:11:08,990 AUDIENCE: No, no, no. 238 00:11:08,990 --> 00:11:11,282 I want to ask you, in the [INAUDIBLE] question section, 239 00:11:11,282 --> 00:11:15,030 is what purpose does it serve, right? 240 00:11:15,030 --> 00:11:17,310 AUDIENCE: It's not a really Turing complete language, 241 00:11:17,310 --> 00:11:21,390 so you're not really trying to create conditions here. 242 00:11:21,390 --> 00:11:22,943 But this is sort of a condition. 243 00:11:22,943 --> 00:11:24,360 GARY GENSLER: So it's a condition. 244 00:11:24,360 --> 00:11:26,010 Because at any point in time, you 245 00:11:26,010 --> 00:11:27,600 might want a condition of payment. 246 00:11:27,600 --> 00:11:30,780 You might want a condition of payment on time. 247 00:11:30,780 --> 00:11:33,540 And we're, in a few slides, going 248 00:11:33,540 --> 00:11:36,360 to talk about the scripting language, the computer 249 00:11:36,360 --> 00:11:39,520 language that allows transactions to happen. 250 00:11:39,520 --> 00:11:43,270 Shimon said, was not Turing complete. 251 00:11:43,270 --> 00:11:47,980 From the readings, anybody want to tell us 252 00:11:47,980 --> 00:11:52,720 what Turing complete is in computer science language? 253 00:11:55,990 --> 00:11:57,190 No? 254 00:11:57,190 --> 00:11:58,660 Anybody know who Turing is? 255 00:11:58,660 --> 00:12:00,787 Anybody see Imitation Games. 256 00:12:03,470 --> 00:12:05,855 Anyone know what the Turing Award is? 257 00:12:05,855 --> 00:12:09,130 It's sort of the Nobel Laureate for computer scientists. 258 00:12:09,130 --> 00:12:11,780 AUDIENCE: Isn't that the award when the machine can 259 00:12:11,780 --> 00:12:14,720 actually can pass as a human being? 260 00:12:14,720 --> 00:12:17,220 GARY GENSLER: Well, that's one thing associated with Turing. 261 00:12:17,220 --> 00:12:20,270 But the Turing Award is an annual award, 262 00:12:20,270 --> 00:12:23,540 sort of like a Nobel Laureate, but for computer scientists. 263 00:12:23,540 --> 00:12:27,470 Turing complete allows you to do loops inside 264 00:12:27,470 --> 00:12:30,120 of computer programs. 265 00:12:30,120 --> 00:12:33,540 And the script language does not allow that to happen. 266 00:12:33,540 --> 00:12:37,800 Every function needs to sort of have some language. 267 00:12:37,800 --> 00:12:40,410 I know Alin looking at me to see I don't dive. 268 00:12:43,700 --> 00:12:47,340 But to answer Emily's question, it's just, 269 00:12:47,340 --> 00:12:50,760 there's so many different ways to condition a transaction. 270 00:12:50,760 --> 00:12:52,410 And Satoshi Nakamoto thought, well, 271 00:12:52,410 --> 00:12:56,490 let's put it in here, right in the transaction format, 272 00:12:56,490 --> 00:13:00,100 that you can condition on time. 273 00:13:00,100 --> 00:13:02,470 And then two parties could do that. 274 00:13:02,470 --> 00:13:05,560 To your question about whether it could inadvertently 275 00:13:05,560 --> 00:13:07,840 lead to double spending, it's a very good question. 276 00:13:07,840 --> 00:13:10,960 Can I hold it until we just do validation for a second? 277 00:13:13,700 --> 00:13:17,650 So this is a unique identifier for the input. 278 00:13:17,650 --> 00:13:23,140 But it's really uniquely identifying from a past output. 279 00:13:23,140 --> 00:13:25,030 What's the block number? 280 00:13:25,030 --> 00:13:30,130 Is it from the 250,000th block or the 300,000th block? 281 00:13:30,130 --> 00:13:33,250 So it takes literally a block ID. 282 00:13:33,250 --> 00:13:36,610 And then within that block, which one of its 1,500 283 00:13:36,610 --> 00:13:39,100 transactions might it be? 284 00:13:39,100 --> 00:13:43,060 So you can find any transaction on an entire blockchain 285 00:13:43,060 --> 00:13:46,330 by knowing the block, and then within the block, which 286 00:13:46,330 --> 00:13:47,620 transaction. 287 00:13:47,620 --> 00:13:52,210 It's just a data mechanism on how to store data. 288 00:13:52,210 --> 00:13:53,710 And through this mechanism, there's 289 00:13:53,710 --> 00:13:58,230 a chain of transactions, as well as a chain of blocks. 290 00:13:58,230 --> 00:14:01,478 And then the value we talked about that-- 291 00:14:01,478 --> 00:14:03,270 I think there's, if I did my numbers right, 292 00:14:03,270 --> 00:14:07,213 10 to the 8th satoshis in every bitcoin. 293 00:14:07,213 --> 00:14:08,630 It's a little hard because I think 294 00:14:08,630 --> 00:14:11,020 there's 10 to the ninth gas in every ether. 295 00:14:13,875 --> 00:14:15,340 And that's a coin. 296 00:14:15,340 --> 00:14:17,290 That is what is a coin. 297 00:14:17,290 --> 00:14:19,480 When Satoshi Nakamoto did this, it wasn't really 298 00:14:19,480 --> 00:14:21,670 a coin, because it was a question whether anybody 299 00:14:21,670 --> 00:14:22,890 would give value. 300 00:14:22,890 --> 00:14:25,240 And until about 18 months later, when 301 00:14:25,240 --> 00:14:29,920 those two pizzas went for 10,000 bitcoin, what was it? 302 00:14:29,920 --> 00:14:34,590 Or until somebody started the first exchange, 303 00:14:34,590 --> 00:14:41,230 a crypto exchange, to exchange that. 304 00:14:41,230 --> 00:14:45,040 So you can also have multiple inputs and multiple outputs. 305 00:14:45,040 --> 00:14:48,610 And I'm going to use an example that I just created of, 306 00:14:48,610 --> 00:14:52,000 I want to send some bitcoins to two different people. 307 00:14:52,000 --> 00:14:55,240 And I need some bitcoins. 308 00:14:55,240 --> 00:14:58,310 So I might grab three former inputs. 309 00:14:58,310 --> 00:15:02,570 And these are just Transaction ID 6, Index 3. 310 00:15:02,570 --> 00:15:04,240 Of course, it wouldn't be ID 6. 311 00:15:04,240 --> 00:15:06,410 It might be ID 300,000. 312 00:15:06,410 --> 00:15:11,290 But the point being is, grab 10 bitcoin, 313 00:15:11,290 --> 00:15:15,140 so I want to find three inputs. 314 00:15:15,140 --> 00:15:20,650 And I needed to send six bitcoins, let's say, to Amanda. 315 00:15:20,650 --> 00:15:21,560 You like that. 316 00:15:21,560 --> 00:15:23,690 And just because you're sitting up front, James, 317 00:15:23,690 --> 00:15:25,220 I'm just going to send three. 318 00:15:25,220 --> 00:15:28,250 But I'm not using Amanda's name and James' name. 319 00:15:28,250 --> 00:15:30,150 I'm using your Bitcoin address-- 320 00:15:30,150 --> 00:15:35,270 Amanda's Bitcoin address, and James' Bitcoin address. 321 00:15:35,270 --> 00:15:37,690 So rather than being account-based, 322 00:15:37,690 --> 00:15:42,430 like if you have an account at Bank of America, and you say, 323 00:15:42,430 --> 00:15:46,120 I want to send $10, I just check to make 324 00:15:46,120 --> 00:15:48,070 sure I have more than $10. 325 00:15:48,070 --> 00:15:52,270 In Bitcoin, I actually have to fine individual transaction 326 00:15:52,270 --> 00:15:55,090 outputs that add up to-- in this case, 327 00:15:55,090 --> 00:16:00,790 I want to send nine bitcoin, six to Amanda, three to James. 328 00:16:00,790 --> 00:16:04,000 Not really, Amanda. 329 00:16:04,000 --> 00:16:06,145 So what's 10 minus 9? 330 00:16:06,145 --> 00:16:06,990 AUDIENCE: 1. 331 00:16:06,990 --> 00:16:07,990 GARY GENSLER: Thank you. 332 00:16:07,990 --> 00:16:09,610 That was really hard. 333 00:16:09,610 --> 00:16:11,800 We are at MIT. 334 00:16:11,800 --> 00:16:15,680 But I might not send one back to me. 335 00:16:15,680 --> 00:16:17,920 I'm going to send 0.9 back to me. 336 00:16:17,920 --> 00:16:21,180 This is kind of my change. 337 00:16:21,180 --> 00:16:25,900 So a Bitcoin transaction can either be equal, 338 00:16:25,900 --> 00:16:28,050 the inputs equal the outputs. 339 00:16:28,050 --> 00:16:30,750 I could have sent one back to me. 340 00:16:30,750 --> 00:16:34,290 But I've decided to incentivize miners and leave 341 00:16:34,290 --> 00:16:39,860 a little extra, 0.1 bitcoin, which would be a lot of fee, 342 00:16:39,860 --> 00:16:40,670 actually. 343 00:16:40,670 --> 00:16:45,403 That would be about, what, $640 or so? 344 00:16:45,403 --> 00:16:46,570 I probably wouldn't do that. 345 00:16:46,570 --> 00:16:54,030 I would probably leave for a fee 10 or 20 or 100 satoshi maybe. 346 00:16:54,030 --> 00:16:57,170 I don't know what the current market is. 347 00:16:57,170 --> 00:16:59,360 But this is a transaction. 348 00:16:59,360 --> 00:17:02,360 Multiple inputs, multiple outputs, but you always 349 00:17:02,360 --> 00:17:06,220 have to send back to yourself. 350 00:17:06,220 --> 00:17:09,010 What happens to these inputs if this transaction actually 351 00:17:09,010 --> 00:17:10,990 happens? 352 00:17:10,990 --> 00:17:12,420 They go away. 353 00:17:12,420 --> 00:17:18,810 The actual inputs disappear once they go through this. 354 00:17:18,810 --> 00:17:21,680 And so inputs always have to equal outputs. 355 00:17:21,680 --> 00:17:26,480 When a transaction is validated, one of the validation methods 356 00:17:26,480 --> 00:17:29,935 is to make sure that the lock time has actually happened, 357 00:17:29,935 --> 00:17:31,310 that you've passed the lock time. 358 00:17:31,310 --> 00:17:35,210 Another validation point is that inputs are greater than 359 00:17:35,210 --> 00:17:36,350 or equal to outputs. 360 00:17:36,350 --> 00:17:38,300 If outputs are greater than inputs, 361 00:17:38,300 --> 00:17:42,020 the transaction will not be validated. 362 00:17:42,020 --> 00:17:45,680 The digital signatures have to be validated, 363 00:17:45,680 --> 00:17:48,710 going just back to the prior slide. 364 00:17:48,710 --> 00:17:54,170 That digital signature has to be validated, as well. 365 00:17:54,170 --> 00:18:00,040 And that, this previous ID and index actually exists. 366 00:18:00,040 --> 00:18:03,050 And this is getting to your question Emily. 367 00:18:03,050 --> 00:18:05,940 It still exists. 368 00:18:05,940 --> 00:18:10,160 But these inputs, once you've used them, 369 00:18:10,160 --> 00:18:12,590 they no longer exist in the database. 370 00:18:12,590 --> 00:18:16,150 They're kind of in the past. 371 00:18:16,150 --> 00:18:17,950 So that's the transactions. 372 00:18:17,950 --> 00:18:18,730 That's the core. 373 00:18:18,730 --> 00:18:21,010 I know it's like eating broccoli, 374 00:18:21,010 --> 00:18:25,590 but it's an important part of all of blockchain technology. 375 00:18:25,590 --> 00:18:28,280 The Coinbase transaction we've already talked about, 376 00:18:28,280 --> 00:18:29,740 so I'll slide over this quickly. 377 00:18:29,740 --> 00:18:32,590 But it's a reward for solving the puzzle. 378 00:18:32,590 --> 00:18:35,170 In the case of Bitcoin, it's solving the proof of work. 379 00:18:35,170 --> 00:18:35,670 Tom. 380 00:18:35,670 --> 00:18:39,287 AUDIENCE: Sorry, can we go back to the miner incentive. 381 00:18:39,287 --> 00:18:43,830 It's the 0.1 or 1 satoshis. 382 00:18:43,830 --> 00:18:46,510 And how is that different than the cost of trust 383 00:18:46,510 --> 00:18:50,590 when you use a financial institution as the [INAUDIBLE]?? 384 00:18:50,590 --> 00:18:56,360 GARY GENSLER: So Tom's question is, the miners' economics. 385 00:18:56,360 --> 00:18:58,880 Of course, they get their 12 and 1/2 bitcoin, 386 00:18:58,880 --> 00:19:02,420 and they might get some transaction fees. 387 00:19:02,420 --> 00:19:04,130 Tom's question's the other side. 388 00:19:04,130 --> 00:19:07,130 Why is that any different than paying 389 00:19:07,130 --> 00:19:08,750 some central intermediary? 390 00:19:12,060 --> 00:19:15,060 Anybody want to try doing this? 391 00:19:15,060 --> 00:19:17,880 I could cold-call, but does anybody want to-- 392 00:19:17,880 --> 00:19:22,410 this is an economics question about markets. 393 00:19:22,410 --> 00:19:24,570 AUDIENCE: I would say that the person that 394 00:19:24,570 --> 00:19:27,720 is sending the transaction has the ability 395 00:19:27,720 --> 00:19:31,800 to choose how much it will pay for that transaction. 396 00:19:31,800 --> 00:19:37,860 And in the bank, it's regulated, and you pay. 397 00:19:37,860 --> 00:19:39,360 Maybe that's one of the differences. 398 00:19:39,360 --> 00:19:41,110 GARY GENSLER: All right, so one difference 399 00:19:41,110 --> 00:19:46,680 is that the bank is setting, generally, a fixed fee 400 00:19:46,680 --> 00:19:51,180 schedule, and this is a decentralized market 401 00:19:51,180 --> 00:19:53,880 mechanism for setting fees. 402 00:19:53,880 --> 00:19:54,525 Sean. 403 00:19:54,525 --> 00:19:58,080 AUDIENCE: The fee for intermediaries a lot higher 404 00:19:58,080 --> 00:20:01,320 compared to Bitcoin, the transaction fees. 405 00:20:01,320 --> 00:20:02,820 GARY GENSLER: So Sean's second point 406 00:20:02,820 --> 00:20:05,790 is that, currently-- this might not be true in the future. 407 00:20:05,790 --> 00:20:08,430 But currently, central intermediaries 408 00:20:08,430 --> 00:20:12,510 are able to charge higher fees than in this decentralized 409 00:20:12,510 --> 00:20:13,476 system. 410 00:20:13,476 --> 00:20:14,390 Alin? 411 00:20:14,390 --> 00:20:16,140 AUDIENCE: I think that, conceptually, it's 412 00:20:16,140 --> 00:20:20,180 not different from a centralized intermediary. 413 00:20:20,180 --> 00:20:22,590 It's just, the functions are the same. 414 00:20:22,590 --> 00:20:25,140 The amounts, maybe the amount that I pay, 415 00:20:25,140 --> 00:20:27,030 the amount of the fees are different, 416 00:20:27,030 --> 00:20:28,972 but the concepts are the same. 417 00:20:28,972 --> 00:20:31,180 GARY GENSLER: All right, so Alin is basically saying, 418 00:20:31,180 --> 00:20:32,597 well, maybe it's not so different. 419 00:20:32,597 --> 00:20:35,820 I mean, though it might be floating rather than fixed, 420 00:20:35,820 --> 00:20:38,437 it might be currently lower rather than higher, 421 00:20:38,437 --> 00:20:40,020 but you're saying, fundamentally, it's 422 00:20:40,020 --> 00:20:41,610 about the same. 423 00:20:41,610 --> 00:20:43,610 One minute, Shimon, let me just-- 424 00:20:43,610 --> 00:20:45,030 because you're faculty. 425 00:20:45,030 --> 00:20:47,090 You're going to actually tell us the answer. 426 00:20:47,090 --> 00:20:50,060 AUDIENCE: Question-- is this mandatory? 427 00:20:50,060 --> 00:20:52,310 GARY GENSLER: Very good question-- it's not mandatory. 428 00:20:52,310 --> 00:20:57,210 AUDIENCE: So that could create a big misalignment of interest. 429 00:20:57,210 --> 00:20:58,710 GARY GENSLER: So it's not mandatory. 430 00:20:58,710 --> 00:21:01,620 Fees are market-based. 431 00:21:01,620 --> 00:21:03,750 And at times, like last December, 432 00:21:03,750 --> 00:21:05,850 they were really high. 433 00:21:05,850 --> 00:21:09,030 And now they're quite low, partly 434 00:21:09,030 --> 00:21:14,670 because the Bitcoin network is not humming at full capacity. 435 00:21:14,670 --> 00:21:19,110 It can readily fit 1,000 to 2,000 transactions a block. 436 00:21:19,110 --> 00:21:21,270 And it's not like there's a jamming 437 00:21:21,270 --> 00:21:24,870 to get 10,000 and 20,000 transactions into a block, 438 00:21:24,870 --> 00:21:27,120 as there was last December. 439 00:21:27,120 --> 00:21:28,351 I'm sorry, Shimon. 440 00:21:31,790 --> 00:21:33,420 AUDIENCE: I think it's very different, 441 00:21:33,420 --> 00:21:35,880 in the sense that you're basically 442 00:21:35,880 --> 00:21:39,180 inflating the fee in most [INAUDIBLE] inflation. 443 00:21:39,180 --> 00:21:41,430 So even if when you don't transact, 444 00:21:41,430 --> 00:21:44,460 you're paying implicitly for other people's transactions, 445 00:21:44,460 --> 00:21:46,230 and that's going to change over time. 446 00:21:46,230 --> 00:21:49,110 So that's actually kind of a clever mechanism 447 00:21:49,110 --> 00:21:52,690 of how, structurally, these are going 448 00:21:52,690 --> 00:21:55,410 to shift over time in the network, 449 00:21:55,410 --> 00:21:57,720 as it hopefully matures. 450 00:21:57,720 --> 00:22:00,650 GARY GENSLER: Why don't we take one more here-- 451 00:22:00,650 --> 00:22:01,960 Alexis. 452 00:22:01,960 --> 00:22:08,070 AUDIENCE: That's like when you have [INAUDIBLE] they're 453 00:22:08,070 --> 00:22:09,420 going to keep the money. 454 00:22:09,420 --> 00:22:14,400 And they could use it for making [INAUDIBLE] or whatever. 455 00:22:14,400 --> 00:22:16,210 So they make money on a spread. 456 00:22:16,210 --> 00:22:18,990 Whereas here, for the case of Bitcoin, 457 00:22:18,990 --> 00:22:20,847 the money's going to stay in the system, 458 00:22:20,847 --> 00:22:22,680 and it's going to flow to another miner, who 459 00:22:22,680 --> 00:22:24,850 is going to use it for another transaction. 460 00:22:24,850 --> 00:22:28,530 It will stay within the same network, I would say, 461 00:22:28,530 --> 00:22:31,088 rather than just going to another destination. 462 00:22:31,088 --> 00:22:32,130 GARY GENSLER: I hear you. 463 00:22:32,130 --> 00:22:34,800 But it might actually leave the network to lawyers, 464 00:22:34,800 --> 00:22:37,430 like Patrick. 465 00:22:37,430 --> 00:22:40,040 Or it might leave the network if you use it at Starbucks, 466 00:22:40,040 --> 00:22:41,660 if Starbucks would accept it. 467 00:22:44,630 --> 00:22:47,420 So I'm challenging your thought, but you 468 00:22:47,420 --> 00:22:48,620 can challenge mine back. 469 00:22:52,355 --> 00:22:57,610 Why don't we close out with Eric, and then just move on. 470 00:22:57,610 --> 00:23:00,370 AUDIENCE: The difference might be in the perspective. 471 00:23:00,370 --> 00:23:04,140 From the person that's originated a transaction, 472 00:23:04,140 --> 00:23:05,770 it's basically the same thing. 473 00:23:05,770 --> 00:23:09,400 But if you think about it as a system, 474 00:23:09,400 --> 00:23:13,720 there's no single entity gathering all the money 475 00:23:13,720 --> 00:23:14,560 from transactions. 476 00:23:14,560 --> 00:23:17,230 You have a network, a [INAUDIBLE] network, 477 00:23:17,230 --> 00:23:21,140 of [INAUDIBLE] that are getting that. 478 00:23:21,140 --> 00:23:25,240 And besides, once we get out past the 21 million 479 00:23:25,240 --> 00:23:30,497 bitcoin generation cap, then all the systems 480 00:23:30,497 --> 00:23:32,080 must be using [INAUDIBLE] in some way. 481 00:23:32,080 --> 00:23:35,260 And transaction fees would be [INAUDIBLE].. 482 00:23:35,260 --> 00:23:40,510 GARY GENSLER: So Eric raises two points that I'll-- 483 00:23:40,510 --> 00:23:42,910 there are other points, but two that I want to repeat. 484 00:23:42,910 --> 00:23:47,380 One is, this is more decentralized possibly today 485 00:23:47,380 --> 00:23:51,670 than the current commercial banking system, for instance, 486 00:23:51,670 --> 00:23:55,900 for transaction processing, or the transaction processing 487 00:23:55,900 --> 00:23:59,650 that Visa and First Data do, which we'll 488 00:23:59,650 --> 00:24:03,010 study when we get to payments. 489 00:24:03,010 --> 00:24:05,028 So it's possibly more decentralized. 490 00:24:05,028 --> 00:24:06,820 I think you said it was more decentralized, 491 00:24:06,820 --> 00:24:08,710 but I'm putting the word possibly in there. 492 00:24:11,890 --> 00:24:16,030 And two is that, at least in bitcoin's case, 493 00:24:16,030 --> 00:24:20,350 there's two revenue sources for providing the services. 494 00:24:20,350 --> 00:24:23,480 The miners do it for the Coinbase transactions, 12 495 00:24:23,480 --> 00:24:27,880 and 1/2 bitcoin per block, approximately $80,000 US 496 00:24:27,880 --> 00:24:29,530 per block currently. 497 00:24:29,530 --> 00:24:32,110 But also, there's this little incentive 498 00:24:32,110 --> 00:24:39,580 of fees, which, over time, will have to grow as you go down. 499 00:24:39,580 --> 00:24:43,300 If it's only going to be one bitcoin per transaction, 500 00:24:43,300 --> 00:24:47,950 and then ultimately almost 0 bitcoin for transaction, 501 00:24:47,950 --> 00:24:50,980 there will have to be more satoshis in the fee side. 502 00:24:50,980 --> 00:24:53,200 And some alternative coins-- 503 00:24:53,200 --> 00:24:56,650 not Bitcoin-- are more modeled on fees, 504 00:24:56,650 --> 00:25:03,610 and some coins are more modeled on mining rewards. 505 00:25:03,610 --> 00:25:06,850 The economics-- Satoshi Nakamoto, 506 00:25:06,850 --> 00:25:11,410 whomever he or she was, or were, if it's a group, 507 00:25:11,410 --> 00:25:14,610 had to think through a bunch of economics. 508 00:25:14,610 --> 00:25:16,050 They've survived for 10 years. 509 00:25:16,050 --> 00:25:18,810 It doesn't mean that's the best set of microeconomics 510 00:25:18,810 --> 00:25:21,420 for a blockchain system. 511 00:25:21,420 --> 00:25:22,770 Tom, you look really skeptical. 512 00:25:22,770 --> 00:25:23,993 AUDIENCE: I'm skeptical. 513 00:25:23,993 --> 00:25:25,410 But maybe this would be the moment 514 00:25:25,410 --> 00:25:27,045 where I take my 12-hour dive. 515 00:25:27,045 --> 00:25:28,170 GARY GENSLER: That's right. 516 00:25:28,170 --> 00:25:31,720 So you're about ready for your rabbit dive? 517 00:25:31,720 --> 00:25:32,220 Maybe. 518 00:25:32,220 --> 00:25:32,887 AUDIENCE: Maybe. 519 00:25:35,432 --> 00:25:37,140 GARY GENSLER: So the Coinbase transaction 520 00:25:37,140 --> 00:25:38,940 we've talked a lot about. 521 00:25:38,940 --> 00:25:44,220 The reward, at least in Bitcoin, halves every 210,000 blocks. 522 00:25:44,220 --> 00:25:46,740 A very important thing that Nakamoto put in 523 00:25:46,740 --> 00:25:51,430 is, you couldn't use your Coinbase reward for 100 blocks. 524 00:25:51,430 --> 00:25:57,420 So it was sort of stale or frozen for 100 blocks. 525 00:25:57,420 --> 00:25:59,660 I can think of two reasons, but maybe you all 526 00:25:59,660 --> 00:26:03,060 would think of another reason. 527 00:26:03,060 --> 00:26:06,280 Anybody want to give it a shot as to why you might do it? 528 00:26:06,280 --> 00:26:07,950 Xiaojian 529 00:26:07,950 --> 00:26:09,480 AUDIENCE: Maybe you mine the block, 530 00:26:09,480 --> 00:26:12,980 and you're trying to mine a block that shouldn't be-- 531 00:26:12,980 --> 00:26:14,520 or it's not legal. 532 00:26:14,520 --> 00:26:17,210 You spend the money right away so you can get-- 533 00:26:17,210 --> 00:26:18,960 GARY GENSLER: This is the principal reason 534 00:26:18,960 --> 00:26:21,240 that's talked about, not only in the literature, 535 00:26:21,240 --> 00:26:26,660 but in the early blog post, was, well, how many blocks does 536 00:26:26,660 --> 00:26:29,370 the chain have to go before everybody really 537 00:26:29,370 --> 00:26:31,590 thinks it's consensus? 538 00:26:31,590 --> 00:26:34,830 You could have said 5, 10, 20. 539 00:26:34,830 --> 00:26:37,650 Satoshi picked 100 blocks, saying, 540 00:26:37,650 --> 00:26:42,160 that, hopefully, is pretty settled, 541 00:26:42,160 --> 00:26:45,930 or about 1,000 minutes. 542 00:26:45,930 --> 00:26:46,430 James. 543 00:26:46,430 --> 00:26:48,500 AUDIENCE: If you mine, and then you can spend, 544 00:26:48,500 --> 00:26:50,690 couldn't you just perpetually create blocks 545 00:26:50,690 --> 00:26:53,270 and then pay yourself and create more and earn more 546 00:26:53,270 --> 00:26:55,085 and then keep building up [INAUDIBLE]?? 547 00:26:55,085 --> 00:26:56,460 GARY GENSLER: So James is asking, 548 00:26:56,460 --> 00:26:58,130 could you just kind of game the system 549 00:26:58,130 --> 00:27:00,695 and keep mining and spending and mining and spending? 550 00:27:03,485 --> 00:27:05,125 AUDIENCE: Then your rewards go down. 551 00:27:05,125 --> 00:27:06,250 GARY GENSLER: Sorry, Aviva? 552 00:27:06,250 --> 00:27:08,933 AUDIENCE: Then your awards go down over time. 553 00:27:08,933 --> 00:27:10,850 GARY GENSLER: Well, that takes 200,000 blocks. 554 00:27:10,850 --> 00:27:12,980 AUDIENCE: But if there are many of you doing that 555 00:27:12,980 --> 00:27:16,330 at the same time, then you accelerate the-- 556 00:27:16,330 --> 00:27:18,830 GARY GENSLER: So in some ways, that's what miners are doing. 557 00:27:18,830 --> 00:27:21,770 But they have to wait 100 blocks. 558 00:27:21,770 --> 00:27:25,100 And so that was what Satoshi was trying to get at. 559 00:27:25,100 --> 00:27:30,000 If you have to wait 100 blocks, it's probably now the consensus 560 00:27:30,000 --> 00:27:31,170 chain. 561 00:27:31,170 --> 00:27:34,740 It's probably been so validated, unless we 562 00:27:34,740 --> 00:27:36,080 got into the problem-- 563 00:27:36,080 --> 00:27:37,200 Patrick, you weren't here. 564 00:27:37,200 --> 00:27:40,140 But some of the students raised the question, well, 565 00:27:40,140 --> 00:27:42,930 what if one country as large as China 566 00:27:42,930 --> 00:27:46,910 walled off their whole network, and just China 567 00:27:46,910 --> 00:27:51,470 went one way and the rest of the world went another? 568 00:27:51,470 --> 00:27:54,230 The theory of the case is that, within 100 blocks 569 00:27:54,230 --> 00:27:58,080 or 1,000 minutes, somehow that would be discovered. 570 00:27:58,080 --> 00:28:02,270 But if it weren't, you might have a little bit of what 571 00:28:02,270 --> 00:28:05,430 James is raising as a question. 572 00:28:05,430 --> 00:28:08,330 But that's at least the theory of the case. 573 00:28:11,290 --> 00:28:14,140 It's always recorded as the first transaction in the Merkle 574 00:28:14,140 --> 00:28:14,980 Tree. 575 00:28:14,980 --> 00:28:17,170 Highly technical point, but it has 576 00:28:17,170 --> 00:28:19,810 to roll up into that darn thing we were talking about, 577 00:28:19,810 --> 00:28:23,640 the data compression at the Merkle Tree. 578 00:28:23,640 --> 00:28:25,980 And here's a little fun fact. 579 00:28:25,980 --> 00:28:31,340 You can add 100 bytes of arbitrary data in a Coinbase. 580 00:28:31,340 --> 00:28:34,850 You might say, why does he raise this? 581 00:28:34,850 --> 00:28:37,460 Well, because it's just a fun little place 582 00:28:37,460 --> 00:28:43,310 that some people express their creative wit, artistic stuff, 583 00:28:43,310 --> 00:28:46,020 send secret messages to each other, 584 00:28:46,020 --> 00:28:49,920 that, buried in the Coinbase transactions, 585 00:28:49,920 --> 00:28:53,120 there is a whole forensics of fun little things 586 00:28:53,120 --> 00:28:56,240 that sometimes miners put in to the Coinbase, 587 00:28:56,240 --> 00:28:59,800 for those of you who are artistic. 588 00:28:59,800 --> 00:29:05,580 The very first genesis block had this sentence. 589 00:29:05,580 --> 00:29:09,990 "The Times, January 3, 2009, Chancellor on brink 590 00:29:09,990 --> 00:29:11,790 of second bailout for banks." 591 00:29:11,790 --> 00:29:13,920 That was a headline out of the Financial Times 592 00:29:13,920 --> 00:29:18,906 that says Satoshi Nakamoto put in the first block of Coinbase. 593 00:29:18,906 --> 00:29:21,400 It's just a little fun place. 594 00:29:21,400 --> 00:29:24,370 There's a playfulness that goes on amongst miners, sometimes 595 00:29:24,370 --> 00:29:25,660 talking to each other. 596 00:29:25,660 --> 00:29:26,830 Did you ever get a message? 597 00:29:26,830 --> 00:29:28,190 Did anybody send you-- 598 00:29:28,190 --> 00:29:29,800 PATRICK MURCK: Not that I'm aware of. 599 00:29:29,800 --> 00:29:31,175 I know what you're talking about. 600 00:29:31,175 --> 00:29:33,820 There's one miner that likes to put Catholic catechisms in. 601 00:29:33,820 --> 00:29:36,220 GARY GENSLER: That has put in the whole catechism? 602 00:29:36,220 --> 00:29:37,540 PATRICK MURCK: In every block, there's a little catechism 603 00:29:37,540 --> 00:29:38,720 that he puts in. 604 00:29:38,720 --> 00:29:39,880 It's a Allegis mining pool. 605 00:29:39,880 --> 00:29:42,160 It's a small one. 606 00:29:42,160 --> 00:29:43,040 So there you go. 607 00:29:43,040 --> 00:29:44,250 GARY GENSLER: There you go. 608 00:29:44,250 --> 00:29:46,930 And do they pay you in bitcoin, too? 609 00:29:46,930 --> 00:29:49,670 PATRICK MURCK: No, they're not a client. 610 00:29:49,670 --> 00:29:52,360 GARY GENSLER: OK. 611 00:29:52,360 --> 00:29:56,230 So it all rolls into a database called the unspent transaction 612 00:29:56,230 --> 00:29:57,050 output. 613 00:29:57,050 --> 00:29:59,020 These are the unspent transactions. 614 00:29:59,020 --> 00:30:01,840 If it's been spent, it's kind of burned. 615 00:30:01,840 --> 00:30:05,650 And bitcoin transactions that haven't been spent 616 00:30:05,650 --> 00:30:08,740 fall into this. 617 00:30:08,740 --> 00:30:11,110 And you can use it. 618 00:30:11,110 --> 00:30:13,840 It's created because it speeds up the whole system. 619 00:30:13,840 --> 00:30:16,930 Instead of going back and looking for all these things, 620 00:30:16,930 --> 00:30:18,430 there's actually a database that has 621 00:30:18,430 --> 00:30:21,220 all the unspent transactions. 622 00:30:21,220 --> 00:30:23,770 I include in here what I find as a sort 623 00:30:23,770 --> 00:30:27,670 of interesting revelation or irony. 624 00:30:27,670 --> 00:30:33,570 When Satoshi built Bitcoin, and for the 16 versions 625 00:30:33,570 --> 00:30:35,790 that have come since over the 10 years, 626 00:30:35,790 --> 00:30:39,060 all the developers, the Bitcoin core developers, 627 00:30:39,060 --> 00:30:41,790 have kept the unspent transaction output, 628 00:30:41,790 --> 00:30:46,080 not on a blockchain specifically, but in a database 629 00:30:46,080 --> 00:30:48,750 called a LevelDB database. 630 00:30:48,750 --> 00:30:51,900 So those of you who are closer to computer science than I 631 00:30:51,900 --> 00:30:55,860 could say all the pros and cons of a LevelDB database. 632 00:30:55,860 --> 00:30:57,990 But I'm just observing that, even 633 00:30:57,990 --> 00:31:06,040 within the most used first central database for blockchain 634 00:31:06,040 --> 00:31:10,210 called Bitcoin, they chose to use not a blockchain, 635 00:31:10,210 --> 00:31:13,900 but, in essence, a more standard database to keep 636 00:31:13,900 --> 00:31:17,260 the unspent transaction output. 637 00:31:17,260 --> 00:31:20,940 Now, in a sense, it's all part of this blockchain solution. 638 00:31:20,940 --> 00:31:24,130 I'm just saying, it's one database 639 00:31:24,130 --> 00:31:26,230 within the blockchain world that's 640 00:31:26,230 --> 00:31:28,333 actually not a blockchain. 641 00:31:30,900 --> 00:31:33,550 It's just sort of an interesting irony. 642 00:31:33,550 --> 00:31:36,370 But it also sort of says, economically 643 00:31:36,370 --> 00:31:39,700 and technologically, Satoshi was trying 644 00:31:39,700 --> 00:31:41,020 to create a money system. 645 00:31:41,020 --> 00:31:42,580 He wasn't trying to use blockchain 646 00:31:42,580 --> 00:31:45,680 for every bit of data. 647 00:31:45,680 --> 00:31:48,865 So this is the actual size of the unspent transaction output. 648 00:31:48,865 --> 00:31:49,990 If you can't see, there's-- 649 00:31:49,990 --> 00:31:53,830 I think it's about 50 to 60 million. 650 00:31:53,830 --> 00:31:55,610 It was higher. 651 00:31:55,610 --> 00:31:57,880 There was about 60 million unspent. 652 00:31:57,880 --> 00:32:00,250 It's not 60 million bitcoins, because there's 653 00:32:00,250 --> 00:32:02,770 about 17 million bitcoin. 654 00:32:02,770 --> 00:32:04,310 So you could average it out. 655 00:32:04,310 --> 00:32:06,190 You could say, well, each transaction 656 00:32:06,190 --> 00:32:08,380 has less than 1 bitcoin. 657 00:32:08,380 --> 00:32:10,990 Well actually, there's been surveys and studies showing 658 00:32:10,990 --> 00:32:14,470 that about half of these 54 million transactions are 659 00:32:14,470 --> 00:32:18,990 so small that they go by the term of-- they're called dust. 660 00:32:18,990 --> 00:32:22,360 That there's so few satoshi that it's not 661 00:32:22,360 --> 00:32:26,130 even worth the fees to try to redeem them. 662 00:32:26,130 --> 00:32:29,040 They only add up to less than a half a percent 663 00:32:29,040 --> 00:32:34,230 of all the outstanding bitcoin, but they're just dust. 664 00:32:34,230 --> 00:32:39,000 So maybe out of these 54 million unspent transaction outputs, 665 00:32:39,000 --> 00:32:42,060 half of them will never be used, because it's not 666 00:32:42,060 --> 00:32:43,800 economically worthwhile. 667 00:32:43,800 --> 00:32:46,800 It's like the pennies in the top dresser drawer 668 00:32:46,800 --> 00:32:50,560 that you all might not spend. 669 00:32:50,560 --> 00:32:53,700 There's this similar thing here. 670 00:32:53,700 --> 00:32:54,330 Hugo. 671 00:32:54,330 --> 00:32:56,755 AUDIENCE: Yeah, I guess I have a question about that. 672 00:32:56,755 --> 00:32:58,380 It might not be feasible now, but maybe 673 00:32:58,380 --> 00:33:01,320 with layers on top of bitcoin that people can 674 00:33:01,320 --> 00:33:03,220 do these micro-transactions. 675 00:33:03,220 --> 00:33:05,730 Because like, our pennies might be worth a lot 676 00:33:05,730 --> 00:33:06,997 of money, [INAUDIBLE]. 677 00:33:06,997 --> 00:33:08,580 GARY GENSLER: So Hugo is raising that, 678 00:33:08,580 --> 00:33:10,620 just like the pennies in your top dresser drawer 679 00:33:10,620 --> 00:33:12,840 might be worth something one day, 680 00:33:12,840 --> 00:33:16,200 what I'm referencing as Bitcoin dust, about half 681 00:33:16,200 --> 00:33:20,250 of these unspent transaction outputs, a satoshi here, 682 00:33:20,250 --> 00:33:23,910 10 satoshi there, might one day be worth something. 683 00:33:23,910 --> 00:33:25,170 Good point. 684 00:33:25,170 --> 00:33:27,960 It's just like the pennies in your top dresser drawer though. 685 00:33:27,960 --> 00:33:30,500 Have you lost them in the meantime? 686 00:33:30,500 --> 00:33:31,750 They might be worth something. 687 00:33:31,750 --> 00:33:34,170 But in the meantime, have you lost the private key 688 00:33:34,170 --> 00:33:37,530 to those little satoshis? 689 00:33:37,530 --> 00:33:39,570 I put on here three moments of time 690 00:33:39,570 --> 00:33:41,940 just to give you the sense of the actual number 691 00:33:41,940 --> 00:33:43,950 of transactions that have happened. 692 00:33:43,950 --> 00:33:47,550 There's been 342 million transactions on the Bitcoin 693 00:33:47,550 --> 00:33:50,340 network to date, or as of a day or two ago 694 00:33:50,340 --> 00:33:53,210 when I put the slides together. 695 00:33:53,210 --> 00:33:55,760 So of the 340 million transactions, 696 00:33:55,760 --> 00:33:59,390 only about 54 million are still outstanding. 697 00:33:59,390 --> 00:34:05,760 The other 290 million have been spent, if you wish. 698 00:34:08,800 --> 00:34:09,300 Yes. 699 00:34:09,300 --> 00:34:12,129 AUDIENCE: So then where are these outstanding transactions 700 00:34:12,129 --> 00:34:12,629 stored? 701 00:34:16,409 --> 00:34:18,895 Are they still being included in the blocks themselves, 702 00:34:18,895 --> 00:34:20,520 where you add them to the [INAUDIBLE]?? 703 00:34:20,520 --> 00:34:23,020 GARY GENSLER: Which outstanding transactions, the 54 million 704 00:34:23,020 --> 00:34:24,870 that are still available? 705 00:34:24,870 --> 00:34:32,420 The 54 million all reside in a database within the Bitcoin 706 00:34:32,420 --> 00:34:37,429 software called the unspent transaction output, UTXO. 707 00:34:37,429 --> 00:34:41,540 And UTXO-- these aren't letters I'm making up-- 708 00:34:41,540 --> 00:34:46,600 that's a database, 54 million transactions. 709 00:34:46,600 --> 00:34:50,830 Separately, they are actually in the blockchain itself. 710 00:34:50,830 --> 00:34:54,940 So all 340 million transactions that have ever happened 711 00:34:54,940 --> 00:34:57,300 are in the blockchain. 712 00:34:57,300 --> 00:35:01,680 But to make it easier for the software, the 54 million 713 00:35:01,680 --> 00:35:08,220 that have never been spent reside in a software. 714 00:35:08,220 --> 00:35:10,710 Does that answer the question? 715 00:35:10,710 --> 00:35:11,300 Work with me. 716 00:35:11,300 --> 00:35:12,967 AUDIENCE: So it's a distributed database 717 00:35:12,967 --> 00:35:14,550 amongst all the different nodes. 718 00:35:14,550 --> 00:35:15,467 GARY GENSLER: Correct. 719 00:35:17,700 --> 00:35:22,200 All 10,000 of the nodes can have the full UTXO set. 720 00:35:22,200 --> 00:35:25,530 Some wallet providers have the full UTXO set, 721 00:35:25,530 --> 00:35:27,740 but they don't have to. 722 00:35:27,740 --> 00:35:33,043 Somewhat lightweight nodes usually don't, but can. 723 00:35:33,043 --> 00:35:34,460 But a lightweight node would never 724 00:35:34,460 --> 00:35:37,550 want to have all 340 million and the full blockchain. 725 00:35:37,550 --> 00:35:41,300 So in essence, they're in multiple places, 726 00:35:41,300 --> 00:35:46,650 because they're in the full blockchain, 10,000 nodes. 727 00:35:46,650 --> 00:35:50,120 And they're also in the UTXO, not only on the 10,000 nodes, 728 00:35:50,120 --> 00:35:51,770 but occasionally elsewhere. 729 00:35:51,770 --> 00:35:53,450 Alin. 730 00:35:53,450 --> 00:35:56,810 AUDIENCE: So I heard you say the word spent transaction, 731 00:35:56,810 --> 00:35:58,610 which is a bit misleading. 732 00:35:58,610 --> 00:36:00,900 Because a transaction can be spent and not spent. 733 00:36:00,900 --> 00:36:03,440 Because, for example, you would have two outputs. 734 00:36:03,440 --> 00:36:05,990 One output is spent by a future transaction, 735 00:36:05,990 --> 00:36:08,070 and the other output is not spent. 736 00:36:08,070 --> 00:36:10,222 So it's a bit misleading to say spent transaction, 737 00:36:10,222 --> 00:36:12,680 because that only happens when the transaction has only one 738 00:36:12,680 --> 00:36:14,480 output and that output is spent. 739 00:36:14,480 --> 00:36:16,730 GARY GENSLER: Well, I'm using it lightly. 740 00:36:16,730 --> 00:36:20,570 I'm saying that, of the 340 million transactions that 741 00:36:20,570 --> 00:36:23,105 have happened, 290-- 742 00:36:23,105 --> 00:36:24,730 AUDIENCE: Have an output that is spent, 743 00:36:24,730 --> 00:36:27,210 or what are you saying exactly? 744 00:36:27,210 --> 00:36:29,310 Because transactions have multiple outputs. 745 00:36:29,310 --> 00:36:32,150 They might have n, and maybe k of them are spent. 746 00:36:32,150 --> 00:36:33,930 So the other outputs are unspent. 747 00:36:33,930 --> 00:36:36,300 GARY GENSLER: I'm saying there's 340 million-- 748 00:36:36,300 --> 00:36:40,440 if I did my data search correctly-- and I'm fallible, 749 00:36:40,440 --> 00:36:42,340 so I might not have. 750 00:36:42,340 --> 00:36:44,100 But if I did my data source correctly, 751 00:36:44,100 --> 00:36:47,490 there was 340 million previous outputs. 752 00:36:47,490 --> 00:36:49,990 290 million of them are gone. 753 00:36:49,990 --> 00:36:52,770 AUDIENCE: OK, so then you should say transaction outputs, 754 00:36:52,770 --> 00:36:55,200 because those are different than transactions. 755 00:36:55,200 --> 00:36:55,950 GARY GENSLER: Yes. 756 00:36:55,950 --> 00:36:59,040 Except for it was easier to put TXS. 757 00:36:59,040 --> 00:37:03,243 But yes, Alin's clarification is, I believe, accurate. 758 00:37:03,243 --> 00:37:05,160 AUDIENCE: [INAUDIBLE] so people can understand 759 00:37:05,160 --> 00:37:07,260 that there is a difference between a transaction 760 00:37:07,260 --> 00:37:08,430 and a transaction output. 761 00:37:08,430 --> 00:37:10,710 GARY GENSLER: In essence, what Alin's saying is, 762 00:37:10,710 --> 00:37:15,180 there's currently 54 million transaction outputs 763 00:37:15,180 --> 00:37:19,360 in the UTXO, which also says outputs. 764 00:37:19,360 --> 00:37:23,320 There had been, in the past, another 290 million outputs 765 00:37:23,320 --> 00:37:27,435 that have already been spent. 766 00:37:27,435 --> 00:37:28,750 Are we together? 767 00:37:31,500 --> 00:37:33,200 So there's a scripting language. 768 00:37:33,200 --> 00:37:36,650 There's a little bit of computer code. 769 00:37:36,650 --> 00:37:38,960 I said there was no prerequisite to take any computer 770 00:37:38,960 --> 00:37:40,880 science before you were here. 771 00:37:40,880 --> 00:37:44,130 And my own computer programming is so old, 772 00:37:44,130 --> 00:37:47,780 because when I was programming, it was in Fortran and APL. 773 00:37:47,780 --> 00:37:48,950 And you can look that up. 774 00:37:48,950 --> 00:37:52,100 It's kind of like around with cuneiform 775 00:37:52,100 --> 00:37:57,350 and you know the Rosetta Stone. 776 00:37:57,350 --> 00:38:02,990 But Satoshi Nakamoto decided to put a little bit 777 00:38:02,990 --> 00:38:06,650 of computer programming inside. 778 00:38:06,650 --> 00:38:08,390 And I'm not going to get the count right, 779 00:38:08,390 --> 00:38:11,030 but there's several hundred, but not several thousand, 780 00:38:11,030 --> 00:38:13,130 little operations and codes that you 781 00:38:13,130 --> 00:38:17,390 can use in the Bitcoin script. 782 00:38:17,390 --> 00:38:19,940 It's not Turing complete, which means 783 00:38:19,940 --> 00:38:23,840 you can't do a lot of the things that you can do in all 784 00:38:23,840 --> 00:38:26,730 the rest of computer science. 785 00:38:26,730 --> 00:38:29,340 But it's more secure. 786 00:38:29,340 --> 00:38:31,980 In essence, it has fewer attack vectors. 787 00:38:31,980 --> 00:38:36,300 It's harder to bring down a little bit. 788 00:38:36,300 --> 00:38:39,570 It's a programming code, as I said. 789 00:38:39,570 --> 00:38:42,090 For those interested, it's called stack-based, 790 00:38:42,090 --> 00:38:46,260 where you sort of move the code over one at a time 791 00:38:46,260 --> 00:38:49,320 as it's being performed. 792 00:38:49,320 --> 00:38:50,940 And it gives some flexibility. 793 00:38:50,940 --> 00:38:53,940 And back to Emily's question about why there was lock time, 794 00:38:53,940 --> 00:38:59,700 or Shimon's, scripting code allows for some conditionality, 795 00:38:59,700 --> 00:39:03,240 that it appears that Nakamoto was trying 796 00:39:03,240 --> 00:39:06,270 to give some ability to condition 797 00:39:06,270 --> 00:39:14,070 a transaction on events, but not so much conditionality, so 798 00:39:14,070 --> 00:39:17,550 much flexibility, that he needed a Turing complete. 799 00:39:17,550 --> 00:39:20,550 So he kind of, I'm going to say, chose a midway place. 800 00:39:23,130 --> 00:39:26,430 I believe, you could have created Bitcoin 801 00:39:26,430 --> 00:39:28,980 and say there was no scripting language. 802 00:39:28,980 --> 00:39:34,750 It was just going to be a straight instruction, moving 803 00:39:34,750 --> 00:39:36,940 this input to another output. 804 00:39:36,940 --> 00:39:42,100 Created a little bit of computer code, but not a lot. 805 00:39:42,100 --> 00:39:45,970 That's what I think of the economics and the marketplace 806 00:39:45,970 --> 00:39:46,900 for this. 807 00:39:46,900 --> 00:39:50,930 And next Tuesday when we talk about smart contracts-- 808 00:39:50,930 --> 00:39:53,140 and I promise you, there's a reason 809 00:39:53,140 --> 00:39:55,390 for the craziness of my talking about Turing 810 00:39:55,390 --> 00:39:57,640 complete and scripting code. 811 00:39:57,640 --> 00:40:01,180 Because next Tuesday, we'll be talking about smart contracts 812 00:40:01,180 --> 00:40:03,910 where they're much more flexible. 813 00:40:03,910 --> 00:40:06,712 And so this is sort of the foundational-- 814 00:40:06,712 --> 00:40:08,170 and you don't need to know anything 815 00:40:08,170 --> 00:40:10,660 more about computer science than you want, 816 00:40:10,660 --> 00:40:12,910 unless you go with Tom down that rabbit hole 817 00:40:12,910 --> 00:40:14,590 and spend more time reading. 818 00:40:17,890 --> 00:40:22,180 So there's four different types of, I call them script types. 819 00:40:22,180 --> 00:40:25,300 They're not actual script words, but you will read 820 00:40:25,300 --> 00:40:27,080 about these from time to time. 821 00:40:27,080 --> 00:40:29,390 And I just wanted to cover these four. 822 00:40:29,390 --> 00:40:32,650 The UTXO, remember, is about 54 million transactions. 823 00:40:32,650 --> 00:40:34,480 And there's been a nice academic paper 824 00:40:34,480 --> 00:40:38,020 that I didn't assign that was written earlier this year 825 00:40:38,020 --> 00:40:41,770 that investigated the whole 54 million, 826 00:40:41,770 --> 00:40:44,470 all of the unspent transactions. 827 00:40:44,470 --> 00:40:45,820 And this is how it broke down. 828 00:40:45,820 --> 00:40:53,050 81% are transactions that send to a hash of a Bitcoin address. 829 00:40:53,050 --> 00:40:56,930 Eight, nine years ago when Satoshi created this, 830 00:40:56,930 --> 00:40:59,210 that was not the most popular instruction. 831 00:40:59,210 --> 00:41:00,860 But it's basically sending an output 832 00:41:00,860 --> 00:41:05,120 to the hash, the compression, the commitment of a Bitcoin 833 00:41:05,120 --> 00:41:07,050 address. 834 00:41:07,050 --> 00:41:09,210 We're now up to 18%. 835 00:41:09,210 --> 00:41:12,240 This didn't exist three and four years ago, really. 836 00:41:12,240 --> 00:41:16,080 But 18% go to a conditional script. 837 00:41:16,080 --> 00:41:18,420 It's a hash of a conditional script. 838 00:41:18,420 --> 00:41:21,840 So somebody saying, Emily, it's not even about time. 839 00:41:21,840 --> 00:41:25,200 It's like, you can only get it when 840 00:41:25,200 --> 00:41:27,810 all these other instructions that are in the scripting 841 00:41:27,810 --> 00:41:30,650 language happen. 842 00:41:30,650 --> 00:41:33,790 And I'm going to hide the conditions in a hash of it. 843 00:41:37,300 --> 00:41:43,720 And then only 0.1% goes the way that he first 844 00:41:43,720 --> 00:41:47,330 envisioned nine years ago, directly to a Bitcoin address. 845 00:41:47,330 --> 00:41:49,840 So it's either to a hash of a Bitcoin address, 846 00:41:49,840 --> 00:41:53,320 a hash of a conditional script, and a little less than 1% now 847 00:41:53,320 --> 00:41:55,900 go to multiple signatures. 848 00:41:55,900 --> 00:42:00,150 Meaning, you need two out of three or three out of five. 849 00:42:00,150 --> 00:42:03,880 Or believe it or not, this academic paper 850 00:42:03,880 --> 00:42:06,730 shows that some say 0 out of 1. 851 00:42:06,730 --> 00:42:09,790 Now, it's hard to believe that somebody mistakenly programmed 852 00:42:09,790 --> 00:42:14,020 something to go to 0 signatures, but apparently somebody did. 853 00:42:14,020 --> 00:42:16,210 So I just wanted to give you a sense there's 854 00:42:16,210 --> 00:42:21,310 some flexibility in the computer code, not a lot of flexibility, 855 00:42:21,310 --> 00:42:26,560 but just enough that you can do things that are really helpful. 856 00:42:26,560 --> 00:42:30,730 And they're going to solve a lot of challenges for Bitcoin. 857 00:42:30,730 --> 00:42:32,410 Hugo mentioned layer 2. 858 00:42:32,410 --> 00:42:35,470 We're going to be talking about layer 2 later in the semester, 859 00:42:35,470 --> 00:42:39,340 where there's a whole way to put technology on top of Bitcoin. 860 00:42:39,340 --> 00:42:41,350 And it's because the scripting language is there 861 00:42:41,350 --> 00:42:42,400 that you can do that. 862 00:42:42,400 --> 00:42:43,690 Any questions on script? 863 00:42:43,690 --> 00:42:50,830 I know I'm trying to cover a big, weighty topic in 120 864 00:42:50,830 --> 00:42:51,635 seconds or less. 865 00:42:55,648 --> 00:42:56,148 No? 866 00:42:58,990 --> 00:43:01,090 So just back to the whole-- 867 00:43:01,090 --> 00:43:02,020 this is just a review. 868 00:43:02,020 --> 00:43:04,030 What have we've talked about? 869 00:43:04,030 --> 00:43:06,610 There's that little graphic again. 870 00:43:06,610 --> 00:43:08,830 It's just a bunch of blocks. 871 00:43:08,830 --> 00:43:10,870 That's what a blockchain is. 872 00:43:10,870 --> 00:43:13,960 Though today, we realize that underneath the blocks, 873 00:43:13,960 --> 00:43:16,170 we have another chain. 874 00:43:16,170 --> 00:43:17,740 I often think of two chains-- 875 00:43:17,740 --> 00:43:20,020 the chains of blocks. 876 00:43:20,020 --> 00:43:24,590 In Bitcoin, there's about a half a million blocks. 877 00:43:24,590 --> 00:43:27,080 But underneath that, there's all the transactions that 878 00:43:27,080 --> 00:43:28,970 are, in fact, chained, as well. 879 00:43:28,970 --> 00:43:32,720 54 million of those outputs have not yet been spent 880 00:43:32,720 --> 00:43:37,280 and 290 million outputs have been spent. 881 00:43:37,280 --> 00:43:40,190 But underneath about a half a million blocks, 882 00:43:40,190 --> 00:43:47,840 there's been 340 million outputs, so to speak. 883 00:43:47,840 --> 00:43:49,680 It creates a database. 884 00:43:49,680 --> 00:43:53,220 Bitcoin is a transaction database. 885 00:43:53,220 --> 00:43:56,700 Next Tuesday, we'll talk about an account-based database 886 00:43:56,700 --> 00:43:58,140 in Ethereum. 887 00:43:58,140 --> 00:44:00,390 But it could be a ledger which is transactions 888 00:44:00,390 --> 00:44:03,420 or a ledger which is balances. 889 00:44:03,420 --> 00:44:06,810 Satoshi Nakamoto decided to do transactions here. 890 00:44:06,810 --> 00:44:10,410 In some ways, I believe it's because it was fewer attack 891 00:44:10,410 --> 00:44:12,330 vectors. 892 00:44:12,330 --> 00:44:14,030 Probably a little bit more secure, 893 00:44:14,030 --> 00:44:16,320 but I'm not entirely sure. 894 00:44:16,320 --> 00:44:21,180 And until you solve the riddle as to who Satoshi Nakamoto is, 895 00:44:21,180 --> 00:44:23,940 we won't know the answer. 896 00:44:23,940 --> 00:44:24,440 Of 897 00:44:24,440 --> 00:44:27,410 Course, hash functions we talked about, 898 00:44:27,410 --> 00:44:30,890 and digital signatures, and a consensus protocol. 899 00:44:30,890 --> 00:44:35,660 So I like to think of it in three buckets, 900 00:44:35,660 --> 00:44:38,060 whether it's for a dinner party conversation 901 00:44:38,060 --> 00:44:40,700 or digging into three lectures. 902 00:44:40,700 --> 00:44:42,740 It's the cryptography itself. 903 00:44:42,740 --> 00:44:45,288 We did that last Thursday. 904 00:44:45,288 --> 00:44:46,830 And if you have to remember anything, 905 00:44:46,830 --> 00:44:49,110 it's only two cryptographic primitives-- 906 00:44:49,110 --> 00:44:51,870 hash functions and digital signatures. 907 00:44:51,870 --> 00:44:54,600 How many people think they kind of roughly have 908 00:44:54,600 --> 00:44:58,070 what a hash function is? 909 00:44:58,070 --> 00:44:59,570 All right, so I lost half of you. 910 00:44:59,570 --> 00:45:00,450 [LAUGHTER] 911 00:45:00,450 --> 00:45:03,910 All right, is there anything I can do for Lauren's table? 912 00:45:03,910 --> 00:45:05,360 I didn't see a single hand go up. 913 00:45:08,930 --> 00:45:10,640 Are you reading your Facebook page 914 00:45:10,640 --> 00:45:12,500 or are you listening to the class? 915 00:45:12,500 --> 00:45:13,760 You've got your computer open. 916 00:45:13,760 --> 00:45:14,580 What's your name. 917 00:45:14,580 --> 00:45:15,330 AUDIENCE: Matthew. 918 00:45:15,330 --> 00:45:16,880 GARY GENSLER: Matthew. 919 00:45:16,880 --> 00:45:17,770 All right, that might have been that you 920 00:45:17,770 --> 00:45:19,062 weren't listening to the class. 921 00:45:19,062 --> 00:45:20,830 Thank you. 922 00:45:20,830 --> 00:45:23,350 How can I help in what a hash function is? 923 00:45:23,350 --> 00:45:26,900 My promise is to bring everybody along. 924 00:45:26,900 --> 00:45:29,540 Nobody at your table said you even roughly got 925 00:45:29,540 --> 00:45:30,720 what a hash function is. 926 00:45:35,520 --> 00:45:37,020 I'm not trying to embarrass anybody. 927 00:45:37,020 --> 00:45:40,150 I'm trying to work this through. 928 00:45:40,150 --> 00:45:44,920 Nicholas, so give me a baseline. 929 00:45:44,920 --> 00:45:47,570 Did you read any of the readings? 930 00:45:47,570 --> 00:45:48,470 Maybe not. 931 00:45:48,470 --> 00:45:49,176 OK. 932 00:45:49,176 --> 00:45:50,480 AUDIENCE: I have, yes. 933 00:45:50,480 --> 00:45:52,760 GARY GENSLER: You have, OK. 934 00:45:52,760 --> 00:45:56,690 Hash functions came along decades ago 935 00:45:56,690 --> 00:46:00,360 to help facilitate database management. 936 00:46:00,360 --> 00:46:03,590 Sometimes it's called a registry. 937 00:46:03,590 --> 00:46:07,520 It's taking a lot of data and shrinking it down, 938 00:46:07,520 --> 00:46:11,780 compressing it, shrinking it, to maybe a series of numbers. 939 00:46:11,780 --> 00:46:17,810 I think of it sometimes as a zip code for information. 940 00:46:17,810 --> 00:46:21,330 Down in Baltimore, Maryland, I'm in 2120-- 941 00:46:21,330 --> 00:46:24,290 well, I grew up in 21208. 942 00:46:24,290 --> 00:46:25,700 It was my parent's zip code. 943 00:46:25,700 --> 00:46:28,370 I won't say my current. 944 00:46:28,370 --> 00:46:29,654 We're being videoed. 945 00:46:29,654 --> 00:46:31,430 [LAUGHTER] 946 00:46:31,430 --> 00:46:34,490 And so I think of it a little bit like that. 947 00:46:34,490 --> 00:46:37,610 So a hash function that has nothing to do with bitcoin 948 00:46:37,610 --> 00:46:42,440 came along to take a bunch of data and create a registry. 949 00:46:42,440 --> 00:46:46,580 But through that, it also became a way to do a commitment. 950 00:46:46,580 --> 00:46:47,810 Yes, your first name? 951 00:46:47,810 --> 00:46:48,768 AUDIENCE: Dana. 952 00:46:48,768 --> 00:46:49,560 GARY GENSLER: Dana? 953 00:46:49,560 --> 00:46:50,435 AUDIENCE: Yeah, Dana. 954 00:46:50,435 --> 00:46:53,510 What goes into the hash function and comes out as a hash, 955 00:46:53,510 --> 00:46:54,920 that's what I don't understand. 956 00:46:54,920 --> 00:46:57,950 GARY GENSLER: So what goes in is any set of data. 957 00:46:57,950 --> 00:47:01,060 Today, that could be an entire movie. 958 00:47:01,060 --> 00:47:03,960 It could be a picture of everybody in this room. 959 00:47:03,960 --> 00:47:12,940 Initially, it was mostly alphanumeric data. 960 00:47:12,940 --> 00:47:15,570 But because, in computer technology, 961 00:47:15,570 --> 00:47:21,570 all data can be broken down to a series of registries 962 00:47:21,570 --> 00:47:22,860 of 0's or 1's-- 963 00:47:28,640 --> 00:47:33,020 computers started with-- literally the first one 964 00:47:33,020 --> 00:47:38,780 started with registries that were either turned on or off. 965 00:47:38,780 --> 00:47:40,880 If they were on, call that a 1. 966 00:47:40,880 --> 00:47:42,590 If it was off, call it a 0. 967 00:47:42,590 --> 00:47:45,110 I'm not sure which way it goes. 968 00:47:45,110 --> 00:47:48,500 I keep looking at Alin. 969 00:47:48,500 --> 00:47:52,500 And so all data can then be brought down 970 00:47:52,500 --> 00:47:55,130 to a series of 0's and 1. 971 00:47:55,130 --> 00:47:59,600 And if you put four 0's and 1's in front of each other, 2 times 972 00:47:59,600 --> 00:48:05,450 2 times 2 times 2, 2 to the fourth is 16, all of a sudden, 973 00:48:05,450 --> 00:48:07,820 you see if you keep going 2 to something, 974 00:48:07,820 --> 00:48:10,140 you can get a lot of data. 975 00:48:10,140 --> 00:48:13,230 So sit back to answer your question 976 00:48:13,230 --> 00:48:16,380 Dana, when we talked last week about The New York Times 977 00:48:16,380 --> 00:48:17,370 crossword puzzle-- 978 00:48:17,370 --> 00:48:20,190 The New York Times may, if they wish, 979 00:48:20,190 --> 00:48:25,860 take the solution of their crossword puzzle and hash it. 980 00:48:25,860 --> 00:48:28,505 And then Stephanie? 981 00:48:28,505 --> 00:48:29,130 AUDIENCE: Yeah. 982 00:48:29,130 --> 00:48:31,380 GARY GENSLER: Stephanie likes to do The New York Times 983 00:48:31,380 --> 00:48:33,370 crossword puzzle. 984 00:48:33,370 --> 00:48:36,880 And she wants to know if she properly completed The New York 985 00:48:36,880 --> 00:48:38,380 Times crossword puzzle. 986 00:48:38,380 --> 00:48:42,400 The New York Times could say to her cell phone, 987 00:48:42,400 --> 00:48:45,490 we're not going to really give you the answer, 988 00:48:45,490 --> 00:48:49,232 but we'll give you the hash of the answer. 989 00:48:49,232 --> 00:48:51,440 And then when she's finished and she pushes a button, 990 00:48:51,440 --> 00:48:57,400 her application could say whether her answer properly 991 00:48:57,400 --> 00:48:59,510 hashes to theirs. 992 00:48:59,510 --> 00:49:00,870 AUDIENCE: [INAUDIBLE] 993 00:49:00,870 --> 00:49:01,180 GARY GENSLER: Please. 994 00:49:01,180 --> 00:49:02,660 We're trying to learn here together. 995 00:49:02,660 --> 00:49:04,243 AUDIENCE: In the case of a blockchain, 996 00:49:04,243 --> 00:49:06,730 it's whatever dated you loaded to the blockchain, 997 00:49:06,730 --> 00:49:09,150 and then in Bitcoin, it's just the transactions. 998 00:49:09,150 --> 00:49:10,570 Is that right? 999 00:49:10,570 --> 00:49:14,260 GARY GENSLER: So in Bitcoin and blockchain, 1000 00:49:14,260 --> 00:49:18,550 they use hash functions in several big ways. 1001 00:49:21,250 --> 00:49:24,220 Everything you said was correct, except for the one 1002 00:49:24,220 --> 00:49:26,320 thing when you said, it's just. 1003 00:49:26,320 --> 00:49:30,310 Because they actually use hash functions in the middle 1004 00:49:30,310 --> 00:49:31,330 of the proof of work. 1005 00:49:34,210 --> 00:49:37,540 Because the hash pointer points-- 1006 00:49:37,540 --> 00:49:41,480 block 3 points to block 2 and there's hashing. 1007 00:49:41,480 --> 00:49:45,220 They use the hash function to compress a bunch of data, what 1008 00:49:45,220 --> 00:49:47,200 I call the Merkle Tree, but it's taking 1009 00:49:47,200 --> 00:49:52,990 1,500 or 2,000 transactions and squeezing it into one hash. 1010 00:49:52,990 --> 00:49:56,160 So they're hashing all of these things up. 1011 00:49:56,160 --> 00:50:01,550 It uses hash functions in the midst of the Bitcoin address. 1012 00:50:01,550 --> 00:50:05,140 So the hash function is like electricity 1013 00:50:05,140 --> 00:50:07,000 in the middle of it, almost. 1014 00:50:07,000 --> 00:50:10,090 It's probably used six or eight places, 1015 00:50:10,090 --> 00:50:13,690 and some I don't, in any way, know myself 1016 00:50:13,690 --> 00:50:15,687 or need to understand. 1017 00:50:15,687 --> 00:50:16,770 Nicolas, how are we doing? 1018 00:50:16,770 --> 00:50:17,930 Did we get a little closer? 1019 00:50:17,930 --> 00:50:18,660 AUDIENCE: Yes. 1020 00:50:18,660 --> 00:50:20,702 GARY GENSLER: Lauren, did we get a little closer? 1021 00:50:20,702 --> 00:50:21,408 AUDIENCE: Yeah. 1022 00:50:21,408 --> 00:50:22,200 GARY GENSLER: Matt? 1023 00:50:22,200 --> 00:50:22,992 AUDIENCE: Oh, yeah. 1024 00:50:22,992 --> 00:50:24,457 GARY GENSLER: You're there. 1025 00:50:24,457 --> 00:50:26,290 AUDIENCE: I'm not there yet, but I'm closer. 1026 00:50:26,290 --> 00:50:27,020 GARY GENSLER: Stop by. 1027 00:50:27,020 --> 00:50:27,860 Send me an email. 1028 00:50:27,860 --> 00:50:29,598 It's gensler@mit.edu. 1029 00:50:29,598 --> 00:50:30,890 I'm here like four days a week. 1030 00:50:30,890 --> 00:50:31,918 Ben. 1031 00:50:31,918 --> 00:50:33,710 AUDIENCE: So I think the time that I really 1032 00:50:33,710 --> 00:50:35,255 understood hash functions was when 1033 00:50:35,255 --> 00:50:37,820 I saw someone do a live demo. 1034 00:50:37,820 --> 00:50:39,950 It's a website called [INAUDIBLE] Brain Wallet. 1035 00:50:39,950 --> 00:50:43,070 But you type in text, and in real time, 1036 00:50:43,070 --> 00:50:44,780 it converts it into a hash. 1037 00:50:44,780 --> 00:50:45,280 So 1038 00:50:45,280 --> 00:50:47,150 GARY GENSLER: Say the website again, Ben. 1039 00:50:47,150 --> 00:50:49,640 AUDIENCE: It's brainwallet.io. 1040 00:50:49,640 --> 00:50:52,620 GARY GENSLER: Brainwallet.io-- a recommendation. 1041 00:50:52,620 --> 00:50:55,840 AUDIENCE: You can also just Google Brain Wallet blockchain. 1042 00:50:55,840 --> 00:50:58,200 And you can type in text, and you see in real time 1043 00:50:58,200 --> 00:50:59,890 it converts it into a Bitcoin address. 1044 00:50:59,890 --> 00:51:02,810 If you change one letter, or make one in uppercase or one 1045 00:51:02,810 --> 00:51:04,550 in lower case, it changes in real time. 1046 00:51:04,550 --> 00:51:08,280 And it's just, you put any text in, you put a password in, 1047 00:51:08,280 --> 00:51:10,248 and it turns it into a blockchain address. 1048 00:51:10,248 --> 00:51:11,290 And that's all a hash is. 1049 00:51:11,290 --> 00:51:14,518 It translates text into a hash. 1050 00:51:14,518 --> 00:51:15,810 GARY GENSLER: Or a whole movie. 1051 00:51:19,510 --> 00:51:21,490 We talked about the network consensus, 1052 00:51:21,490 --> 00:51:28,000 how to actually agree on the state of information 1053 00:51:28,000 --> 00:51:32,020 with no centralized authority. 1054 00:51:32,020 --> 00:51:35,200 You don't have a central bank or a commercial bank 1055 00:51:35,200 --> 00:51:42,490 or a Facebook or a parental unit, if you wish. 1056 00:51:42,490 --> 00:51:44,680 It's all of us out there on the playground 1057 00:51:44,680 --> 00:51:46,550 figuring it out together somehow. 1058 00:51:50,350 --> 00:51:52,670 AUDIENCE: Sorry, I have a quesiton here. 1059 00:51:52,670 --> 00:51:55,085 For the proof of work, my question 1060 00:51:55,085 --> 00:51:57,320 is, for example, I make a transaction, 1061 00:51:57,320 --> 00:52:00,440 does it mean I have to wait for 10 minutes for the transaction 1062 00:52:00,440 --> 00:52:02,780 to be completed? 1063 00:52:02,780 --> 00:52:07,340 Or for example, when we do Venmo, it's like instantaneous. 1064 00:52:07,340 --> 00:52:10,030 I can immediately get the results. 1065 00:52:10,030 --> 00:52:11,780 GARY GENSLER: Anton's question is, does it 1066 00:52:11,780 --> 00:52:13,280 mean I have to wait 10 minutes? 1067 00:52:13,280 --> 00:52:18,050 Venmo and so many other payment practices can go more quickly. 1068 00:52:18,050 --> 00:52:22,250 The answer to you is, yes. 1069 00:52:22,250 --> 00:52:26,350 And that is one of the commercial challenges 1070 00:52:26,350 --> 00:52:30,770 to blockchain as we know it in Bitcoin. 1071 00:52:30,770 --> 00:52:36,980 There are certain approaches to that, layering in technology 1072 00:52:36,980 --> 00:52:39,552 on top of, they call it layer 2 or lightning network. 1073 00:52:39,552 --> 00:52:41,510 We're not going to dive into lightning network. 1074 00:52:41,510 --> 00:52:45,160 But everybody should hold onto Anton's question. 1075 00:52:45,160 --> 00:52:47,930 It's the right question, that if all of you 1076 00:52:47,930 --> 00:52:52,130 bring your critical reasoning to this class about markets 1077 00:52:52,130 --> 00:52:56,300 and about commercial realities, a little bit about the law 1078 00:52:56,300 --> 00:52:57,905 and a little bit about technology. 1079 00:52:57,905 --> 00:52:59,530 Because that's what we're trying to do. 1080 00:52:59,530 --> 00:53:01,940 It's like, oh, well does this really matter? 1081 00:53:01,940 --> 00:53:02,510 Will it work? 1082 00:53:02,510 --> 00:53:03,080 Hugo. 1083 00:53:03,080 --> 00:53:05,458 AUDIENCE: So just a point of contention here-- 1084 00:53:05,458 --> 00:53:06,500 GARY GENSLER: Contention? 1085 00:53:06,500 --> 00:53:07,110 AUDIENCE: A little bit. 1086 00:53:07,110 --> 00:53:08,277 GARY GENSLER: Oh, very good. 1087 00:53:08,277 --> 00:53:08,840 I like that. 1088 00:53:08,840 --> 00:53:11,750 AUDIENCE: So if you know the person 1089 00:53:11,750 --> 00:53:14,390 that you're transacting with, you can accept the transaction 1090 00:53:14,390 --> 00:53:15,590 with 0 confirmations. 1091 00:53:15,590 --> 00:53:17,330 As long as it gets into the mem pool 1092 00:53:17,330 --> 00:53:19,910 and has a reasonable fee attached to it, 1093 00:53:19,910 --> 00:53:22,260 then it will eventually get included on a block, 1094 00:53:22,260 --> 00:53:24,260 and that's probably good enough for most people. 1095 00:53:24,260 --> 00:53:27,950 GARY GENSLER: So what Hugo is saying is, 1096 00:53:27,950 --> 00:53:34,060 another approach is to take counterparty risk, 1097 00:53:34,060 --> 00:53:34,863 that Anton's-- 1098 00:53:34,863 --> 00:53:36,280 back to Anton's question-- does it 1099 00:53:36,280 --> 00:53:37,990 mean you have to wait 10 minutes? 1100 00:53:37,990 --> 00:53:40,030 The actual technical answer to that 1101 00:53:40,030 --> 00:53:45,880 is, no, you don't have to wait 10 minutes, unless you want 1102 00:53:45,880 --> 00:53:50,020 final settlement, if you want finality 1103 00:53:50,020 --> 00:53:56,140 with no counterparty risk, no commercial risk. 1104 00:53:56,140 --> 00:53:58,060 Hugo is saying, well, if I'm willing to take 1105 00:53:58,060 --> 00:54:01,660 some economic or commercial counterparty risk, which is, 1106 00:54:01,660 --> 00:54:05,470 in finance, you take it all the time, 1107 00:54:05,470 --> 00:54:08,380 then maybe I can do it in less than 10 minutes. 1108 00:54:08,380 --> 00:54:10,720 And in fact, even Starbucks, when 1109 00:54:10,720 --> 00:54:14,470 Starbucks accepts your credit card swipe, 1110 00:54:14,470 --> 00:54:17,170 they're taking a little bit of counterparty risk 1111 00:54:17,170 --> 00:54:21,010 from the payment processing company First Data. 1112 00:54:21,010 --> 00:54:24,010 I don't mean to say that it might not be also 1113 00:54:24,010 --> 00:54:25,450 from Visa and the banks. 1114 00:54:25,450 --> 00:54:28,210 But I'll just say the payment system 1115 00:54:28,210 --> 00:54:30,490 has some counterparty risk. 1116 00:54:30,490 --> 00:54:32,890 Because final settlement in our payment 1117 00:54:32,890 --> 00:54:36,420 system doesn't happen within seconds. 1118 00:54:36,420 --> 00:54:39,210 So the actual answer-- thank you, Hugo. 1119 00:54:39,210 --> 00:54:40,470 Clean me up again. 1120 00:54:40,470 --> 00:54:42,750 You should do this all the time. 1121 00:54:42,750 --> 00:54:44,520 Everybody should clean me up. 1122 00:54:44,520 --> 00:54:46,720 It's that final settlement can happen. 1123 00:54:46,720 --> 00:54:48,433 And so you have to find other solutions, 1124 00:54:48,433 --> 00:54:50,100 whether it's some commercial arrangement 1125 00:54:50,100 --> 00:54:53,228 with counterparty risk or other technical commercial 1126 00:54:53,228 --> 00:54:53,770 arrangements. 1127 00:54:53,770 --> 00:54:54,290 Shimon? 1128 00:54:54,290 --> 00:54:56,770 AUDIENCE: Well, I'll make a counterargument, 1129 00:54:56,770 --> 00:55:00,065 which is that, 10 minutes is not final settlement, right? 1130 00:55:00,065 --> 00:55:01,290 Because it could be forking. 1131 00:55:01,290 --> 00:55:04,630 So it's basically the probability 1132 00:55:04,630 --> 00:55:08,380 of the finality goes up with time, 1133 00:55:08,380 --> 00:55:12,120 which is associated with how many blocks are attached to 1134 00:55:12,120 --> 00:55:13,500 [INAUDIBLE]. 1135 00:55:13,500 --> 00:55:17,980 GARY GENSLER: So Shimon's point is that, even in 10 minutes, 1136 00:55:17,980 --> 00:55:23,000 your probability of finality is not complete. 1137 00:55:23,000 --> 00:55:27,470 Because the block might not be the block 1138 00:55:27,470 --> 00:55:29,840 that's included in the longest chain. 1139 00:55:29,840 --> 00:55:32,510 And many people have said, you maybe 1140 00:55:32,510 --> 00:55:34,620 should wait three blocks or six blocks. 1141 00:55:34,620 --> 00:55:35,495 I think the longest-- 1142 00:55:38,810 --> 00:55:40,370 I'm going to use the term loosely-- 1143 00:55:40,370 --> 00:55:46,110 orphaned chain has been five blocks long. 1144 00:55:46,110 --> 00:55:49,140 And in Bitcoin, what's the longest orphan? 1145 00:55:49,140 --> 00:55:50,640 AUDIENCE: It was an accidental fork. 1146 00:55:50,640 --> 00:55:53,240 It was 20 or more blocks. 1147 00:55:53,240 --> 00:55:56,127 Due to some crazy things that miners were doing, 1148 00:55:56,127 --> 00:55:57,710 they accidentally forked a blockchain. 1149 00:55:57,710 --> 00:55:59,720 AUDIENCE: OK, and what year was that? 1150 00:55:59,720 --> 00:56:02,020 AUDIENCE: So look up March 2013 fork. 1151 00:56:02,020 --> 00:56:04,815 And there's another fork after that, in 2015 maybe. 1152 00:56:04,815 --> 00:56:07,190 GARY GENSLER: Yeah, somebody's been down the rabbit hole. 1153 00:56:07,190 --> 00:56:07,690 [LAUGHTER] 1154 00:56:07,690 --> 00:56:12,830 All right, there's some still probabilistic risk. 1155 00:56:12,830 --> 00:56:15,155 Aviva, and then we'll move on. 1156 00:56:15,155 --> 00:56:17,450 AUDIENCE: What's a fork. 1157 00:56:17,450 --> 00:56:19,400 GARY GENSLER: So I don't have the chart up. 1158 00:56:19,400 --> 00:56:23,590 But if you remember, there was the slide 1159 00:56:23,590 --> 00:56:25,270 that showed the longest block. 1160 00:56:25,270 --> 00:56:26,920 It was in black. 1161 00:56:26,920 --> 00:56:29,740 And it had little purple blocks. 1162 00:56:29,740 --> 00:56:31,780 That's a fork. 1163 00:56:31,780 --> 00:56:39,030 There's some forks that end up being that both chains continue 1164 00:56:39,030 --> 00:56:40,140 for a long time. 1165 00:56:40,140 --> 00:56:41,550 They're called hard forks. 1166 00:56:41,550 --> 00:56:43,110 And there is something called-- 1167 00:56:43,110 --> 00:56:46,080 Bitcoin and Bitcoin Cash have come out of that. 1168 00:56:46,080 --> 00:56:52,963 Most, the way that Alin was using it, are discarded. 1169 00:56:52,963 --> 00:56:55,130 I'm sorry, behind you, remind me of your first name? 1170 00:56:55,130 --> 00:56:57,538 AUDIENCE: Erin. 1171 00:56:57,538 --> 00:56:58,080 I don't know. 1172 00:56:58,080 --> 00:56:59,550 Let me know if it's not the right time 1173 00:56:59,550 --> 00:57:00,425 to ask this question. 1174 00:57:00,425 --> 00:57:03,922 But I'm getting a bit confused between the most important 1175 00:57:03,922 --> 00:57:05,880 things we now need to know, and the differences 1176 00:57:05,880 --> 00:57:09,870 between the technology of mining Bitcoins versus just 1177 00:57:09,870 --> 00:57:14,220 transacting on the blockchain. 1178 00:57:18,622 --> 00:57:20,580 What are the major differences we need to know? 1179 00:57:20,580 --> 00:57:23,460 GARY GENSLER: Very good question, Erin. 1180 00:57:23,460 --> 00:57:27,300 Erin's question of, what do I need to know about mining, 1181 00:57:27,300 --> 00:57:29,010 what do I need to know about transaction, 1182 00:57:29,010 --> 00:57:30,510 is it one and the same? 1183 00:57:30,510 --> 00:57:31,680 I apologize. 1184 00:57:31,680 --> 00:57:33,940 They overlap, but they're not the same. 1185 00:57:33,940 --> 00:57:35,805 So think of a Venn diagram. 1186 00:57:35,805 --> 00:57:38,610 But it's a very good question. 1187 00:57:38,610 --> 00:57:46,710 The essence of mining is creating an incentive structure 1188 00:57:46,710 --> 00:57:50,520 where there is no central authority 1189 00:57:50,520 --> 00:57:56,850 to validate and put a new set of transactions or data-- 1190 00:57:56,850 --> 00:57:59,310 I'm going to say broadly, data-- 1191 00:57:59,310 --> 00:58:06,050 into the ledger, into the accepted state of what 1192 00:58:06,050 --> 00:58:08,480 reality is. 1193 00:58:08,480 --> 00:58:10,010 So mining helps with that. 1194 00:58:10,010 --> 00:58:12,030 That's that whole process. 1195 00:58:12,030 --> 00:58:15,770 In essence, who gets to decide the next block of data? 1196 00:58:15,770 --> 00:58:21,430 Transactions are included in the data, but it's not identical. 1197 00:58:21,430 --> 00:58:27,880 Mining is really critical, but it's not the only component. 1198 00:58:27,880 --> 00:58:30,233 In terms of transactions, then you 1199 00:58:30,233 --> 00:58:31,900 have to actually think of, well, there's 1200 00:58:31,900 --> 00:58:33,300 this other thing going on. 1201 00:58:33,300 --> 00:58:36,970 Well, there's been 340 million of these in the Bitcoin network 1202 00:58:36,970 --> 00:58:38,470 so far. 1203 00:58:38,470 --> 00:58:41,280 And has it been used already? 1204 00:58:41,280 --> 00:58:43,270 Has it been spent? 1205 00:58:43,270 --> 00:58:45,880 Does it have an appropriate digital signature? 1206 00:58:45,880 --> 00:58:49,570 If there's a time lock on it, is there a condition? 1207 00:58:49,570 --> 00:58:52,030 Might there even be this little bit of scripting code 1208 00:58:52,030 --> 00:58:53,830 that puts other conditions, like there 1209 00:58:53,830 --> 00:58:57,150 has to be multiple signatures? 1210 00:58:57,150 --> 00:58:58,275 Has it been double spent? 1211 00:59:01,810 --> 00:59:04,230 It was a very good question. 1212 00:59:04,230 --> 00:59:06,492 They overlap a lot. 1213 00:59:06,492 --> 00:59:07,950 One thing you can just remember is, 1214 00:59:07,950 --> 00:59:10,830 mining is about that there's been a half a million blocks. 1215 00:59:10,830 --> 00:59:13,800 Transactions, there's been 340 million. 1216 00:59:13,800 --> 00:59:16,370 So there must be something else going on 1217 00:59:16,370 --> 00:59:17,730 in all those transactions. 1218 00:59:17,730 --> 00:59:20,430 Does that help a bit? 1219 00:59:20,430 --> 00:59:21,865 Kelly? 1220 00:59:21,865 --> 00:59:23,490 AUDIENCE: Since we're talking about all 1221 00:59:23,490 --> 00:59:25,590 the technical features and how they overlap, 1222 00:59:25,590 --> 00:59:30,270 one of the questions was, what part of blockchain 1223 00:59:30,270 --> 00:59:32,927 is novel to Satoshi? 1224 00:59:37,310 --> 00:59:39,480 Is the novelty in bringing it all together, 1225 00:59:39,480 --> 00:59:41,190 or is it one specific thing? 1226 00:59:41,190 --> 00:59:43,770 Because the paper talked about the ledger 1227 00:59:43,770 --> 00:59:46,650 and creating the incentive, and then 1228 00:59:46,650 --> 00:59:49,350 sort of solving the whole Byzantine Generals Problem. 1229 00:59:49,350 --> 00:59:51,570 But I don't really understand. 1230 00:59:51,570 --> 00:59:54,810 Was there a specific thing that unlocked the-- 1231 00:59:54,810 --> 00:59:56,520 GARY GENSLER: So Kelly's question, 1232 00:59:56,520 --> 01:00:00,420 which is the heart of the study questions is, what makes 1233 01:00:00,420 --> 01:00:02,610 the whole Satoshi paper novel? 1234 01:00:02,610 --> 01:00:06,070 Is it just bringing it all together, 1235 01:00:06,070 --> 01:00:08,430 or is there something more? 1236 01:00:08,430 --> 01:00:09,370 I'll give you a hint. 1237 01:00:09,370 --> 01:00:10,630 I think there's one other-- 1238 01:00:10,630 --> 01:00:15,540 I think that the genius can be just in bringing together 1239 01:00:15,540 --> 01:00:16,880 things. 1240 01:00:16,880 --> 01:00:19,770 That in itself can be sheer genius. 1241 01:00:19,770 --> 01:00:21,840 AUDIENCE: There's also the part that, creating 1242 01:00:21,840 --> 01:00:23,430 the value in the currency. 1243 01:00:23,430 --> 01:00:24,770 But I don't know if that's-- 1244 01:00:24,770 --> 01:00:27,870 GARY GENSLER: Right, so creating an incentive structure 1245 01:00:27,870 --> 01:00:30,500 within there. 1246 01:00:30,500 --> 01:00:32,420 Others? 1247 01:00:32,420 --> 01:00:32,920 Derek. 1248 01:00:32,920 --> 01:00:33,930 AUDIENCE: Yeah, I have a question. 1249 01:00:33,930 --> 01:00:35,790 GARY GENSLER: Question, or answer to Kelly's question? 1250 01:00:35,790 --> 01:00:36,790 AUDIENCE: No, I had a question. 1251 01:00:36,790 --> 01:00:37,890 GARY GENSLER: All right, I'm going to hold your question, 1252 01:00:37,890 --> 01:00:38,932 Derek, just for a second. 1253 01:00:38,932 --> 01:00:41,040 Who's going to help me with-- because it's 1254 01:00:41,040 --> 01:00:44,492 central to the study questions. 1255 01:00:44,492 --> 01:00:46,975 AUDIENCE: Just another thing to add-- 1256 01:00:46,975 --> 01:00:50,950 the proof of work that Satoshi did on the consensus 1257 01:00:50,950 --> 01:00:52,748 to [INAUDIBLE] it's also novel. 1258 01:00:52,748 --> 01:00:53,540 GARY GENSLER: Yeah. 1259 01:00:53,540 --> 01:00:56,790 But novel, though, didn't Adam Back 1260 01:00:56,790 --> 01:00:59,760 already do some of it in hashcash. 1261 01:00:59,760 --> 01:01:02,700 So its application was novel. 1262 01:01:02,700 --> 01:01:05,250 I come out where I think the genius is bringing it 1263 01:01:05,250 --> 01:01:11,530 all together, and using Adam Back's proof of work 1264 01:01:11,530 --> 01:01:15,370 in a way that really addressed double spend. 1265 01:01:15,370 --> 01:01:17,890 Adam Back was not dealing with a double spend issue. 1266 01:01:17,890 --> 01:01:19,260 He had different challenges. 1267 01:01:19,260 --> 01:01:21,340 It was about email spam. 1268 01:01:21,340 --> 01:01:23,152 Frankly, it didn't even work with email. 1269 01:01:23,152 --> 01:01:25,360 AUDIENCE: [INAUDIBLE] we weren't counting [INAUDIBLE] 1270 01:01:25,360 --> 01:01:27,475 work as a completely novel thing, 1271 01:01:27,475 --> 01:01:30,550 his article discussed how it was already published. 1272 01:01:30,550 --> 01:01:31,960 So I guess it's the application. 1273 01:01:31,960 --> 01:01:34,000 GARY GENSLER: It's the application specifically 1274 01:01:34,000 --> 01:01:36,865 to the double spend. 1275 01:01:36,865 --> 01:01:41,470 Now, Patrick Murck, would you answer it then? 1276 01:01:41,470 --> 01:01:43,840 This is somebody who ran the Bitcoin Foundation, 1277 01:01:43,840 --> 01:01:46,832 but he's a lawyer, you know? 1278 01:01:46,832 --> 01:01:48,790 PATRICK MURCK: I think that's absolutely right. 1279 01:01:48,790 --> 01:01:51,870 So usually-- and I think they address it in that particular 1280 01:01:51,870 --> 01:01:52,480 paper-- 1281 01:01:52,480 --> 01:01:54,050 when somebody says, what was a thing 1282 01:01:54,050 --> 01:01:56,300 that was different about Bitcoin from everything else, 1283 01:01:56,300 --> 01:01:59,020 the answer is Nakamoto Consensus, right? 1284 01:01:59,020 --> 01:02:02,410 And Nakamoto Consensus, being the incentive structure 1285 01:02:02,410 --> 01:02:05,050 pulling everything together and aligning everybody 1286 01:02:05,050 --> 01:02:07,720 to actually create trusted signing parties 1287 01:02:07,720 --> 01:02:10,900 for this database, without having to actually trust 1288 01:02:10,900 --> 01:02:12,990 or identify those parties. 1289 01:02:12,990 --> 01:02:16,670 And that's something that really hadn't existed before. 1290 01:02:16,670 --> 01:02:18,520 And so that's really novel. 1291 01:02:18,520 --> 01:02:21,280 That's sort of the breakthrough that I 1292 01:02:21,280 --> 01:02:25,570 think you can attribute to that particular white paper. 1293 01:02:25,570 --> 01:02:27,730 I also use this as a way to give a clean definition 1294 01:02:27,730 --> 01:02:30,550 of blockchain, which is a badly abused term, as I'm 1295 01:02:30,550 --> 01:02:33,955 sure you'll discover through the rest of this class. 1296 01:02:33,955 --> 01:02:36,100 It's saying, to me, a blockchain is 1297 01:02:36,100 --> 01:02:40,248 something that is born from Nakamoto Consensus. 1298 01:02:40,248 --> 01:02:42,040 Blockchains, as they discuss in that paper, 1299 01:02:42,040 --> 01:02:44,080 have existed for decades. 1300 01:02:44,080 --> 01:02:46,590 That's not even a novel data structure. 1301 01:02:46,590 --> 01:02:49,270 But using that to form Nakamoto Con-- that's the thing. 1302 01:02:49,270 --> 01:02:52,520 It always sort of comes back to that. 1303 01:02:52,520 --> 01:02:55,348 Anyways, maybe a longer elaboration than you wanted. 1304 01:02:55,348 --> 01:02:56,890 GARY GENSLER: Good, there we have it. 1305 01:02:56,890 --> 01:03:00,145 You now know that you're in a class that has guest speakers. 1306 01:03:02,657 --> 01:03:04,990 I do want to say on the guest speaker point-- and Derek, 1307 01:03:04,990 --> 01:03:07,830 I cognizant that you have a question, 1308 01:03:07,830 --> 01:03:12,790 but we have 12 minutes or so. 1309 01:03:12,790 --> 01:03:16,030 Next Tuesday, Larry Lessig, who sometimes 1310 01:03:16,030 --> 01:03:20,530 floats into this class, has agreed to guest lecture 1311 01:03:20,530 --> 01:03:21,530 with me. 1312 01:03:21,530 --> 01:03:23,440 And so we're going to co-do smart contracts. 1313 01:03:23,440 --> 01:03:26,080 And let me just say a little bit about Larry. 1314 01:03:26,080 --> 01:03:27,160 You sort of see him here. 1315 01:03:27,160 --> 01:03:29,270 He bicycles over from Harvard. 1316 01:03:29,270 --> 01:03:31,150 He's a constitutional scholar that's 1317 01:03:31,150 --> 01:03:33,080 a remarkable constitutional scholar. 1318 01:03:33,080 --> 01:03:36,250 And even though he came to be a full professor at Harvard, 1319 01:03:36,250 --> 01:03:38,110 they stole him away from Stanford actually, 1320 01:03:38,110 --> 01:03:41,420 where he was a full professor, too, readily. 1321 01:03:41,420 --> 01:03:44,530 But he's extensively written. 1322 01:03:44,530 --> 01:03:48,110 Anybody in here a West Wing fan, the television series 1323 01:03:48,110 --> 01:03:50,660 West Wing? 1324 01:03:50,660 --> 01:03:52,350 He is the only-- 1325 01:03:52,350 --> 01:03:54,822 do you know that Larry Lessig was in West Wing? 1326 01:03:54,822 --> 01:03:55,970 AUDIENCE: No. 1327 01:03:55,970 --> 01:03:57,470 GARY GENSLER: Now you do. 1328 01:03:57,470 --> 01:03:58,340 Well, Larry was. 1329 01:03:58,340 --> 01:04:02,360 And Christopher Lloyd played Larry Lessig in one episode. 1330 01:04:02,360 --> 01:04:05,400 But you'll go back and you'll find the episode. 1331 01:04:05,400 --> 01:04:09,560 And Larry has a funny story about watching 1332 01:04:09,560 --> 01:04:10,940 the filming of it. 1333 01:04:10,940 --> 01:04:15,320 But Larry is a constitutional scholar over at Harvard. 1334 01:04:15,320 --> 01:04:20,030 He clerked for Justice Scalia on the Supreme Court. 1335 01:04:20,030 --> 01:04:25,580 He clerked for Posner over in Chicago. 1336 01:04:25,580 --> 01:04:27,140 He knows a lot about contracts. 1337 01:04:27,140 --> 01:04:29,810 And so I asked him if he'd help teach smart contracts 1338 01:04:29,810 --> 01:04:30,600 next Tuesday. 1339 01:04:30,600 --> 01:04:32,933 So Larry's going to-- we're going to Mutt and Jeff it up 1340 01:04:32,933 --> 01:04:35,240 here next Tuesday. 1341 01:04:35,240 --> 01:04:36,140 Watch out, Patrick. 1342 01:04:36,140 --> 01:04:38,237 You might be up here one day. 1343 01:04:38,237 --> 01:04:40,070 This is not a class that we're going to have 1344 01:04:40,070 --> 01:04:42,110 a lot of guest lecturers. 1345 01:04:42,110 --> 01:04:44,340 Later in this semester, I hope, we're 1346 01:04:44,340 --> 01:04:46,310 still in confirming Jeff Sprecher, who's 1347 01:04:46,310 --> 01:04:49,825 the chief executive officer of Intercontinental Exchange, 1348 01:04:49,825 --> 01:04:51,200 runs the New York Stock Exchange. 1349 01:04:51,200 --> 01:04:56,480 Jeff is probably joining us on, I think it's November 15. 1350 01:04:56,480 --> 01:05:01,700 But I really want to stay to the content and so forth. 1351 01:05:01,700 --> 01:05:03,418 Derek, what's your question? 1352 01:05:03,418 --> 01:05:05,210 AUDIENCE: I can follow up with you on that. 1353 01:05:05,210 --> 01:05:06,377 GARY GENSLER: OK, follow up. 1354 01:05:06,377 --> 01:05:08,790 All right, and then we just did transactions today. 1355 01:05:08,790 --> 01:05:12,110 Remember, the hash function is The New York Times. 1356 01:05:12,110 --> 01:05:16,400 We're in a little better place with Lauren's table now. 1357 01:05:16,400 --> 01:05:19,730 All right, good, good. 1358 01:05:19,730 --> 01:05:21,765 My goal is not to embarrass anybody 1359 01:05:21,765 --> 01:05:22,890 when I ask these questions. 1360 01:05:22,890 --> 01:05:25,790 My goal is that we all come along on this journey, 1361 01:05:25,790 --> 01:05:27,890 that we somehow have some basis. 1362 01:05:27,890 --> 01:05:29,990 Because it does relate to understanding 1363 01:05:29,990 --> 01:05:33,800 the commercial reality and the economics. 1364 01:05:33,800 --> 01:05:38,020 We talked about the time stamping and the blocks, 1365 01:05:38,020 --> 01:05:41,940 the Merkle Trees, which is not a deep part of it, and of course, 1366 01:05:41,940 --> 01:05:43,280 digital signatures. 1367 01:05:43,280 --> 01:05:45,470 Part of the reason I replay this each time 1368 01:05:45,470 --> 01:05:47,090 is because, in politics, I, of course, 1369 01:05:47,090 --> 01:05:49,710 learned that repetition is a really important thing. 1370 01:05:49,710 --> 01:05:50,210 [LAUGHTER] 1371 01:05:50,210 --> 01:05:55,280 But I also think it's true in academic settings. 1372 01:05:55,280 --> 01:05:58,160 And then Bitcoin addresses, and that's just a cleanup, 1373 01:05:58,160 --> 01:06:00,890 that it's not an identical to a public. 1374 01:06:00,890 --> 01:06:03,800 And then the proof of work, and back to the questions. 1375 01:06:03,800 --> 01:06:06,770 Nakamoto Consensus is, yes, all of this 1376 01:06:06,770 --> 01:06:11,060 and an incentive structure, but it's this proof of work. 1377 01:06:11,060 --> 01:06:14,450 And to Erin's question, proof of work 1378 01:06:14,450 --> 01:06:16,550 is a little bit different than the transactions 1379 01:06:16,550 --> 01:06:20,160 that we talked about, but there's a lot of overlap. 1380 01:06:20,160 --> 01:06:22,230 And then it creates the native currency. 1381 01:06:22,230 --> 01:06:28,810 And I've corrected this slide to 2140, of course. 1382 01:06:28,810 --> 01:06:31,490 The network is really critical, too, 1383 01:06:31,490 --> 01:06:33,360 and that there's all these different actors 1384 01:06:33,360 --> 01:06:37,530 on a network, 10,000 nodes and this many light nodes 1385 01:06:37,530 --> 01:06:39,690 and the miners and the mining pool operators. 1386 01:06:39,690 --> 01:06:43,320 And they all have their separate economics. 1387 01:06:43,320 --> 01:06:45,900 And so if anybody wants to come and get office hours, 1388 01:06:45,900 --> 01:06:49,440 talk about those economics, please come on in. 1389 01:06:49,440 --> 01:06:51,870 If there's nothing you remember from the reading, 1390 01:06:51,870 --> 01:06:55,200 I've now read this paper probably six times 1391 01:06:55,200 --> 01:06:57,930 to kind of slowly get it through my head. 1392 01:06:57,930 --> 01:07:00,870 But every time I read the Clark paper, 1393 01:07:00,870 --> 01:07:03,930 I go, wow, that really helps me. 1394 01:07:03,930 --> 01:07:07,590 Because it's not like Satoshi Nakamoto just flipped 1395 01:07:07,590 --> 01:07:09,570 his fingers and there it-- 1396 01:07:09,570 --> 01:07:13,020 it was on the backs of a lot of cryptography, 1397 01:07:13,020 --> 01:07:15,850 a lot of technology earlier. 1398 01:07:15,850 --> 01:07:17,610 But this is the chart I turn back to, 1399 01:07:17,610 --> 01:07:19,140 and it's sometimes helps me. 1400 01:07:19,140 --> 01:07:22,590 Ah, there's time-stamping, there's digital cache, 1401 01:07:22,590 --> 01:07:26,280 there's proof of work, and how these things. 1402 01:07:26,280 --> 01:07:28,200 Maybe 10 years from now, they'll look back 1403 01:07:28,200 --> 01:07:30,600 and Nakamoto's stuff will just be built upon. 1404 01:07:30,600 --> 01:07:32,100 That's the central question. 1405 01:07:32,100 --> 01:07:34,290 That's what some of our colleagues 1406 01:07:34,290 --> 01:07:36,600 are doing over at the Digital Currency Initiative 1407 01:07:36,600 --> 01:07:38,610 or over at the Computer Sciences Lab. 1408 01:07:38,610 --> 01:07:41,190 They're saying, can they build upon this 1409 01:07:41,190 --> 01:07:44,283 and take it to another level? 1410 01:07:44,283 --> 01:07:46,450 Right now-- and you'll see throughout the semester-- 1411 01:07:46,450 --> 01:07:49,290 there's not a lot of full scale applications 1412 01:07:49,290 --> 01:07:52,310 of this technology. 1413 01:07:52,310 --> 01:07:54,380 But it might just be in a whole line of this. 1414 01:07:54,380 --> 01:07:56,422 Yes, and I can't remember your first name, Aviva? 1415 01:07:56,422 --> 01:07:57,820 AUDIENCE: I'm Aviva. 1416 01:07:57,820 --> 01:07:59,340 And that is the other Indian woman. 1417 01:07:59,340 --> 01:08:00,170 GARY GENSLER: That's the other Aviva. 1418 01:08:00,170 --> 01:08:01,837 AUDIENCE: That's the other Indian woman. 1419 01:08:01,837 --> 01:08:03,545 GARY GENSLER: Yeah, yeah, it's important. 1420 01:08:03,545 --> 01:08:05,783 Out of 100 people, we can have two Avivas, you know? 1421 01:08:05,783 --> 01:08:07,700 We could even have two Aviva's if there's two. 1422 01:08:07,700 --> 01:08:09,550 AUDIENCE: [INAUDIBLE] name's Priya. 1423 01:08:09,550 --> 01:08:10,466 AUDIENCE: I'm Priya. 1424 01:08:10,466 --> 01:08:12,950 GARY GENSLER: Pria-- oh, I'm sorry. 1425 01:08:12,950 --> 01:08:13,900 Thank you. 1426 01:08:13,900 --> 01:08:15,770 AUDIENCE: So I was just thinking about, 1427 01:08:15,770 --> 01:08:19,229 as one of the applications of the hash function-- 1428 01:08:19,229 --> 01:08:24,040 so does the hash function actually replace the data? 1429 01:08:24,040 --> 01:08:27,319 So I'm thinking now that everyone's saving data 1430 01:08:27,319 --> 01:08:30,560 in the cloud, so can you save a hash function instead 1431 01:08:30,560 --> 01:08:34,142 of your actual data, and then so it can be compressed? 1432 01:08:34,142 --> 01:08:36,350 GARY GENSLER: Very good question-- my summary of it-- 1433 01:08:36,350 --> 01:08:38,899 though others' probably would be more expert. 1434 01:08:38,899 --> 01:08:40,470 My summary is, you get a choice. 1435 01:08:40,470 --> 01:08:41,970 You could do either. 1436 01:08:41,970 --> 01:08:46,840 So let's take it in blockchain rather than in the cloud. 1437 01:08:46,840 --> 01:08:51,790 You can choose to save in a blockchain just hashes, 1438 01:08:51,790 --> 01:08:55,569 and have the full picture somewhere else. 1439 01:08:55,569 --> 01:08:59,580 Let's say you were going to do a whole database of-- 1440 01:08:59,580 --> 01:09:00,623 AUDIENCE: A library. 1441 01:09:00,623 --> 01:09:01,540 GARY GENSLER: Of what? 1442 01:09:01,540 --> 01:09:02,915 AUDIENCE: I'd say like a library. 1443 01:09:02,915 --> 01:09:04,600 GARY GENSLER: A library-- 1444 01:09:04,600 --> 01:09:07,510 and so there's 100,000 books in the library. 1445 01:09:07,510 --> 01:09:09,760 You could hash all 100,000 books, 1446 01:09:09,760 --> 01:09:15,069 and then store the hashes in the blockchain and not the books 1447 01:09:15,069 --> 01:09:16,720 themselves. 1448 01:09:16,720 --> 01:09:19,420 And that would form a blockchain of the commitments 1449 01:09:19,420 --> 01:09:20,680 to those books. 1450 01:09:20,680 --> 01:09:25,149 Or you could actually, I guess, put the books themselves 1451 01:09:25,149 --> 01:09:26,229 into the blockchain. 1452 01:09:26,229 --> 01:09:29,359 Now, I saw some shaking heads in the middle. 1453 01:09:29,359 --> 01:09:33,010 Alin from the Digital Currency Initiative. 1454 01:09:33,010 --> 01:09:35,670 AUDIENCE: The way I understand this is, the answer is no. 1455 01:09:38,979 --> 01:09:40,720 You can store the hash. 1456 01:09:40,720 --> 01:09:42,556 That doesn't replace the data. 1457 01:09:42,556 --> 01:09:45,460 The storing of the hash allows you to prove 1458 01:09:45,460 --> 01:09:47,500 that you have the data. 1459 01:09:47,500 --> 01:09:49,460 But the fact that you store the hash 1460 01:09:49,460 --> 01:09:51,520 doesn't mean that you store the data. 1461 01:09:51,520 --> 01:09:57,190 GARY GENSLER: But Alin, there's a two-part question. 1462 01:09:57,190 --> 01:09:59,050 You shouldn't get rid of the book, 1463 01:09:59,050 --> 01:10:02,140 because hashes are one way. 1464 01:10:02,140 --> 01:10:06,370 You're not able to take the hash and recreate the look. 1465 01:10:06,370 --> 01:10:09,040 You can't take the hash and recreate The New York Times 1466 01:10:09,040 --> 01:10:11,070 crossword puzzle. 1467 01:10:11,070 --> 01:10:13,275 But you don't need to store them in the same place. 1468 01:10:15,840 --> 01:10:18,090 Thank you, because it's two parts the question. 1469 01:10:18,090 --> 01:10:20,850 You could store the hashes in the cloud 1470 01:10:20,850 --> 01:10:23,400 and store the books somewhere else. 1471 01:10:23,400 --> 01:10:25,455 But you still need to store the book, maybe. 1472 01:10:29,390 --> 01:10:32,450 So who is Satoshi Nakamoto? 1473 01:10:32,450 --> 01:10:35,480 We have just a handful of minutes, but can every table-- 1474 01:10:35,480 --> 01:10:37,550 each table's going to take four minutes. 1475 01:10:37,550 --> 01:10:40,550 And amongst yourselves, decide who is Satoshi Nakamoto. 1476 01:10:40,550 --> 01:10:41,910 So how are we doing? 1477 01:10:41,910 --> 01:10:45,108 Who's your answer to Satoshi Nakamoto? 1478 01:10:45,108 --> 01:10:47,150 AUDIENCE: We would say, probably multiple people, 1479 01:10:47,150 --> 01:10:48,500 led by Hal Finney. 1480 01:10:48,500 --> 01:10:51,740 GARY GENSLER: OK, so the first one is, multiple people, 1481 01:10:51,740 --> 01:10:53,480 probably led by Hal Finney. 1482 01:11:00,561 --> 01:11:01,190 AUDIENCE: NSA. 1483 01:11:01,190 --> 01:11:05,100 AUDIENCE: Something within the government or a government. 1484 01:11:05,100 --> 01:11:11,370 GARY GENSLER: So table number 2 is, government actor, US 1485 01:11:11,370 --> 01:11:12,796 or foreign. 1486 01:11:12,796 --> 01:11:15,080 AUDIENCE: I don't know if it matters, but maybe US. 1487 01:11:15,080 --> 01:11:17,630 GARY GENSLER: US, but it might not matter. 1488 01:11:17,630 --> 01:11:18,860 How are we doing over here? 1489 01:11:18,860 --> 01:11:20,138 AUDIENCE: Dorian Nakamoto. 1490 01:11:20,138 --> 01:11:21,430 GARY GENSLER: Dorian Nakamoto-- 1491 01:11:24,020 --> 01:11:27,860 so you're going with the Newsweek story. 1492 01:11:27,860 --> 01:11:30,620 Pria, sorry about the name thing before. 1493 01:11:30,620 --> 01:11:31,620 Your table, who-- 1494 01:11:31,620 --> 01:11:33,473 AUDIENCE: A bunch of crypto punks. 1495 01:11:33,473 --> 01:11:35,390 GARY GENSLER: What's that, a bunch of crypto-- 1496 01:11:35,390 --> 01:11:36,260 AUDIENCE: Punks. 1497 01:11:36,260 --> 01:11:36,545 GARY GENSLER: Punks. 1498 01:11:36,545 --> 01:11:39,140 AUDIENCE: Plus economists-- it's like a group of people. 1499 01:11:39,140 --> 01:11:42,380 GARY GENSLER: So a group of crypto punks and economists. 1500 01:11:42,380 --> 01:11:44,720 And how do you spell crypto punks, though? 1501 01:11:44,720 --> 01:11:46,220 AUDIENCE: Cipher punks. 1502 01:11:46,220 --> 01:11:49,565 GARY GENSLER: Cipher punks, cipher punks, actually. 1503 01:11:52,490 --> 01:11:53,770 Where are we here? 1504 01:11:53,770 --> 01:11:55,280 AUDIENCE: NSA. 1505 01:11:55,280 --> 01:11:57,320 GARY GENSLER: Oh, the NSA, all right. 1506 01:11:57,320 --> 01:11:59,697 AUDIENCE: People with incentive and the capability. 1507 01:11:59,697 --> 01:12:01,655 GARY GENSLER: Oh, so incentives and capability, 1508 01:12:01,655 --> 01:12:03,170 you think it's NSA. 1509 01:12:03,170 --> 01:12:03,930 AUDIENCE: MIT. 1510 01:12:03,930 --> 01:12:08,150 GARY GENSLER: MIT, oh! 1511 01:12:08,150 --> 01:12:10,520 AUDIENCE: We said, a guy named Gary Gensler. 1512 01:12:10,520 --> 01:12:12,255 [LAUGHTER] 1513 01:12:12,255 --> 01:12:13,880 GARY GENSLER: There is a word for that, 1514 01:12:13,880 --> 01:12:17,090 but I can't say that on tape. 1515 01:12:17,090 --> 01:12:21,940 Kelly, Anton, two Alins what do we have here, Jihee. 1516 01:12:21,940 --> 01:12:24,410 AUDIENCE: Let's go with Nick Szabo. 1517 01:12:24,410 --> 01:12:28,610 GARY GENSLER: Nick Szabo, who wrote the first paper 1518 01:12:28,610 --> 01:12:31,410 on smart contracts. 1519 01:12:31,410 --> 01:12:34,300 So Aviva is Nick Szabo. 1520 01:12:34,300 --> 01:12:35,928 My hash table-- 1521 01:12:35,928 --> 01:12:38,020 [LAUGHTER] 1522 01:12:40,487 --> 01:12:41,820 GARY GENSLER: Who do you go for? 1523 01:12:41,820 --> 01:12:44,390 AUDIENCE: We actually did have him, Szabo. 1524 01:12:44,390 --> 01:12:46,940 GARY GENSLER: All right, you can say 1525 01:12:46,940 --> 01:12:48,430 another table for Nick Szabo. 1526 01:12:48,430 --> 01:12:50,620 Put another vote next to him. 1527 01:12:50,620 --> 01:12:51,400 Here? 1528 01:12:51,400 --> 01:12:52,942 AUDIENCE: We're going with you, Gary. 1529 01:12:52,942 --> 01:12:54,800 GARY GENSLER: No, no, come on. 1530 01:12:54,800 --> 01:12:57,860 AUDIENCE: We think it's Craig Steven Wright. 1531 01:12:57,860 --> 01:13:00,710 GARY GENSLER: Craig Wright, the Australian. 1532 01:13:00,710 --> 01:13:02,270 Oh, my god. 1533 01:13:02,270 --> 01:13:06,350 Oh, Patrick Murck's table is going to go last. 1534 01:13:06,350 --> 01:13:08,600 Here? 1535 01:13:08,600 --> 01:13:11,340 AUDIENCE: Bill Belichick. 1536 01:13:11,340 --> 01:13:13,362 GARY GENSLER: Bill Belichick. 1537 01:13:13,362 --> 01:13:16,630 AUDIENCE: Alan Greenspan. 1538 01:13:16,630 --> 01:13:18,630 GARY GENSLER: Alan Greenspan. 1539 01:13:18,630 --> 01:13:20,270 I actually know Alan. 1540 01:13:20,270 --> 01:13:23,110 He's really talented, but I don't think 1541 01:13:23,110 --> 01:13:25,420 Andrea would let him do this. 1542 01:13:25,420 --> 01:13:28,030 Here, who do we have? 1543 01:13:32,935 --> 01:13:33,810 AUDIENCE: [INAUDIBLE] 1544 01:13:33,810 --> 01:13:34,560 GARY GENSLER: Who? 1545 01:13:34,560 --> 01:13:35,830 AUDIENCE: [INAUDIBLE], me. 1546 01:13:35,830 --> 01:13:37,510 GARY GENSLER: No, no, but who-- 1547 01:13:37,510 --> 01:13:39,210 you're saying you invented it? 1548 01:13:39,210 --> 01:13:40,450 No, no, this table. 1549 01:13:40,450 --> 01:13:42,450 AUDIENCE: This table is saying he invented it. 1550 01:13:42,450 --> 01:13:46,990 GARY GENSLER: Nick Szabo, so another for Nick Szabo. 1551 01:13:46,990 --> 01:13:51,160 So does anybody want to tell us why it's the NSA? 1552 01:13:54,130 --> 01:13:55,960 Oh, Patrick Murck, I'm sorry. 1553 01:13:55,960 --> 01:13:57,430 What's this table? 1554 01:13:57,430 --> 01:13:59,980 PATRICK MURCK: Well, I said, if I-- 1555 01:13:59,980 --> 01:14:00,850 so I don't know. 1556 01:14:00,850 --> 01:14:04,720 But if I did know, I would say that I don't know, 1557 01:14:04,720 --> 01:14:06,865 and I would create as much obfuscation as possible. 1558 01:14:06,865 --> 01:14:08,320 So I think I was the worst person 1559 01:14:08,320 --> 01:14:09,970 to have at a table for this. 1560 01:14:09,970 --> 01:14:12,220 And I did them nothing but a disservice in their hunt 1561 01:14:12,220 --> 01:14:13,620 for Satoshi. 1562 01:14:13,620 --> 01:14:15,185 But I'll let somebody else speak. 1563 01:14:15,185 --> 01:14:16,335 No, I don't know. 1564 01:14:16,335 --> 01:14:17,710 GARY GENSLER: But if he did know, 1565 01:14:17,710 --> 01:14:20,480 he would say he doesn't know. 1566 01:14:20,480 --> 01:14:24,000 Derek, who did your table go for? 1567 01:14:24,000 --> 01:14:26,670 AUDIENCE: We said Craig Wright. 1568 01:14:26,670 --> 01:14:28,780 PATRICK MURCK: You can see my influence. 1569 01:14:28,780 --> 01:14:31,570 GARY GENSLER: Oh, my gosh. 1570 01:14:31,570 --> 01:14:35,680 So let me ask this, because it's just for fun, one minute. 1571 01:14:35,680 --> 01:14:38,020 Somebody said that it was the NSA. 1572 01:14:38,020 --> 01:14:39,760 Do you want to say why? 1573 01:14:39,760 --> 01:14:41,410 AUDIENCE: Because, arguably, they 1574 01:14:41,410 --> 01:14:44,830 have the most advanced cryptography in the world. 1575 01:14:44,830 --> 01:14:49,973 And if anybody was doing this, to have a system where 1576 01:14:49,973 --> 01:14:52,390 all the dark money in the world was flowing around instead 1577 01:14:52,390 --> 01:14:56,220 of $100 bills, you would create this and create it in a way 1578 01:14:56,220 --> 01:14:58,497 where you could hack it backwards and figure it out. 1579 01:14:58,497 --> 01:14:59,705 And they have the capability. 1580 01:14:59,705 --> 01:15:00,455 GARY GENSLER: Wow. 1581 01:15:00,455 --> 01:15:00,955 Hugo. 1582 01:15:00,955 --> 01:15:03,497 AUDIENCE: More [INAUDIBLE] money is going through [INAUDIBLE] 1583 01:15:03,497 --> 01:15:05,230 right now than it is through Bitcoin. 1584 01:15:05,230 --> 01:15:08,820 GARY GENSLER: So Hugo would say that, if it was the NSA, 1585 01:15:08,820 --> 01:15:12,790 it didn't work out for them too well. 1586 01:15:12,790 --> 01:15:14,860 And those of you who said Craig Wright, 1587 01:15:14,860 --> 01:15:17,380 I heard some others in the room that said, no. 1588 01:15:17,380 --> 01:15:19,600 So who said Craig Wright? 1589 01:15:19,600 --> 01:15:23,190 Which two tables said Craig Wright? 1590 01:15:23,190 --> 01:15:25,460 Isabella and Ben, why did you pick Craig Wright? 1591 01:15:25,460 --> 01:15:26,906 AUDIENCE: So I read a bunch of it. 1592 01:15:26,906 --> 01:15:30,762 And basically, people analyzed the English used in the email, 1593 01:15:30,762 --> 01:15:32,792 and they think it traced back to Australia. 1594 01:15:32,792 --> 01:15:34,250 And then we [INAUDIBLE] from there. 1595 01:15:34,250 --> 01:15:37,622 GARY GENSLER: All right, so just the language analytics 1596 01:15:37,622 --> 01:15:38,330 for Craig Wright. 1597 01:15:38,330 --> 01:15:39,950 And those who said it can't possibly 1598 01:15:39,950 --> 01:15:43,920 be Craig Wright, who did that? 1599 01:15:43,920 --> 01:15:45,120 Alin? 1600 01:15:45,120 --> 01:15:47,150 AUDIENCE: Yeah, so he started a website. 1601 01:15:47,150 --> 01:15:49,160 He said, oh, here's cryptographic proof 1602 01:15:49,160 --> 01:15:50,440 that I'm Satoshi Nakamoto. 1603 01:15:50,440 --> 01:15:51,860 But he actually botched it. 1604 01:15:51,860 --> 01:15:53,570 Like, if you are Nakamoto, you can 1605 01:15:53,570 --> 01:15:56,540 prove you're Nakamoto by spending the first coin. 1606 01:15:56,540 --> 01:15:59,030 But he couldn't do that, so come on, man. 1607 01:15:59,030 --> 01:16:01,010 GARY GENSLER: So Alin says, he failed the test. 1608 01:16:01,010 --> 01:16:04,100 He didn't spend the first coins from 2009. 1609 01:16:04,100 --> 01:16:06,740 Three tables picked Nick Szabo. 1610 01:16:06,740 --> 01:16:11,263 Why'd you pick Nick Szabo, just any one of the three tables? 1611 01:16:11,263 --> 01:16:13,680 Because we're going to talk about him in the next lecture. 1612 01:16:13,680 --> 01:16:14,440 No? 1613 01:16:14,440 --> 01:16:15,790 All right, look, this was a bit of fun. 1614 01:16:15,790 --> 01:16:17,373 I just thought it would be worthwhile. 1615 01:16:17,373 --> 01:16:19,420 Because the only person in the room 1616 01:16:19,420 --> 01:16:21,970 that really knows who Satoshi Nakamoto 1617 01:16:21,970 --> 01:16:23,380 is isn't going to tell us. 1618 01:16:23,380 --> 01:16:25,750 [LAUGHTER] But you're welcome back any time. 1619 01:16:25,750 --> 01:16:26,400 Thank you. 1620 01:16:26,400 --> 01:16:27,550 We'll see you next Tuesday. 1621 01:16:27,550 --> 01:16:32,230 Remember, Larry Lessig is here, so please do the readings. 1622 01:16:32,230 --> 01:16:35,160 Please, let's have a good time.