1
00:02:30,000 --> 00:02:40,000
I also belong to IOB Labs, which is the organization that supports my research, and
within IOB Labs, I'm part of their research and innovation team.

2
00:02:40,000 --> 00:02:52,000
This is an awesome team, 10 people that we work in state-of-the-art research on
Bitcoin, RSK, and things blockchain in general.

3
00:02:53,000 --> 00:03:02,000
So what is Rustok? Rustok was the first Bitcoin sidechain. It was launched on
January 2018.

4
00:03:02,000 --> 00:03:14,000
The native currency of RSK is, of course, Bitcoin, and currently there are about
113,000 accounts in RSK.

5
00:03:14,000 --> 00:03:29,000
It has fast log times, around 32 seconds, and it runs EVM contracts. So basically
you can take an Ethereum application and port it to RSK and run it with a Bitcoin as
the native currency.

6
00:03:29,000 --> 00:03:46,000
Currently there are 15,000 contracts deployed on RSK. It is merged mining with
Bitcoin with approximately currently 57% of Bitcoin hash rate merged mining RSK.

7
00:03:46,000 --> 00:04:01,000
Generally this goes from 50% to 80% of the Bitcoin hash rate. And it has a two-way
peg with Bitcoin that is HSM-enforced, and this is an SPB bridge.

8
00:04:01,000 --> 00:04:18,000
Currently there are 3,000 BTC locked in the RSK peg. And the blockchain is used by
thousands of people. There are about 70,000 monthly active users in RSK.

9
00:04:18,000 --> 00:04:30,000
So there are a lot of Rustokers using RSK for DeFi, also for remittances, for
savings in stablecoins, and all kinds of very interesting applications.

10
00:04:30,000 --> 00:04:41,000
So the first question that I'm interested in talking today is what is a sidechain?
So this is my definition of a sidechain.

11
00:04:41,000 --> 00:04:55,000
A Bitcoin sidechain is an independent blockchain that can securely transfer Bitcoins
internally and from to the Bitcoin network without an ability to support a money
token different from Bitcoin.

12
00:04:56,000 --> 00:05:08,000
This is my own definition of what a sidechain is. But it seems like this is
cheating, right? Because maybe I choose the definition of sidechain to match RSK.

13
00:05:08,000 --> 00:05:27,000
Well, this is not the case because I'm basing my definition on the original
sidechain definition from 2012 that was posted on Bitcoin Talk that defines a
sidechain not by its technology but by the future it should have.

14
00:05:28,000 --> 00:05:53,000
So basically this user said that a sidechain does not require separate currencies,
so there is no other native money, does not need full or significant adoption by the
Bitcoin network, so it's isolated, does not mirror the Bitcoin blockchain block by
block, so this is a separate blockchain, it's not an overlay protocol, and it uses
and is dependent on Bitcoin.

15
00:05:53,000 --> 00:06:00,000
So probably this could be that it is two-way bridge or it is merge mine with
Bitcoin.

16
00:06:01,000 --> 00:06:15,000
So if we take a look at my definition and we compare it with the original definition
of the sidechain, basically what I did is simply put all these properties into a
single sentence that you can easily remember.

17
00:06:15,000 --> 00:06:21,000
So this is a very straightforward definition for a sidechain.

18
00:06:22,000 --> 00:06:26,000
But you probably think, OK, but what about decentralization?

19
00:06:27,000 --> 00:06:29,000
The definition does not say anything about decentralization.

20
00:06:30,000 --> 00:06:45,000
Well, the bad news is that with the extremely limited tools that Bitcoin has today,
which is basically multi-sig, time locks, and taproot, we cannot achieve the desired
level of decentralization.

21
00:06:46,000 --> 00:06:51,000
But with this limitation, Rusto has reached a very high level of decentralization.

22
00:06:52,000 --> 00:07:01,000
And I will show you in this talk all the properties that RSK has to be able to
become really decentralized.

23
00:07:02,000 --> 00:07:12,000
So the properties are basically that it must be run, it should be able to run in a
laptop, you should be able to run a full node in a laptop.

24
00:07:13,000 --> 00:07:19,000
You should have like clients, so nobody needs to depend on a centralized
infrastructure.

25
00:07:20,000 --> 00:07:21,000
It is a merge mine.

26
00:07:22,000 --> 00:07:31,000
It has a PoPeg, which is a very, very interesting type of peg with Bitcoin that I
will talk a little bit more about.

27
00:07:32,000 --> 00:07:33,000
It should be self-sufficiency.

28
00:07:34,000 --> 00:07:40,000
So if it grows, then it should be able to pay for its own growth.

29
00:07:41,000 --> 00:07:48,000
It should not grow in an unbounded way where people are left out of the network
because they cannot run a full node.

30
00:07:49,000 --> 00:07:55,000
And, of course, it needs to have a community-driven procedure to create network
upgrades.

31
00:07:56,000 --> 00:08:02,000
So let's go and see each one of these items one by one and what happens in RSK.

32
00:08:04,000 --> 00:08:14,000
So the technical features that enable decentralization relate to that full nodes
should be able to be run on a laptop.

33
00:08:15,000 --> 00:08:25,000
So one of the properties of RSK is that it is using a very, very efficient data
structure to store the state.

34
00:08:26,000 --> 00:08:30,000
So basically each account in RSK consumes only 12 bytes.

35
00:08:30,000 --> 00:08:44,000
So we can put millions of accounts and store all this data structure in memory and
provide very, very efficient management of this data structure.

36
00:08:45,000 --> 00:08:56,000
Second, we developed a specific database engine, a key value database engine, to
store this data structure that is very efficient.

37
00:08:57,000 --> 00:09:03,000
Its performance is about 20 times faster than LevelDB, for example.

38
00:09:04,000 --> 00:09:06,000
And writing speed is twice as fast as LevelDB.

39
00:09:07,000 --> 00:09:12,000
So this is the data structure that we plan to use in the future for storing the
state when the state grows.

40
00:09:13,000 --> 00:09:22,000
And we have a proposal to do compact block propagation that is already there in the
proposal repository.

41
00:09:23,000 --> 00:09:26,000
And, of course, there's still a lot of room to optimize blockchain synchronization.

42
00:09:27,000 --> 00:09:31,000
We think we can do better, but this is very promising.

43
00:09:32,000 --> 00:09:38,000
So we will always be able to run a full node in a standard laptop.

44
00:09:39,000 --> 00:09:44,000
So the other interesting aspect of RSK is how it relates to merge mining.

45
00:09:45,000 --> 00:09:50,000
So currently between 50 and 60 percent of the Bitcoin hash rate is mining RSK.

46
00:09:51,000 --> 00:09:52,000
So it is very secure.

47
00:09:53,000 --> 00:10:07,000
But, of course, if you know all the details about merge mining, the security of
merge mining is only assured if there is a high incentive alignment between
Bitcoiners and the sidechain community.

48
00:10:08,000 --> 00:10:10,000
And this is true for RSK.

49
00:10:11,000 --> 00:10:15,000
For instance, we have support from Bitcoin maxi pools such as RAINS.

50
00:10:16,000 --> 00:10:21,000
And we have also open source plugins to connect to different pools.

51
00:10:22,000 --> 00:10:29,000
One of the properties of RSK is that it is protected from fee snipping by using a
smooth revenue.

52
00:10:30,000 --> 00:10:37,000
So basically each block does not pay the full amount to the miner of that block, but
it pays a percentage of the fees collected.

53
00:10:38,000 --> 00:10:43,000
And the rest is postponed and is paid for the following miners.

54
00:10:44,000 --> 00:10:50,000
And one system that we put in place in RSK is called the universal merge mining
subsystem.

55
00:10:51,000 --> 00:10:56,000
And the idea of this system is to provide Blind Merged Mining for other sidechains.

56
00:10:57,000 --> 00:11:00,000
Currently there are no other sidechains, so this system is not being used.

57
00:11:00,000 --> 00:11:07,000
But we thought, how can we use RSK to facilitate the creation of new sidechains?

58
00:11:08,000 --> 00:11:10,000
So we put this system in RSK.

59
00:11:11,000 --> 00:11:20,000
Another interesting thing that needs to be very good decentralized is the POPEG, or
in case of other sidechains, the federation.

60
00:11:21,000 --> 00:11:27,000
So the federation or the POPEG in RSK is based on a 7 out of 13 multisig.

61
00:11:27,000 --> 00:11:39,000
But the main difference between RSK and other kind of federations is that the
private keys are stored in portable HSM devices that are based on ledger wallet
technology.

62
00:11:40,000 --> 00:11:46,000
So the POHSM validates the RSK chain in SPB node.

63
00:11:47,000 --> 00:11:56,000
So essentially each one of these devices is validating certain rules of the
consensus of RSK.

64
00:11:57,000 --> 00:12:06,000
So the devices themselves are able to control the release of BTC, the pegout
process.

65
00:12:07,000 --> 00:12:20,000
So each pegout command requires 4000 RSK confirmation blocks, which is about, in
comparing cumulative difficulty, it's about 100 Bitcoin blocks of hash rate, right?

66
00:12:21,000 --> 00:12:30,000
So if you want to cheat one device, even if you steal one of these devices, you will
not be able to command any pegout on the device.

67
00:12:30,000 --> 00:12:41,000
So this is very important because if you think that, you know, in the future states
around the world can try to censor the pegins or censor the pegouts,

68
00:12:42,000 --> 00:12:47,000
it's very important that the pegnatory, the one, the functionaries holding one of
these devices,

69
00:12:48,000 --> 00:12:52,000
it's very important that they can be and they can move to any part of the world.

70
00:12:52,000 --> 00:13:03,000
And with a single laptop and with the HSM devices, they can provide the services for
approving or for signing the pegouts.

71
00:13:04,000 --> 00:13:12,000
And finally, the POPEC nodes can be securely run on laptops anywhere on earth.

72
00:13:13,000 --> 00:13:17,000
They could in the future even connect to RSK over Tor.

73
00:13:17,000 --> 00:13:26,000
And it is very, very important that POPEC nodes remain secure, even when you run
them on laptops,

74
00:13:27,000 --> 00:13:37,000
because one of the most important centralization forces in any blockchain that is
not decentralized enough is that to run one validator,

75
00:13:37,000 --> 00:13:45,000
you need so much resources that people end up running these validators on the cloud
or on data centers,

76
00:13:46,000 --> 00:13:49,000
so they are prone to censorship by governments.

77
00:13:50,000 --> 00:13:57,000
The POHSMs are based on the ST33 tamper-resistant secure element,

78
00:13:57,000 --> 00:14:04,000
but if you have some doubts about these secure elements that are not open source,

79
00:14:05,000 --> 00:14:11,000
then we are ready to port the firmware to the TropicSquare open source secure
element whenever it is ready.

80
00:14:12,000 --> 00:14:20,000
So probably it will be on a second layer multisig so that both multisig need to
approve a pegout,

81
00:14:20,000 --> 00:14:30,000
one multisig using the POHSM devices and the new multisig using probably something
like the TropicSquare open source secure element.

82
00:14:31,000 --> 00:14:36,000
So we are trying to make this as open source as possible.

83
00:14:37,000 --> 00:14:43,000
The firmware of the POHSM will be soon open source, probably next week, I think.

84
00:14:44,000 --> 00:14:55,000
And we have a very interesting community proposal to move to threshold signatures to
be able to add more pegnatories in the future while keeping the pegout cost low.

85
00:14:56,000 --> 00:15:10,000
And we added in two Firefox ago, we added a 3.0 to 4.0 backup multisig that is time
locked for one year.

86
00:15:10,000 --> 00:15:16,000
So because these hardware devices have full control of the private keys, there is no
way to tamper with them,

87
00:15:17,000 --> 00:15:24,000
we needed to provide some safeguard in case the devices have some malfunction.

88
00:15:25,000 --> 00:15:29,000
So essentially, if for one year the devices are unable to process pegout,

89
00:15:30,000 --> 00:15:38,000
then there is a backup federation that will take care of these funds and
redistribute them or put them back in the peg.

90
00:15:39,000 --> 00:15:46,000
So one of the important things about RSK is that we plan to be self-sufficient in
the future.

91
00:15:47,000 --> 00:15:52,000
So we plan to use payment channels to pay for infrastructure for like client
services.

92
00:15:53,000 --> 00:16:00,000
We develop a technology which is called proof of unique blockchain storage to
incentivize decentralized storage of historical blockchain data.

93
00:16:01,000 --> 00:16:10,000
Pegnotories are already incentivized with a percent change on transaction fees, so
we think this is a sustained level in the network.

94
00:16:11,000 --> 00:16:15,000
And we implemented, and this is going to be activated in the following half work,

95
00:16:16,000 --> 00:16:22,000
we implemented a storage rent to keep the state under control so that it can fit in
RAM,

96
00:16:23,000 --> 00:16:28,000
so you will be always able to run your node on laptop, on a personal computer.

97
00:16:29,000 --> 00:16:35,000
Well, I'm getting run out of time, so this is about like clients.

98
00:16:36,000 --> 00:16:40,000
We also provided or are developing technologies to allow like clients.

99
00:16:41,000 --> 00:16:49,000
And finally, we have a very interesting process of upgrading the Rootstock platform

100
00:16:50,000 --> 00:16:55,000
because Rootstock development is much less politicized than Bitcoin development

101
00:16:55,000 --> 00:17:00,000
or than Ethereum development or than mostly any other chain development

102
00:17:01,000 --> 00:17:04,000
because basically we cannot change Bitcoin's money supply function.

103
00:17:05,000 --> 00:17:10,000
So anything you do on RSK is basically a technological change,

104
00:17:11,000 --> 00:17:14,000
something to optimize the way the resources are used,

105
00:17:15,000 --> 00:17:21,000
and there is no proposal you can create to alter something that,

106
00:17:21,000 --> 00:17:27,000
you know, alter decentralization or alter to change Bitcoin's supply function,

107
00:17:28,000 --> 00:17:30,000
which will be very, very controversial, right?

108
00:17:31,000 --> 00:17:38,000
So our Bitcoin community has a very strong stance to put decentralization before
scalability,

109
00:17:39,000 --> 00:17:43,000
so no change in the node will happen that affects decentralization.

110
00:17:44,000 --> 00:17:47,000
And we have more than 300 improvement proposals in our repository.

111
00:17:47,000 --> 00:17:52,000
Obviously, the core developers are trying to code them

112
00:17:53,000 --> 00:17:58,000
and to include them for community analysis as fast as they can,

113
00:17:59,000 --> 00:18:02,000
but this is a lot of proposals for improvement platform.

114
00:18:03,000 --> 00:18:07,000
And we perform approximately one upgrade per year.

115
00:18:08,000 --> 00:18:13,000
Okay, what about the decentralization of the two-way peg in the future?

116
00:18:14,000 --> 00:18:21,000
Well, I created the HashRay DeskCrow proposal, which is BIP-11,

117
00:18:22,000 --> 00:18:29,000
that uses a new opcode called opcode-axe to provide HashRay DeskCrow functionality
to Bitcoin.

118
00:18:30,000 --> 00:18:36,000
I hope that in the future I can present this proposal again when the time comes,

119
00:18:36,000 --> 00:18:43,000
when the community is ready to evaluate different alternatives to create sidechains.

120
00:18:44,000 --> 00:18:49,000
Also, we could switch to BIP300 if this is what the Bitcoin community prefers,

121
00:18:50,000 --> 00:18:56,000
or maybe there's a new technology, zero-knowledge peg or whatever you want to call
it,

122
00:18:57,000 --> 00:19:00,000
zero-knowledge rollups, and RSK could switch to this technology.

123
00:19:00,000 --> 00:19:05,000
We are working on very cool features that we'll be activating in the following hard
fork.

124
00:19:06,000 --> 00:19:08,000
One is, I mentioned before, Storage Rent.

125
00:19:09,000 --> 00:19:12,000
We have a prototype ready for parallel transaction processing

126
00:19:13,000 --> 00:19:15,000
to improve the transaction processing capabilities.

127
00:19:16,000 --> 00:19:19,000
And we have a protocol which is called Flyover protocol,

128
00:19:20,000 --> 00:19:21,000
which is a trustless peg-in and peg-out protocol.

129
00:19:22,000 --> 00:19:25,000
This is very interesting. It's already activated in the network.

130
00:19:26,000 --> 00:19:27,000
Okay, this is all.

131
00:19:27,000 --> 00:19:31,000
Thank you very much for this remote invitation.

132
00:19:32,000 --> 00:19:35,000
If you want to research more about RSK, you can go to the research forum

133
00:19:36,000 --> 00:19:38,000
or the developer site or their Discord

134
00:19:39,000 --> 00:19:42,000
or many other communication channels set up by the community.

135
00:19:43,000 --> 00:19:51,000
So, I invite you to go there, download your RSK full node, and have fun.

136
00:19:52,000 --> 00:19:53,000
Thank you.

137
00:19:58,000 --> 00:20:04,000
All right, we've got the next one coming up.

138
00:20:05,000 --> 00:20:07,000
This one is about unpopular opinions.

139
00:20:08,000 --> 00:20:09,000
I don't know if everyone's got them.

140
00:20:10,000 --> 00:20:14,000
Altcoins, LARPing, tail emissions.

141
00:20:15,000 --> 00:20:16,000
Let's have it out.

142
00:20:17,000 --> 00:20:18,000
We're going to have a panel.

143
00:20:19,000 --> 00:20:20,000
Marty's going to host it.

144
00:20:21,000 --> 00:20:22,000
We've got a few great panelists.

145
00:20:23,000 --> 00:20:25,000
We've got Peter Todd, Eric Voskil, Paul Sztorc.

146
00:20:28,000 --> 00:20:31,000
Giacomo, and I think that's it.

147
00:20:32,000 --> 00:20:33,000
Welcome them up.

148
00:20:34,000 --> 00:20:35,000
Marty's going to introduce everyone.

149
00:20:36,000 --> 00:20:37,000
Thank you so much.

150
00:20:39,000 --> 00:20:41,000
All right, we're audibling.

151
00:20:42,000 --> 00:20:45,000
We're still doing unpopular opinions, but we only have 30 minutes for this time
slot.

152
00:20:46,000 --> 00:20:47,000
Matt couldn't be here for our HR.

153
00:20:48,000 --> 00:20:50,000
He sends his love from across the pond.

154
00:20:51,000 --> 00:20:53,000
So, we're just going to extend this topic for the next hour

155
00:20:54,000 --> 00:20:56,000
because there's a lot of unpopular opinions.

156
00:20:59,000 --> 00:21:03,000
I mean, I don't want to shit on the dude who just presented,

157
00:21:04,000 --> 00:21:06,000
but Rootstock seems to be pretty unpopular with people.

158
00:21:07,000 --> 00:21:08,000
Why is that?

159
00:21:10,000 --> 00:21:12,000
I think a lot of people don't know what Rootstock does.

160
00:21:13,000 --> 00:21:15,000
He did a great job in that talk of listing a lot of things,

161
00:21:16,000 --> 00:21:18,000
but he never actually explained that it was supposed to be like an Ethereum.

162
00:21:19,000 --> 00:21:22,000
He never actually explained a basic feature of it.

163
00:21:23,000 --> 00:21:24,000
Well, you can explain why it's so unpopular,

164
00:21:25,000 --> 00:21:27,000
because it's for the same reason why your own ideas are so unpopular.

165
00:21:28,000 --> 00:21:31,000
This is starting, like, very good.

166
00:21:32,000 --> 00:21:35,000
Well, you could say we need the neutral zone.

167
00:21:36,000 --> 00:21:37,000
I was expecting a slower introduction.

168
00:21:38,000 --> 00:21:43,000
I actually have a real reason why RSK probably is not that popular,

169
00:21:44,000 --> 00:21:47,000
which is because Ethereum has its own token that you can buy,

170
00:21:48,000 --> 00:21:49,000
so it's not good enough.

171
00:21:49,000 --> 00:21:52,000
Even if RSK tied for features with Ethereum,

172
00:21:53,000 --> 00:21:54,000
people would still say,

173
00:21:55,000 --> 00:21:56,000
well, if I believe in these features,

174
00:21:57,000 --> 00:21:59,000
why would I go for the multi-sig,

175
00:22:00,000 --> 00:22:02,000
sort of crippled version on BTC

176
00:22:03,000 --> 00:22:05,000
when I could just invest in ETH, join the ETH community.

177
00:22:06,000 --> 00:22:09,000
And so even if it tied for features, it would lose to ETH,

178
00:22:10,000 --> 00:22:13,000
but it doesn't even tie because Sergio likes to invent

179
00:22:14,000 --> 00:22:15,000
a lot of weird things and plug them into RSK.

180
00:22:16,000 --> 00:22:17,000
Sergio likes to stay out of jail,

181
00:22:17,000 --> 00:22:19,000
and because of that, RSK will never moon.

182
00:22:20,000 --> 00:22:23,000
Peter, you just had a couple choice tweets about HSMs,

183
00:22:24,000 --> 00:22:25,000
riffing on that.

184
00:22:26,000 --> 00:22:27,000
Yeah, you pro-HSM now, right?

185
00:22:28,000 --> 00:22:29,000
Well, I mean,

186
00:22:30,000 --> 00:22:31,000
HSMs have their place,

187
00:22:32,000 --> 00:22:33,000
but to go on the same slide,

188
00:22:34,000 --> 00:22:36,000
say decentralization and HSM is crazy.

189
00:22:37,000 --> 00:22:39,000
Yeah, I think the HSMs are a little bit of a cop-out,

190
00:22:40,000 --> 00:22:41,000
so I agree with Peter on that one.

191
00:22:42,000 --> 00:22:43,000
Why is it crazy?

192
00:22:44,000 --> 00:22:45,000
Well, first of all, I've got to explain what is an HSM.

193
00:22:46,000 --> 00:22:48,000
And HSM stands for Hardware Security Module,

194
00:22:49,000 --> 00:22:51,000
and the idea there is you have this piece of hardware

195
00:22:52,000 --> 00:22:55,000
that's a security module that does something trusted

196
00:22:56,000 --> 00:23:00,000
in a way where even the person in physical possession of the HSM

197
00:23:01,000 --> 00:23:02,000
can't violate what it's doing.

198
00:23:03,000 --> 00:23:04,000
And, you know, of course, like Liquid uses these,

199
00:23:05,000 --> 00:23:07,000
like Blockstream Liquid, they use HSMs for the signers,

200
00:23:08,000 --> 00:23:11,000
but they use HSMs because people in possession of the signers

201
00:23:12,000 --> 00:23:14,000
really do not want the ability to go steal all that money.

202
00:23:15,000 --> 00:23:16,000
Well, at least most of them probably.

203
00:23:17,000 --> 00:23:18,000
Most of them don't.

204
00:23:19,000 --> 00:23:20,000
Because, you know, it creates risks.

205
00:23:21,000 --> 00:23:22,000
Like if you're in possession of someone else's money,

206
00:23:23,000 --> 00:23:24,000
you really don't want to be the guy who screws up

207
00:23:25,000 --> 00:23:26,000
and loses it all, so you build HSMs

208
00:23:27,000 --> 00:23:28,000
that are a little piece of hardware, they're secure,

209
00:23:29,000 --> 00:23:30,000
they have the anti-tamper effects, whatever.

210
00:23:31,000 --> 00:23:32,000
But the fact of the matter is,

211
00:23:33,000 --> 00:23:34,000
HSMs, you're trusting the manufacturer.

212
00:23:35,000 --> 00:23:37,000
And that's often better than just trusting the guy

213
00:23:38,000 --> 00:23:39,000
who's writing the computer.

214
00:23:40,000 --> 00:23:41,000
But that's still not decentralized.

215
00:23:41,000 --> 00:23:44,000
That, by the way, was also the way that

216
00:23:45,000 --> 00:23:47,000
Hal Finney was imagining reusable proof-of-work,

217
00:23:48,000 --> 00:23:49,000
because Hal Finney could not solve

218
00:23:50,000 --> 00:23:51,000
the double-spelling problem in a decentralized context,

219
00:23:52,000 --> 00:23:53,000
so he proposed to use trusted computing,

220
00:23:54,000 --> 00:23:55,000
which is basically HSMs.

221
00:23:56,000 --> 00:23:58,000
But it was like a second best to a decentralized solution,

222
00:23:59,000 --> 00:24:00,000
and then Satoshi Nakamoto invented.

223
00:24:01,000 --> 00:24:02,000
So as many other things,

224
00:24:03,000 --> 00:24:04,000
things proposed in Bitcoin right now,

225
00:24:05,000 --> 00:24:06,000
they are just marketing rediscoveries

226
00:24:07,000 --> 00:24:08,000
of things already proposed before,

227
00:24:08,000 --> 00:24:09,000
like the way that I proposed proof-of-stake,

228
00:24:10,000 --> 00:24:11,000
basically, with the second b-money.

229
00:24:12,000 --> 00:24:13,000
It was broken.

230
00:24:14,000 --> 00:24:15,000
People noticed it was broken.

231
00:24:16,000 --> 00:24:17,000
And so in 2014,

232
00:24:18,000 --> 00:24:19,000
shit-coiners rediscovered it

233
00:24:20,000 --> 00:24:21,000
and re-marketed it,

234
00:24:22,000 --> 00:24:23,000
just like everything else,

235
00:24:24,000 --> 00:24:25,000
like DAG chains or whatever.

236
00:24:26,000 --> 00:24:27,000
I mean, proof-of-stake in particular is hilarious,

237
00:24:28,000 --> 00:24:29,000
because in the early days

238
00:24:30,000 --> 00:24:31,000
of people trying to invent Bitcoin,

239
00:24:32,000 --> 00:24:33,000
it was sort of an obvious idea.

240
00:24:34,000 --> 00:24:35,000
Of course, you could have people with coins

241
00:24:36,000 --> 00:24:37,000
who sign it, but obviously that wouldn't work.

242
00:24:38,000 --> 00:24:39,000
How would you ever get consensus

243
00:24:40,000 --> 00:24:41,000
when you just go rewrite history at will?

244
00:24:42,000 --> 00:24:43,000
Well, it turns out if you go publish the consensus

245
00:24:44,000 --> 00:24:45,000
on Twitter or something, I guess that's enough.

246
00:24:46,000 --> 00:24:47,000
Yeah. Eric, we were talking about proof-of-stake,

247
00:24:48,000 --> 00:24:49,000
I believe it was yesterday or the day before,

248
00:24:50,000 --> 00:24:51,000
I think, the way you describe it.

249
00:24:52,000 --> 00:24:53,000
How would you say he who owns the history

250
00:24:54,000 --> 00:24:55,000
owns the future?

251
00:24:56,000 --> 00:24:57,000
Is that how you...?

252
00:24:58,000 --> 00:24:59,000
Yeah.

253
00:25:00,000 --> 00:25:01,000
I've never heard it framed that way.

254
00:25:03,576 --> 00:25:06,076
I don't think it's an uncommon viewpoint.

255
00:25:07,576 --> 00:25:08,676
It's certainly not an opinion.

256
00:25:09,276 --> 00:25:11,776
I don't know if it qualifies for this panel. This is factual.

257
00:25:12,976 --> 00:25:16,976
Sorry, Eric is offended because his are never opinions, just facts,

258
00:25:16,976 --> 00:25:18,576
so you put him in a couple of opinions.

259
00:25:18,576 --> 00:25:19,476
A man who knows everything.

260
00:25:19,476 --> 00:25:22,776
I think I've been on this panel every year, and I don't know why.

261
00:25:22,776 --> 00:25:27,276
I understand Giacomo, but even Peter.

262
00:25:28,176 --> 00:25:29,876
Even Paul over here, I don't understand.

263
00:25:33,376 --> 00:25:36,376
I generalize as much as I can, just about everything I can,

264
00:25:36,376 --> 00:25:39,376
and I look at stake versus work.

265
00:25:39,876 --> 00:25:42,576
People might imagine there's all these different kinds of proof,

266
00:25:42,576 --> 00:25:44,776
and it gets really confusing and everything, and it is, right?

267
00:25:44,976 --> 00:25:47,176
But when it comes right down to it, there's two kinds of proof,

268
00:25:47,476 --> 00:25:49,476
and I call it internal and external.

269
00:25:49,676 --> 00:25:52,276
There's what's internal to the chain, the history,

270
00:25:52,676 --> 00:25:53,576
and there's what's external.

271
00:25:53,576 --> 00:25:55,376
It has nothing to do with history.

272
00:25:55,376 --> 00:25:58,476
Work is just a name for what's external,

273
00:25:58,976 --> 00:26:00,576
and stake is a word for what's internal,

274
00:26:00,576 --> 00:26:03,276
and even if you combine them, one is going to dominate.

275
00:26:03,876 --> 00:26:04,976
What was it? Decred?

276
00:26:05,076 --> 00:26:05,676
Decred.

277
00:26:06,076 --> 00:26:07,376
Yeah, so I did some analysis on Decred,

278
00:26:07,376 --> 00:26:08,776
and I was like, you know, stake dominates,

279
00:26:08,776 --> 00:26:09,776
so it's proof of stake coin.

280
00:26:09,876 --> 00:26:11,076
It doesn't matter how much work you do.

281
00:26:11,576 --> 00:26:13,176
So if you want to simplify things,

282
00:26:13,176 --> 00:26:15,376
that's what this world divides into,

283
00:26:15,376 --> 00:26:19,076
and stake means that the history controls the future,

284
00:26:19,676 --> 00:26:23,576
and given that you don't have censorship resistance

285
00:26:23,576 --> 00:26:26,976
because once somebody controls the history,

286
00:26:26,976 --> 00:26:28,876
there's no way you can remove them.

287
00:26:29,476 --> 00:26:31,876
Yeah, they can buy old keys,

288
00:26:31,976 --> 00:26:37,276
and they can also just make millions of fake versions of history.

289
00:26:37,576 --> 00:26:38,976
So I don't know how many people here,

290
00:26:38,976 --> 00:26:41,276
hopefully we're in a room full of very smart people

291
00:26:41,276 --> 00:26:43,876
who are very skeptical of proof of stake

292
00:26:43,876 --> 00:26:47,876
and do not believe that it is a significant improvement

293
00:26:47,876 --> 00:26:49,376
over proof of work at all.

294
00:26:49,676 --> 00:26:52,076
I mean, I think, who here is,

295
00:26:52,276 --> 00:26:53,476
can we ask people to raise their hand

296
00:26:53,476 --> 00:26:54,676
if they think that proof of stake,

297
00:26:54,676 --> 00:26:55,976
they're open-minded to proof of stake?

298
00:26:55,976 --> 00:26:57,376
If you raise your hand, you come on stage

299
00:26:57,376 --> 00:26:58,976
because that's unpopular here.

300
00:26:59,376 --> 00:27:01,176
That's locally unpopular.

301
00:27:02,276 --> 00:27:03,976
Special room for you out back.

302
00:27:04,176 --> 00:27:06,976
The main selling point is environmentalism,

303
00:27:07,176 --> 00:27:10,376
and Paul's blog had a great article

304
00:27:10,376 --> 00:27:12,276
because you have to know that Paul's articles

305
00:27:12,276 --> 00:27:14,676
before it became anti-Maximilians were very good,

306
00:27:14,776 --> 00:27:16,276
and there's an article called

307
00:27:16,376 --> 00:27:18,476
There's Nothing Cheaper Than Proof of Work,

308
00:27:18,676 --> 00:27:20,776
which explains to you that you don't even get

309
00:27:20,776 --> 00:27:22,876
any environmentalist advantage.

310
00:27:22,976 --> 00:27:24,676
It's not just circular and broken.

311
00:27:24,776 --> 00:27:26,076
It's also mostly useless.

312
00:27:26,676 --> 00:27:30,576
Yeah, I think if you put an enormous amount of R&D

313
00:27:30,576 --> 00:27:32,376
into proof of stake, you could probably make it

314
00:27:32,376 --> 00:27:34,776
like very much basically equal to proof of work

315
00:27:34,776 --> 00:27:36,476
as far as in every respect,

316
00:27:36,476 --> 00:27:38,476
but I don't think that I'm very skeptical.

317
00:27:38,676 --> 00:27:41,076
For years, I thought we would do this joke

318
00:27:41,076 --> 00:27:42,976
about it being three months away,

319
00:27:42,976 --> 00:27:47,176
and we have that great Twitter thread about that,

320
00:27:47,576 --> 00:27:50,676
and so that joke went on and on and on forever,

321
00:27:50,676 --> 00:27:52,776
back in 2015, they would be like,

322
00:27:53,276 --> 00:27:55,376
Ethereum was going to switch to proof of stake,

323
00:27:55,376 --> 00:27:59,776
and now it seems like they may finally actually do it,

324
00:27:59,776 --> 00:28:01,076
which I think there's that metaphor

325
00:28:01,076 --> 00:28:02,976
of when the dog's chasing the car,

326
00:28:02,976 --> 00:28:04,376
and then the dog finally catches the car,

327
00:28:04,376 --> 00:28:05,576
and the dog doesn't know what to do,

328
00:28:05,576 --> 00:28:07,176
and it's like, I don't know what will happen

329
00:28:07,176 --> 00:28:08,776
if they actually turn it on.

330
00:28:08,776 --> 00:28:10,676
No one will be more interested than I

331
00:28:10,676 --> 00:28:12,876
if they actually switch to it

332
00:28:12,876 --> 00:28:13,776
because I just kind of thought

333
00:28:13,776 --> 00:28:15,276
they would just kind of never switch to it,

334
00:28:15,376 --> 00:28:16,476
but we'll see what happens.

335
00:28:16,476 --> 00:28:18,176
Well, it's fascinating that right now,

336
00:28:18,676 --> 00:28:20,976
hashing power on Ethereum is actually going up.

337
00:28:22,076 --> 00:28:23,276
It's not like they're jumping ship,

338
00:28:23,276 --> 00:28:24,776
thinking, oh, this is all going to end.

339
00:28:24,776 --> 00:28:27,076
For whatever reason, they're jumping onto Ethereum,

340
00:28:27,076 --> 00:28:29,076
thinking, well, either we can make some money

341
00:28:29,076 --> 00:28:30,376
or we're going to plot something.

342
00:28:31,576 --> 00:28:33,676
Sorry, but we are all agreeing right now.

343
00:28:33,676 --> 00:28:35,576
Yeah, this is very popular.

344
00:28:35,576 --> 00:28:38,976
Let's talk about this inflation attacker.

345
00:28:38,976 --> 00:28:40,876
Yeah, Giacomo, I was wondering

346
00:28:40,876 --> 00:28:42,276
about what you were going to say

347
00:28:42,276 --> 00:28:44,176
about the Peter Tau's inflation thing

348
00:28:44,176 --> 00:28:46,776
because normally I feel like the two of you always agree,

349
00:28:46,776 --> 00:28:48,976
and that was my first thought when I read this thing

350
00:28:48,976 --> 00:28:52,176
because Giacomo's like the biggest anti-never

351
00:28:52,176 --> 00:28:55,276
at any violation of the 21 million coins,

352
00:28:55,276 --> 00:28:56,576
and then I was like, oh my gosh,

353
00:28:56,576 --> 00:28:59,676
this is finally the divorce between the marriage

354
00:28:59,676 --> 00:29:01,776
of Giacomo and Peter Tau.

355
00:29:01,776 --> 00:29:02,776
Just one Satoshi.

356
00:29:02,776 --> 00:29:03,776
She's going to.

357
00:29:04,976 --> 00:29:07,376
All right, yeah, so for this topic, Peter,

358
00:29:07,376 --> 00:29:09,376
obviously you wrote to the mailing list

359
00:29:09,376 --> 00:29:12,476
about your basically belief that there will need to be.

360
00:29:12,476 --> 00:29:14,176
We're calling it tail coin now.

361
00:29:14,176 --> 00:29:15,476
Tail coin.

362
00:29:15,476 --> 00:29:17,476
Tail coin is going to need to be created.

363
00:29:18,376 --> 00:29:19,476
Why do you believe that is so?

364
00:29:19,476 --> 00:29:20,676
Remember, it's not just tail coin,

365
00:29:20,676 --> 00:29:22,376
it's continuous tail coin.

366
00:29:22,376 --> 00:29:23,876
I like more emission coin

367
00:29:23,876 --> 00:29:26,676
because like a fart is an emission of some gas.

368
00:29:28,176 --> 00:29:30,576
It's TLC for those who want to look it up.

369
00:29:31,576 --> 00:29:33,576
So why do we need tail emissions?

370
00:29:34,676 --> 00:29:37,876
Well, I mean, tail emissions is just one way.

371
00:29:37,876 --> 00:29:42,276
I think the bigger issue is Satoshi designed Bitcoin

372
00:29:42,276 --> 00:29:46,776
so that the inflationary reward that you get

373
00:29:46,776 --> 00:29:48,976
in every block would eventually disappear.

374
00:29:48,976 --> 00:29:50,676
And that means that Bitcoin has a state change.

375
00:29:50,676 --> 00:29:52,476
You start off in one state

376
00:29:52,476 --> 00:29:56,176
where every block will go make you money guaranteed,

377
00:29:56,176 --> 00:29:57,576
maybe not very much in the very beginning,

378
00:29:57,576 --> 00:29:58,976
but certainly like after the first year

379
00:29:58,976 --> 00:30:00,676
or two bootstrapping,

380
00:30:00,676 --> 00:30:02,576
you mine a block, you're going to make money.

381
00:30:02,576 --> 00:30:05,876
Thus, your incentive is keep mining the blockchain forward.

382
00:30:05,876 --> 00:30:07,576
And then in the future, it changes this thing

383
00:30:07,576 --> 00:30:09,876
where now it's all about transaction fees.

384
00:30:09,876 --> 00:30:11,776
And there is no coin in the world

385
00:30:11,776 --> 00:30:14,576
that's proof of work that relies just on transaction fees.

386
00:30:15,376 --> 00:30:16,676
Like nobody's done this.

387
00:30:16,676 --> 00:30:18,076
Nobody's proven it to work.

388
00:30:18,076 --> 00:30:19,876
And it's just from that point alone,

389
00:30:19,876 --> 00:30:21,476
I would say this is crazy.

390
00:30:21,476 --> 00:30:23,776
Going from this thing that is working

391
00:30:23,776 --> 00:30:26,776
to this thing that is completely untested.

392
00:30:27,476 --> 00:30:29,276
Why would you ever go do that

393
00:30:29,276 --> 00:30:32,976
for the sake of having, you know, 0.1% less inflation?

394
00:30:32,976 --> 00:30:35,976
Why going from PayPal to Bitcoin, right?

395
00:30:35,976 --> 00:30:37,976
Nobody did that before.

396
00:30:38,976 --> 00:30:42,076
But why would you do two novel things in once

397
00:30:42,076 --> 00:30:43,476
when you only have to do one?

398
00:30:43,476 --> 00:30:46,876
Well, I think there's an important point here.

399
00:30:46,876 --> 00:30:51,276
Like the addition of tail emissions,

400
00:30:51,276 --> 00:30:53,776
the amount of emissions is arbitrary to begin with.

401
00:30:53,776 --> 00:30:56,276
So you open a Pandora's box of arbitrariness

402
00:30:56,276 --> 00:30:59,976
that if you do it once, it's like, oh, we did 0.1.

403
00:30:59,976 --> 00:31:02,076
There's a lot of arbitrary decisions we go make.

404
00:31:02,076 --> 00:31:04,876
And certainly, I mean, I think the easier,

405
00:31:04,876 --> 00:31:07,576
or I should say the more popular viewpoint

406
00:31:08,276 --> 00:31:10,376
is that Satoshi should have done this at the beginning.

407
00:31:10,376 --> 00:31:12,976
It should have been like one Bitcoin per block forever.

408
00:31:12,976 --> 00:31:14,776
I mean, the exact number, of course, doesn't matter.

409
00:31:14,776 --> 00:31:16,776
But, you know, some amount per block forever

410
00:31:16,776 --> 00:31:18,576
would have been simple and would have worked.

411
00:31:18,576 --> 00:31:20,376
Some, if you want to argue about it,

412
00:31:20,376 --> 00:31:21,976
maybe some amount increasing

413
00:31:21,976 --> 00:31:24,176
so it actually is inflationary.

414
00:31:24,176 --> 00:31:26,276
But like the point is just don't do something

415
00:31:26,276 --> 00:31:28,276
where it has a state change from,

416
00:31:28,276 --> 00:31:29,476
yeah, there's always incentive

417
00:31:29,476 --> 00:31:30,876
to move the blockchain forward

418
00:31:30,876 --> 00:31:34,076
to actually there probably isn't in a whole lot of scenarios.

419
00:31:34,076 --> 00:31:35,776
Yeah, there are two different discussions here.

420
00:31:35,776 --> 00:31:39,176
One is would telemissions or stable emission

421
00:31:39,176 --> 00:31:40,176
be better or worse?

422
00:31:40,176 --> 00:31:41,276
And we can discuss that.

423
00:31:41,276 --> 00:31:44,476
And the other is should we change Bitcoin now

424
00:31:44,476 --> 00:31:45,376
to achieve that?

425
00:31:45,376 --> 00:31:46,776
And I think on the second part,

426
00:31:46,776 --> 00:31:48,976
we are really ready to punch each other

427
00:31:48,976 --> 00:31:50,576
because that's, I mean, that's the no-go.

428
00:31:50,576 --> 00:31:51,576
The first part is-

429
00:31:51,576 --> 00:31:52,876
Sometimes sneaky ways around this

430
00:31:52,876 --> 00:31:55,076
where you don't have to go and, you know,

431
00:31:55,076 --> 00:31:56,476
get into that argument.

432
00:31:56,476 --> 00:31:57,976
Exactly.

433
00:31:57,976 --> 00:32:01,376
Yeah, but as the, if it's a fixed amount of coins per block,

434
00:32:01,376 --> 00:32:03,876
eventually it will be such a small percentage

435
00:32:03,876 --> 00:32:05,476
that it will have no effect anyway, right?

436
00:32:05,476 --> 00:32:06,476
So it won't even be able to pay for-

437
00:32:06,476 --> 00:32:08,676
Which is exactly why I wrote telemission.

438
00:32:08,676 --> 00:32:10,376
And funny enough, it's non-inflationary

439
00:32:10,376 --> 00:32:11,376
because it isn't-

440
00:32:11,376 --> 00:32:12,976
But doesn't that mean also that the idea

441
00:32:12,976 --> 00:32:13,776
serves no purpose?

442
00:32:13,776 --> 00:32:14,676
Well, the thing is, no, no.

443
00:32:14,676 --> 00:32:16,376
The point is, the point of that article

444
00:32:16,376 --> 00:32:19,676
is to show that rather than it converging towards zero,

445
00:32:19,676 --> 00:32:22,276
like you would expect if the coin supply increased forever,

446
00:32:22,276 --> 00:32:24,676
in reality, it doesn't converge towards zero

447
00:32:24,676 --> 00:32:26,576
because people inevitably lose coins.

448
00:32:26,576 --> 00:32:28,576
You know, what it actually converges towards

449
00:32:28,576 --> 00:32:30,876
is a rate of lost coins,

450
00:32:30,876 --> 00:32:33,276
which is, you know, going to be small

451
00:32:33,376 --> 00:32:36,076
because we're not all losing all our money in boat accidents

452
00:32:36,076 --> 00:32:37,576
as much as we say we are.

453
00:32:38,976 --> 00:32:40,076
But it's not going to be zero

454
00:32:40,076 --> 00:32:41,976
because some of us actually do lose our money

455
00:32:41,976 --> 00:32:43,776
in boating accidents or other accidents.

456
00:32:43,776 --> 00:32:45,776
Even if it was zero, your proof still holds

457
00:32:45,776 --> 00:32:48,476
because it just infinitely increases the supply of money

458
00:32:48,476 --> 00:32:50,176
and then at time equals infinity, then finally-

459
00:32:50,176 --> 00:32:52,676
But the supply of money is not infinite in reality

460
00:32:52,676 --> 00:32:53,976
because people lose their coins.

461
00:32:53,976 --> 00:32:56,776
Well, of course, I'm just saying that your math

462
00:32:56,776 --> 00:32:59,876
holds even in that case when it's zero.

463
00:32:59,876 --> 00:33:00,476
Well, I mean, all right.

464
00:33:00,476 --> 00:33:02,576
So in that case, we're so far off in the future,

465
00:33:02,676 --> 00:33:03,576
who really cares?

466
00:33:03,576 --> 00:33:04,376
We're all dead.

467
00:33:04,376 --> 00:33:06,476
This whole article is kind of silly anyway

468
00:33:06,476 --> 00:33:09,076
because it's just saying that eventually

469
00:33:09,476 --> 00:33:12,376
the coins added will be a very small portion.

470
00:33:12,376 --> 00:33:13,676
Yeah, don't you want to build a system

471
00:33:13,676 --> 00:33:15,476
that isn't going to just arbitrarily stop?

472
00:33:15,476 --> 00:33:16,676
The article is really important

473
00:33:16,676 --> 00:33:17,876
because everyone needs to know

474
00:33:17,876 --> 00:33:21,076
that Peter Todd understands differential equations.

475
00:33:21,076 --> 00:33:21,576
That's-

476
00:33:22,376 --> 00:33:23,076
No, I don't.

477
00:33:23,076 --> 00:33:26,076
I used Wolfram Alpha to solve them, just like you.

478
00:33:26,476 --> 00:33:28,276
Why would you ever solve them yourself?

479
00:33:28,276 --> 00:33:29,576
Wolfram Alpha is brilliant.

480
00:33:29,576 --> 00:33:31,176
If you saw his presentation earlier today-

481
00:33:31,276 --> 00:33:32,276
Unpopular opinion.

482
00:33:32,276 --> 00:33:33,276
Don't learn math.

483
00:33:33,276 --> 00:33:34,476
Just use Wolfram Alpha.

484
00:33:34,476 --> 00:33:36,876
I use Wolfram Alpha to be clear.

485
00:33:36,876 --> 00:33:38,676
I don't solve differential equations.

486
00:33:38,676 --> 00:33:39,476
If you understand math,

487
00:33:39,476 --> 00:33:40,876
you definitely want to use Wolfram Alpha.

488
00:33:40,876 --> 00:33:43,776
Sometimes I ask someone else to use Wolfram Alpha for me.

489
00:33:46,576 --> 00:33:47,476
Anyway, sorry.

490
00:33:48,476 --> 00:33:49,076
No, I was going to say,

491
00:33:49,676 --> 00:33:52,076
the way you phrased it or framed it, Peter,

492
00:33:52,176 --> 00:33:57,676
was it's insane to depend on the security of a chain

493
00:33:58,076 --> 00:33:59,376
on transaction fees alone

494
00:33:59,376 --> 00:34:01,676
because we've never seen anybody do it?

495
00:34:01,976 --> 00:34:04,776
Does that mean that you believe it cannot be done?

496
00:34:05,076 --> 00:34:06,476
Is there a possibility in your mind

497
00:34:06,676 --> 00:34:08,976
that a fee market could develop and it could sustain?

498
00:34:09,476 --> 00:34:10,376
I don't know.

499
00:34:10,476 --> 00:34:12,776
And I don't think it's a good idea to risk,

500
00:34:13,776 --> 00:34:14,776
especially in the future,

501
00:34:14,776 --> 00:34:18,376
a multi-trillion dollar asset on this crazy idea

502
00:34:18,376 --> 00:34:20,076
just to go save a few billion dollars.

503
00:34:20,476 --> 00:34:21,576
Is it that crazy, though?

504
00:34:21,876 --> 00:34:22,676
Yeah, I think it is.

505
00:34:22,876 --> 00:34:26,276
And also, there's technical reasons why this is dubious.

506
00:34:26,276 --> 00:34:30,176
One reason is, with transaction fees,

507
00:34:30,476 --> 00:34:32,976
if that's the significant source of revenue,

508
00:34:33,376 --> 00:34:35,876
if someone goes in, has more transaction fees

509
00:34:35,876 --> 00:34:36,976
than average in one period,

510
00:34:36,976 --> 00:34:38,476
it can make sense to reorg blocks.

511
00:34:38,776 --> 00:34:40,376
So, I have a bunch of fees here,

512
00:34:40,976 --> 00:34:42,776
further back in the chain, there's less fees.

513
00:34:42,776 --> 00:34:47,676
Well, it can be worthwhile to re-mine these blocks out of order.

514
00:34:47,776 --> 00:34:49,376
So now I get those transaction fees

515
00:34:49,376 --> 00:34:50,376
rather than some other miner.

516
00:34:50,776 --> 00:34:53,376
And one issue why this, of course, is bad

517
00:34:53,376 --> 00:34:54,676
is because it's very disruptive.

518
00:34:54,776 --> 00:34:57,576
The more subtle issue is because only big miners can do this.

519
00:34:58,076 --> 00:34:59,476
If I have 1% hashing power,

520
00:34:59,476 --> 00:35:01,176
I don't have enough hashing power to reorg.

521
00:35:01,576 --> 00:35:03,976
But if I have 30%, it might be profitable.

522
00:35:04,876 --> 00:35:06,876
Just to clarify, this fee sniping,

523
00:35:06,876 --> 00:35:10,776
since it's been hyped a lot by the Stanford guys that discovered,

524
00:35:10,976 --> 00:35:12,876
this goes to equilibrium anyway,

525
00:35:12,876 --> 00:35:14,376
because if you have a fee sniping,

526
00:35:14,576 --> 00:35:15,576
eventually, in the main pool,

527
00:35:15,576 --> 00:35:18,676
you build up transactions that want to enter and cannot,

528
00:35:18,876 --> 00:35:21,576
and then you build up enough fee to just evolve the chain.

529
00:35:21,576 --> 00:35:23,876
But it makes mining less reality.

530
00:35:24,676 --> 00:35:27,676
But again, I'm not so worried about the fee sniping itself.

531
00:35:27,876 --> 00:35:28,676
That would kind of suck.

532
00:35:28,676 --> 00:35:32,176
But the real issue is that bigger miners make more money on this.

533
00:35:32,176 --> 00:35:33,476
That's the dangerous thing.

534
00:35:33,476 --> 00:35:36,576
Reorgs make mining more centralized.

535
00:35:36,676 --> 00:35:39,076
So the more incentives to reorgs you have.

536
00:35:39,276 --> 00:35:42,476
But the other point is that Eric has very good explanations.

537
00:35:42,876 --> 00:35:45,576
The first I read about that, about the attack of censorship.

538
00:35:45,576 --> 00:35:47,476
So you may have a fee sniping,

539
00:35:47,576 --> 00:35:49,676
but the second big attack you can have is

540
00:35:50,476 --> 00:35:53,576
governments paying miners in order to produce anti-blocks.

541
00:35:53,776 --> 00:35:58,476
In this case, an inflation subsidy will do nothing to prevent it.

542
00:35:58,476 --> 00:36:00,876
It will just diminish the cost of the attack.

543
00:36:00,976 --> 00:36:03,676
And the only answer to this is fees.

544
00:36:03,876 --> 00:36:07,176
So if we assume that nobody is going to pay for block space,

545
00:36:07,776 --> 00:36:10,376
the block space market is going to suck anyway.

546
00:36:10,476 --> 00:36:11,876
We cannot invent money.

547
00:36:11,876 --> 00:36:13,876
We cannot print usage.

548
00:36:13,976 --> 00:36:15,276
We cannot print demand.

549
00:36:15,376 --> 00:36:16,576
We can just print supply,

550
00:36:16,676 --> 00:36:18,976
but we cannot bring demand in artificially.

551
00:36:19,076 --> 00:36:22,376
We just have to accept that nobody cares for block space.

552
00:36:22,476 --> 00:36:25,876
And if nobody does, why do we really have to provide block space?

553
00:36:25,976 --> 00:36:28,876
See, I don't think that really works out because...

554
00:36:30,176 --> 00:36:35,476
Alright, in your realistic scenario where some government wants to spend money to go
screw over Bitcoin,

555
00:36:35,576 --> 00:36:37,676
what's more likely to happen is either

556
00:36:38,076 --> 00:36:42,576
their budget is enough money that they're not trying to go and earn a bit of return

557
00:36:42,576 --> 00:36:44,276
by getting some subsidy in the process.

558
00:36:44,376 --> 00:36:47,276
They're saying, alright, we have a budget for like 10 billion dollars.

559
00:36:47,676 --> 00:36:49,976
Is that enough to go screw over Bitcoin? Yes or no?

560
00:36:50,076 --> 00:36:51,076
Oh, it is? Okay, good.

561
00:36:51,076 --> 00:36:54,876
Let's go build a bunch of mining equipment or buy a bunch or steal a bunch or
whatever.

562
00:36:55,576 --> 00:36:58,476
But what's more important to defeat that is ensuring that

563
00:36:59,776 --> 00:37:02,276
there's a lot of revenue going to miners, period,

564
00:37:02,576 --> 00:37:06,776
that are already out there existing doing something really boring like Bitcoin
mining.

565
00:37:07,276 --> 00:37:09,976
And it's really boring because there aren't shenanigans to go on.

566
00:37:09,976 --> 00:37:13,276
You just go mine things, you get some transaction fees, you get a subsidy,

567
00:37:13,276 --> 00:37:16,576
and it's a very boring process that anyone, you know,

568
00:37:17,176 --> 00:37:20,176
with a bunch of electricity and somewhere to go put all the heat can do.

569
00:37:20,576 --> 00:37:22,976
Like, you want mining to be boring. You don't want it to be interesting.

570
00:37:23,176 --> 00:37:25,376
You don't want rootstock on top of all this stuff, too.

571
00:37:25,576 --> 00:37:28,876
I don't want to have to go install 10 different MergeMine clients for all this.

572
00:37:29,076 --> 00:37:34,176
I want to do one thing, very simple, install Bitcoin Core, hit go, run my mining
pool or whatever.

573
00:37:34,176 --> 00:37:40,076
Oh, but you know that BlindMerge Mining is designed to make it so that the miners
don't have to run any other clients

574
00:37:40,076 --> 00:37:44,476
and they just do what they're doing now, which is include the highest fee-paying
transactions

575
00:37:44,476 --> 00:37:46,376
and then everything else just kind of happens behind the scenes.

576
00:37:46,376 --> 00:37:48,176
And they can make more money by being clever.

577
00:37:48,376 --> 00:37:52,476
So to clarify, before it was me and Paul and Eric against...

578
00:37:52,476 --> 00:37:56,176
No, me and Paul against Eric and Peter, I guess, because you're a little bit...

579
00:37:56,176 --> 00:37:57,276
I don't take sides, man.

580
00:37:58,976 --> 00:38:00,976
You literally haven't taken sides.

581
00:38:01,976 --> 00:38:04,376
I really try not to make opinions and somehow...

582
00:38:04,376 --> 00:38:05,576
We did the whole talk about it.

583
00:38:06,576 --> 00:38:08,176
That is very unpopular.

584
00:38:08,576 --> 00:38:09,476
Maybe that is.

585
00:38:09,676 --> 00:38:10,776
It frustrates people.

586
00:38:10,776 --> 00:38:15,776
But boring or otherwise, when it comes right down to it, if there's a 51% attacker,

587
00:38:17,376 --> 00:38:25,276
there's the only economically rational way they are challenged is by some difference

588
00:38:25,276 --> 00:38:32,976
between the reward to mine, you know, illicit blocks versus others, right?

589
00:38:33,376 --> 00:38:37,676
And that difference can't come from subsidy, right?

590
00:38:37,676 --> 00:38:38,976
So the subsidy doesn't contribute to that.

591
00:38:38,976 --> 00:38:40,376
It's a kind of an orthogonal question.

592
00:38:40,376 --> 00:38:41,976
I view these as independent issues.

593
00:38:42,876 --> 00:38:46,176
Do you think a 51% attacker will be economically rational?

594
00:38:46,476 --> 00:38:50,076
Or have an incentive that's entirely outside of our economic model?

595
00:38:51,076 --> 00:38:53,276
I think both those can be true.

596
00:38:53,576 --> 00:38:56,276
It can be completely rational and outside of our economic model.

597
00:38:56,276 --> 00:38:58,076
That's basically the state, right?

598
00:38:58,376 --> 00:39:03,076
So a state wants to preserve the revenue it gets from its own money, which is
significant.

599
00:39:03,576 --> 00:39:08,976
And if they don't do something about Bitcoin, etc., then they will give that up.

600
00:39:09,176 --> 00:39:14,376
And there's a huge, presumably, financial incentive to prevent that from happening.

601
00:39:14,376 --> 00:39:19,376
So whatever budget they allocate, the only thing that challenges that budget

602
00:39:19,376 --> 00:39:21,176
is some difference between these fee levels.

603
00:39:21,176 --> 00:39:24,476
I mean, first we might presume they'll just ban certain types of transactions.

604
00:39:24,476 --> 00:39:28,976
You can see it's already kind of creeping its way along with KYC requirements on big
miners, etc.

605
00:39:28,976 --> 00:39:34,576
But at some point, somebody's got to pay the miners to take the risk to mine the
illicit transactions.

606
00:39:35,176 --> 00:39:36,476
Subsidy or no subsidy?

607
00:39:36,476 --> 00:39:38,676
This is kind of an independent question, which I find is interesting

608
00:39:38,676 --> 00:39:41,976
because it doesn't really challenge this tail coin thing.

609
00:39:42,076 --> 00:39:46,776
Some people think that a fee market doesn't develop as long as we have a big
subsidy.

610
00:39:46,776 --> 00:39:49,276
I find this a really interesting perspective.

611
00:39:49,276 --> 00:39:52,376
But we have a big subsidy and a fee market has developed on a fee market.

612
00:39:52,376 --> 00:39:53,776
Well, why wouldn't it, right?

613
00:39:53,776 --> 00:39:55,976
People are paying for block space.

614
00:39:55,976 --> 00:39:58,576
The amount of block space is not affected by the subsidy.

615
00:39:58,576 --> 00:40:00,176
And you're competing to get in.

616
00:40:00,176 --> 00:40:01,976
It has nothing to do with the subsidy.

617
00:40:02,176 --> 00:40:06,176
And the amount of the subsidy or the reward, the combined subsidy and fee level,

618
00:40:06,176 --> 00:40:08,576
has nothing to do with miner profitability.

619
00:40:09,376 --> 00:40:10,576
It's a competitive market.

620
00:40:10,776 --> 00:40:11,776
It's a competitive market.

621
00:40:11,776 --> 00:40:14,376
They're going to make the interest rate, the cost of capital, period.

622
00:40:14,376 --> 00:40:16,376
It doesn't matter what the reward level is.

623
00:40:16,376 --> 00:40:18,376
So these are completely unrelated issues.

624
00:40:18,376 --> 00:40:23,376
The question of whether fees will rise to the point where they become significant

625
00:40:23,376 --> 00:40:27,976
in proportion to reward is a question of how much demand there is for block space,

626
00:40:27,976 --> 00:40:31,576
which means demand to transact.

627
00:40:31,576 --> 00:40:31,976
That's it.

628
00:40:31,976 --> 00:40:35,176
That's the only thing that's going to cause fees to rise to some level.

629
00:40:35,176 --> 00:40:38,576
Now, if it happens, doesn't happen, that is going to have an effect, right,

630
00:40:38,576 --> 00:40:42,576
if subsidy drops off on the total hash rate, but not on miner profitability.

631
00:40:42,576 --> 00:40:43,976
That's not going to be affected.

632
00:40:44,376 --> 00:40:48,176
But let me remember, though, we don't actually care about miner profitability

633
00:40:48,176 --> 00:40:49,176
in a lot of these tax centers.

634
00:40:49,176 --> 00:40:53,576
What we actually care about is how much money does it cost to do an attack.

635
00:40:53,576 --> 00:40:53,976
Right.

636
00:40:53,976 --> 00:40:56,776
And that is set by fees.

637
00:40:56,776 --> 00:40:58,376
It's nothing to do with the profitability.

638
00:40:58,376 --> 00:41:01,176
It's the difference between the censored fee level,

639
00:41:01,576 --> 00:41:04,176
the going fee rate for censored fee transactions,

640
00:41:04,576 --> 00:41:06,576
and what people are willing to pay above that

641
00:41:06,576 --> 00:41:08,576
to incent miners to take those transactions.

642
00:41:08,576 --> 00:41:13,376
That fee differential is the only thing that will pay miners to take the risk.

643
00:41:13,976 --> 00:41:16,776
But, well, so one thing I want to unpack is, first of all,

644
00:41:17,776 --> 00:41:23,976
there's a huge difference between 95% of miners refuse to go mine your transaction

645
00:41:24,376 --> 00:41:29,176
and 51% or 100%, of course, if they do an attack,

646
00:41:29,576 --> 00:41:33,376
like, sorry, 100% or 51% if they do an attack of miners refusing.

647
00:41:33,576 --> 00:41:35,176
And the difference there is that,

648
00:41:35,776 --> 00:41:38,976
if miners simply aren't interested in my transaction,

649
00:41:39,176 --> 00:41:42,776
but I can go pay that remaining 5% to go and include it,

650
00:41:42,776 --> 00:41:45,576
and those other miners will build on top of those blocks.

651
00:41:46,376 --> 00:41:47,776
So this is the budget.

652
00:41:49,176 --> 00:41:51,776
Point is, pretty quickly, my transaction will get mined.

653
00:41:52,176 --> 00:41:54,576
And this is just a minor inconvenience.

654
00:41:54,976 --> 00:41:58,176
You know, I pay a little bit more fee, maybe I have to wait a little longer.

655
00:41:58,176 --> 00:42:00,976
But especially with things like Lightning, it doesn't really matter that much.

656
00:42:01,576 --> 00:42:03,576
You know, my transaction will get confirmed,

657
00:42:03,576 --> 00:42:05,576
and my Lightning channel will open, I'll be able to do what I want.

658
00:42:05,576 --> 00:42:08,576
Yeah, the real attack we're talking about is actually reorging

659
00:42:08,576 --> 00:42:12,176
when your transaction, when your illicit transaction is inside one previous block.

660
00:42:12,176 --> 00:42:14,176
As well as attackers who just want to say,

661
00:42:14,176 --> 00:42:16,176
all right, Bitcoin is now AML KYC.

662
00:42:16,376 --> 00:42:19,776
And I don't think that attack is going to be one driven by profit.

663
00:42:19,776 --> 00:42:21,576
I think that attack is going to be one driven by,

664
00:42:21,576 --> 00:42:22,776
well, we are going to force you.

665
00:42:22,976 --> 00:42:26,576
And, you know, either they can go do that in ways

666
00:42:26,576 --> 00:42:28,376
where mining revenue can't help you at all,

667
00:42:28,576 --> 00:42:30,576
or they do that in ways where they're...

668
00:42:31,976 --> 00:42:33,576
We're talking about this.

669
00:42:33,576 --> 00:42:37,976
So what do you guys think the probability of an actual 51% attack is?

670
00:42:37,976 --> 00:42:41,776
Has that probability gone up or down over the last three years?

671
00:42:41,776 --> 00:42:44,976
Well, it is a function of the cost of doing so, right?

672
00:42:44,976 --> 00:42:48,176
So that right now, it's very, very expensive to do such a thing.

673
00:42:48,176 --> 00:42:51,776
Like the spiteful, there's like the spiteful attack

674
00:42:51,776 --> 00:42:55,376
where someone just wants to shut the Bitcoin network off for a while.

675
00:42:55,576 --> 00:42:57,176
And they don't care about,

676
00:42:57,176 --> 00:42:59,776
they don't care if they're just spending money and losing it.

677
00:43:00,176 --> 00:43:01,776
That is a very sort of spiteful thing.

678
00:43:01,776 --> 00:43:03,776
That would depend on how expensive it is.

679
00:43:04,576 --> 00:43:07,576
Do you think that probability would go up or down

680
00:43:07,576 --> 00:43:13,376
if we cut the overall mining reward by say, you know, like 10 times?

681
00:43:14,376 --> 00:43:16,976
Oh, slightly down, irrelevantly down.

682
00:43:16,976 --> 00:43:18,576
There is just a little bit of money.

683
00:43:18,576 --> 00:43:20,776
It costs a little bit more to perform the attack.

684
00:43:20,776 --> 00:43:21,976
What about a hundred times?

685
00:43:22,376 --> 00:43:23,376
What about a thousand times?

686
00:43:23,376 --> 00:43:26,576
Like where's the threshold where this amount of money going into mining

687
00:43:26,576 --> 00:43:28,376
is low enough that people will do it?

688
00:43:28,376 --> 00:43:29,976
Because they have a bit of budget left over,

689
00:43:29,976 --> 00:43:32,576
and you know, the CIA's budget for the end of the year.

690
00:43:32,576 --> 00:43:34,576
There's no cost to achieving 51%.

691
00:43:34,576 --> 00:43:35,576
Mining's profitable.

692
00:43:35,576 --> 00:43:37,176
The bigger you are, the more profitable it is.

693
00:43:37,176 --> 00:43:38,976
No, but he's saying the 51% attack.

694
00:43:39,376 --> 00:43:42,976
So once you're not building on somebody else's blocks

695
00:43:42,976 --> 00:43:45,176
and you've got 51%, what's the cost to you?

696
00:43:45,176 --> 00:43:47,976
Now, you could presume price goes down.

697
00:43:47,976 --> 00:43:49,776
So you're going to lose some capital costs.

698
00:43:49,776 --> 00:43:52,776
But we might see price go up in the face of a 51% attack.

699
00:43:52,976 --> 00:43:56,576
Okay, to be clear, you're saying that there's a capital cost to the 51% attack.

700
00:43:56,576 --> 00:43:59,576
No, I'm saying mining is profitable,

701
00:43:59,576 --> 00:44:02,176
and mining at 51% will be especially profitable.

702
00:44:02,176 --> 00:44:04,576
Well, why can't I do a 51% attack?

703
00:44:04,576 --> 00:44:05,576
You can.

704
00:44:05,976 --> 00:44:06,976
Me, right now?

705
00:44:07,376 --> 00:44:08,576
Just raise the capital.

706
00:44:08,576 --> 00:44:10,176
Ah, okay, but I've got to raise the capital.

707
00:44:10,176 --> 00:44:11,176
Capital is not...

708
00:44:11,176 --> 00:44:12,176
And it has to be at risk.

709
00:44:12,176 --> 00:44:13,176
That's not zero cost.

710
00:44:13,176 --> 00:44:17,776
If you raise the capital, you'll recoup your capital costs

711
00:44:17,776 --> 00:44:19,176
because mining is profitable.

712
00:44:19,376 --> 00:44:21,976
So, for example, the primary threat, the state,

713
00:44:21,976 --> 00:44:26,176
can easily raise the capital and mine profitably.

714
00:44:26,176 --> 00:44:28,176
Otherwise, Bitcoin's still alive.

715
00:44:28,576 --> 00:44:30,376
It's not fair that me and you are not fighting,

716
00:44:30,376 --> 00:44:32,176
so let's do Samurai Wallet, please.

717
00:44:33,176 --> 00:44:35,176
I mean, people have in mind the 51% attack

718
00:44:35,176 --> 00:44:37,776
that would reduce the value of Bitcoin to zero, I think.

719
00:44:37,776 --> 00:44:38,776
That's like the...

720
00:44:38,776 --> 00:44:41,776
The only entity that has the incentive financially

721
00:44:41,776 --> 00:44:44,176
to execute a 51% attack is the state

722
00:44:44,176 --> 00:44:46,176
because it's competing with state money.

723
00:44:46,176 --> 00:44:47,176
Are they competent enough?

724
00:44:48,176 --> 00:44:49,176
Wait, wait.

725
00:44:49,176 --> 00:44:52,176
Well, you say yes, but we have an example last year

726
00:44:52,176 --> 00:44:55,176
of China literally kicked out the whole mining industry,

727
00:44:55,176 --> 00:44:59,576
had an opportune time to seize the mining equipment at the border,

728
00:44:59,576 --> 00:45:03,176
and China arguably has the highest incentive to destroy Bitcoin,

729
00:45:03,176 --> 00:45:04,176
and they didn't do it.

730
00:45:04,176 --> 00:45:06,576
These things will evolve, presumably, over time.

731
00:45:06,576 --> 00:45:08,176
If China doesn't want people using Bitcoin,

732
00:45:08,176 --> 00:45:09,176
they can just ban Bitcoin,

733
00:45:09,176 --> 00:45:11,176
they can get rid of most of the use, right?

734
00:45:11,176 --> 00:45:13,176
But if it continues to get used...

735
00:45:14,176 --> 00:45:16,176
So, the easiest thing for the state to do,

736
00:45:16,176 --> 00:45:17,176
the cheapest thing is pass a law, ban it.

737
00:45:17,176 --> 00:45:18,176
Most of the white...

738
00:45:18,176 --> 00:45:20,176
All the white market disappears overnight.

739
00:45:20,176 --> 00:45:21,176
Gone.

740
00:45:21,176 --> 00:45:23,176
Right, Coinbase, everybody, all the big miners,

741
00:45:23,176 --> 00:45:24,176
they're just gone.

742
00:45:24,176 --> 00:45:25,176
So, China did that.

743
00:45:26,176 --> 00:45:28,176
But as Bitcoin evolves

744
00:45:28,176 --> 00:45:30,176
as a more distributed black market money,

745
00:45:31,176 --> 00:45:32,176
the next phase will...

746
00:45:32,176 --> 00:45:35,176
If it's taking enough revenue away from state monies,

747
00:45:35,176 --> 00:45:37,176
the next step, logical step,

748
00:45:37,176 --> 00:45:40,176
is to attack the entire network from a single point on the earth,

749
00:45:40,176 --> 00:45:41,176
which is a 51% attack.

750
00:45:41,176 --> 00:45:44,176
Do you think Bitcoin's stronger or weaker than it was

751
00:45:44,176 --> 00:45:45,176
before China kicked it out?

752
00:45:45,176 --> 00:45:46,176
I don't have a way to measure that.

753
00:45:46,176 --> 00:45:47,176
I don't know.

754
00:45:48,176 --> 00:45:49,176
I don't like opinions.

755
00:45:49,176 --> 00:45:50,176
That's an opinion, and I'm not going to...

756
00:45:50,176 --> 00:45:51,176
You don't like opinions.

757
00:45:51,176 --> 00:45:52,176
But I can...

758
00:45:52,176 --> 00:45:55,176
Presumably, if you act on economic rationality,

759
00:45:55,176 --> 00:45:58,176
you have to look at the rationality of the state

760
00:45:58,176 --> 00:46:00,176
is not in making money, double spending.

761
00:46:00,176 --> 00:46:03,176
It's not in even shutting down the network.

762
00:46:03,176 --> 00:46:06,176
It's preserving the signage and transparency

763
00:46:06,176 --> 00:46:09,176
of its own money as tax revenue.

764
00:46:09,176 --> 00:46:12,176
And if they don't do it, what's going to happen?

765
00:46:12,176 --> 00:46:13,176
That will go away.

766
00:46:13,176 --> 00:46:14,176
It literally will go away.

767
00:46:14,176 --> 00:46:16,176
The state won't have state money anymore,

768
00:46:16,176 --> 00:46:18,176
and we'll have our Bitcoin paradise.

769
00:46:18,176 --> 00:46:21,176
But we have to assume that they might want to protect it.

770
00:46:21,176 --> 00:46:22,176
Hang on.

771
00:46:22,176 --> 00:46:26,176
So you seem to be talking about the state being financially rational.

772
00:46:26,176 --> 00:46:27,176
Remind me again.

773
00:46:27,176 --> 00:46:30,176
How is it financially rational to go shut down the economy

774
00:46:30,176 --> 00:46:32,176
for two years to go combat the flu?

775
00:46:33,176 --> 00:46:35,176
These things may not look rational to you,

776
00:46:35,176 --> 00:46:37,176
but you're not the state.

777
00:46:37,176 --> 00:46:40,176
I'm pretty sure there's more than just financial rationality

778
00:46:40,176 --> 00:46:41,176
going on here.

779
00:46:41,176 --> 00:46:43,176
I think there are political considerations.

780
00:46:43,176 --> 00:46:47,176
We should build a system that's resilient to crazy people.

781
00:46:47,176 --> 00:46:51,176
Because then they could spend even more money than they would ever create.

782
00:46:51,176 --> 00:46:52,176
Yeah, they could.

783
00:46:52,176 --> 00:46:53,176
They could do a suicide attack.

784
00:46:53,176 --> 00:46:54,176
Absolutely.

785
00:46:54,176 --> 00:46:57,176
We have to consider irrational attacks as well, of course, by the state.

786
00:46:57,176 --> 00:46:59,176
I'll just bring back the point that it can be done.

787
00:46:59,176 --> 00:47:00,176
It's possible.

788
00:47:00,176 --> 00:47:02,176
We can't reject that.

789
00:47:02,176 --> 00:47:05,176
The security model of Bitcoin has to be based on,

790
00:47:05,176 --> 00:47:08,176
probable or not, dealing with it.

791
00:47:08,176 --> 00:47:11,176
And the way, the only way it's dealt with

792
00:47:11,176 --> 00:47:14,176
is by paying people to mine transactions when it's not allowed.

793
00:47:14,176 --> 00:47:17,176
So we need lots of security budget and lots of fees being paid.

794
00:47:17,176 --> 00:47:19,176
It doesn't come from security.

795
00:47:19,176 --> 00:47:22,176
It doesn't come from block subsidy,

796
00:47:22,176 --> 00:47:24,176
and it doesn't come from the kind of prevailing fee rate.

797
00:47:24,176 --> 00:47:27,176
It comes from the amount that people are willing to pay

798
00:47:27,176 --> 00:47:31,176
above the prevailing fee rate to get unapproved transactions approved.

799
00:47:31,176 --> 00:47:33,176
That delta is it.

800
00:47:33,176 --> 00:47:36,176
It's the only additional amount that you're paying to miners

801
00:47:36,176 --> 00:47:37,176
to take those transactions.

802
00:47:37,176 --> 00:47:40,176
So discussion will go to how to increase that delta by, for example,

803
00:47:40,176 --> 00:47:42,176
aggregated demand.

804
00:47:42,176 --> 00:47:44,176
We may have many black market users

805
00:47:44,176 --> 00:47:47,176
that individually cannot go to that threshold.

806
00:47:47,176 --> 00:47:50,176
So we want to aggregate with stuff like multi-party channels.

807
00:47:50,176 --> 00:47:52,176
No, no, we should get rid of Lightning.

808
00:47:52,176 --> 00:47:54,176
No, no, we should get rid of Lightning.

809
00:47:54,176 --> 00:47:56,176
Lightning is the thing that's preventing all these fees

810
00:47:56,176 --> 00:47:58,176
from going on to the main chain.

811
00:47:58,176 --> 00:48:00,176
Far more fees would get paid to the main chain

812
00:48:00,176 --> 00:48:01,176
if we got rid of Lightning.

813
00:48:01,176 --> 00:48:02,176
You know what's unpopular?

814
00:48:02,176 --> 00:48:04,176
It's very popular to say that Bitcoin is going to fail

815
00:48:04,176 --> 00:48:07,176
because nobody will have the block space to open Lightning channels

816
00:48:07,176 --> 00:48:09,176
because it would be too expensive.

817
00:48:09,176 --> 00:48:11,176
But also, there would be fees so low

818
00:48:11,176 --> 00:48:13,176
that nobody would pay for miners.

819
00:48:13,176 --> 00:48:14,176
So these things are popular.

820
00:48:14,176 --> 00:48:17,176
What's unpopular is just, we'll be fine, guys.

821
00:48:17,176 --> 00:48:20,176
We have seen something really interesting happen.

822
00:48:22,176 --> 00:48:25,176
If you go look at fees, they did go up a lot

823
00:48:25,176 --> 00:48:28,176
to the point where transactions on average

824
00:48:28,176 --> 00:48:32,176
would frequently cost, I mean, even $10, $20.

825
00:48:32,176 --> 00:48:34,176
But how much of that can be applied to that spam attack

826
00:48:34,176 --> 00:48:36,176
that happened during that period?

827
00:48:37,176 --> 00:48:38,176
But hang on.

828
00:48:38,176 --> 00:48:41,176
But on top of that, people were willing to go pay these big fees.

829
00:48:41,176 --> 00:48:44,176
And suddenly, of course, Lightning gets better and better

830
00:48:44,176 --> 00:48:46,176
and more and more stuff happens on Lightning

831
00:48:46,176 --> 00:48:48,176
and the fees go drop like crazy.

832
00:48:48,176 --> 00:48:51,176
It's really not clear how does this play out.

833
00:48:52,176 --> 00:48:54,176
I think that just in general,

834
00:48:54,176 --> 00:48:56,176
I think it's not clear at all

835
00:48:56,176 --> 00:48:59,176
that people will pay the high fees in a sustainable way.

836
00:48:59,176 --> 00:49:02,176
They sometimes will in the short term.

837
00:49:02,176 --> 00:49:05,176
They very often will when the price of Bitcoin

838
00:49:05,176 --> 00:49:07,176
is surging upwards.

839
00:49:08,176 --> 00:49:11,176
Also, we have very clever things we can do

840
00:49:11,176 --> 00:49:13,176
to make opening Lightning channels

841
00:49:13,176 --> 00:49:15,176
even cheaper than it is right now.

842
00:49:15,176 --> 00:49:18,176
For one UTXO for an arbitrary number of channel openings.

843
00:49:18,176 --> 00:49:21,176
Well, Eric, this is something you and I talked about last night

844
00:49:21,176 --> 00:49:24,176
and something I've been putting out there as a theory.

845
00:49:24,176 --> 00:49:26,176
Actually, Pierre Rochard, I think, was the first one

846
00:49:26,176 --> 00:49:28,176
to bring this idea to me.

847
00:49:28,176 --> 00:49:30,176
It's the fact that you can apply Jevin's paradox

848
00:49:30,176 --> 00:49:32,176
to a Bitcoin UTXO.

849
00:49:32,176 --> 00:49:34,176
UTXO gets more efficient.

850
00:49:34,176 --> 00:49:35,176
You can do more things with it.

851
00:49:35,176 --> 00:49:37,176
Will that drive demand in the long run

852
00:49:37,176 --> 00:49:40,176
which will then drive a thriving fee market?

853
00:49:40,176 --> 00:49:43,176
Is this an unpopular opinion or just a dumb opinion?

854
00:49:43,176 --> 00:49:47,176
Well, I think there may be some of that

855
00:49:47,176 --> 00:49:49,176
but the main point here is that

856
00:49:49,176 --> 00:49:51,176
these economic effects will not go high enough

857
00:49:51,176 --> 00:49:54,176
to sustain a market where nobody needs block space.

858
00:49:54,176 --> 00:49:56,176
I think that the main point is

859
00:49:56,176 --> 00:49:59,176
is there demand for censorship resistance

860
00:49:59,176 --> 00:50:03,176
using a non-double spending blockchain mechanism?

861
00:50:03,176 --> 00:50:05,176
If the answer is yes...

862
00:50:05,185 --> 00:50:07,185
If the answer is yes, then the market will figure it out.

863
00:50:07,185 --> 00:50:12,185
If the answer is no, there is no trick we can play or economical paradox we can hope
for.

864
00:50:12,185 --> 00:50:18,185
The point is that do we think that people will need Bitcoin censorship resistance?

865
00:50:18,185 --> 00:50:23,185
If yes, great. If not, that means that it's not useful to them.

866
00:50:23,185 --> 00:50:30,185
I can't agree with that. I think part of the issue here is that the amount of money
that will naturally get paid there

867
00:50:30,185 --> 00:50:35,185
in how people actually use the system could easily be much less than an amount of
money

868
00:50:35,185 --> 00:50:37,185
you could easily tolerate with a security tax.

869
00:50:37,185 --> 00:50:42,185
Because right now, the security tax on Bitcoin is approximately like 2%.

870
00:50:42,185 --> 00:50:45,185
That's what the inflationary reward works out to be.

871
00:50:45,185 --> 00:50:47,185
It will spend security, not censorship resistance.

872
00:50:47,185 --> 00:50:51,185
No, I disagree with you on how that works.

873
00:50:51,185 --> 00:50:59,185
But the tax we are paying that goes to pay miners works out to be about 2% of the
entire value of all Bitcoins per year.

874
00:51:00,185 --> 00:51:11,185
At the same time, I suspect an amount of say 0.1% would be more than enough to go
and pay for security against pretty much any realistic attacker.

875
00:51:11,185 --> 00:51:12,185
Maybe 0.01%.

876
00:51:12,185 --> 00:51:17,185
Not the one we are talking about. It's just a security against fee sniping, but not
the one we are talking about.

877
00:51:17,185 --> 00:51:22,185
The point is, that's the amount of money that's paying to protect against these
types of attacks.

878
00:51:22,185 --> 00:51:25,185
It would be very easy for something like Lightning to come out where individually,

879
00:51:25,185 --> 00:51:28,185
there's no reason for me to go pay more fees than I have to.

880
00:51:28,185 --> 00:51:32,185
And my individual payments could be effectively zero of my total value of my
Bitcoin.

881
00:51:32,185 --> 00:51:38,185
And in that scenario, we would be underpaying for security because everyone can just
go pay the lowest amount.

882
00:51:38,185 --> 00:51:42,185
But not just Lightning, even a hodler. A hodler doesn't need fast confirmation.

883
00:51:42,185 --> 00:51:49,185
Exactly. That is exactly why the amounts we may end up paying due to fees may be
much less than something that would be tolerated.

884
00:51:49,185 --> 00:51:52,185
Communist. You want to tax capitalists in order to subsidize...

885
00:51:52,185 --> 00:51:55,185
There's nothing wrong with taxing people for security.

886
00:51:55,185 --> 00:51:56,185
How about the greater good?

887
00:51:57,185 --> 00:52:01,185
I have an unpopular fact. I want to state an unpopular fact.

888
00:52:01,185 --> 00:52:02,185
What's that?

889
00:52:02,185 --> 00:52:07,185
What Peter is referring to as a tax is not a tax. And we talked about this last
night.

890
00:52:07,185 --> 00:52:10,185
I think there's a chance where I'm coming from.

891
00:52:10,185 --> 00:52:14,185
Mining Bitcoin is not a tax on other people who own Bitcoin.

892
00:52:14,185 --> 00:52:18,185
Bitcoin miners pay the full price of every Bitcoin they buy.

893
00:52:18,185 --> 00:52:22,185
They don't get it at a discount like the Fed getting $100 bills for $0.20.

894
00:52:23,185 --> 00:52:26,185
Like a power company. And the power company doesn't give them Bitcoin back.

895
00:52:26,185 --> 00:52:30,185
When a miner mines gold, they pay the power companies, truck drivers, employees.

896
00:52:30,185 --> 00:52:32,185
They pay all these people to produce the gold.

897
00:52:32,185 --> 00:52:39,185
Miners in Bitcoin are paying their own money to produce the Bitcoin at the price
that everybody else is currently paying.

898
00:52:39,185 --> 00:52:41,185
Which is not like any other thing in the world.

899
00:52:41,185 --> 00:52:49,185
When the price of something rises, like gold, just a canonical example, the cost of
producing it has not risen.

900
00:52:50,185 --> 00:52:51,185
Those are not related.

901
00:52:51,185 --> 00:52:57,185
So as the price rises, competition comes in and undercuts each other until that
price is brought back down to the production cost.

902
00:52:57,185 --> 00:53:01,185
Which means instead of your price of gold continuing to stay high,

903
00:53:01,185 --> 00:53:07,185
it's brought back down through increased people willing to let it go at the cost
price.

904
00:53:07,185 --> 00:53:11,185
Whereas with Bitcoin, a miner pays the current price.

905
00:53:11,185 --> 00:53:14,185
He's not going to let it go at half that price because that was his cost.

906
00:53:15,185 --> 00:53:19,185
He has the same demand as existed when he mined it that everybody else had.

907
00:53:19,185 --> 00:53:20,185
Hang on, let's be clear.

908
00:53:20,185 --> 00:53:27,185
Government contractors, when they go get my taxpayer dollar, they go and spend all
the taxpayer dollar of me,

909
00:53:27,185 --> 00:53:30,185
of things like their power, their ranch, etc, etc, etc.

910
00:53:30,185 --> 00:53:32,185
All that money, that's not a tax.

911
00:53:32,185 --> 00:53:36,185
Those government contractors are paying full value for all that return.

912
00:53:36,185 --> 00:53:39,185
For me, I should think good that it's not actually a tax.

913
00:53:39,185 --> 00:53:45,185
The real difference is that if you don't pay mining fees, your transaction is not
included, but you're not going to jail, basically.

914
00:53:45,185 --> 00:53:48,185
In my value, my Bitcoin is still getting diluted.

915
00:53:48,185 --> 00:53:54,185
Look, I love how I've managed to go get someone like you to go say that inflation
isn't actually a tax.

916
00:53:54,185 --> 00:53:56,185
It's absolutely a tax. It's a tax on savings.

917
00:53:56,185 --> 00:54:05,185
Monetary inflation with dollars is a tax because the dollars are produced at 20
cents per hundred, 5 cents per one.

918
00:54:06,185 --> 00:54:09,185
If they were produced at cost, producing them would not be a tax.

919
00:54:09,185 --> 00:54:12,185
Just like producing cars is not a tax on people who own cars.

920
00:54:12,185 --> 00:54:16,185
It doesn't matter how much it costs to go and do the accounting thing.

921
00:54:16,185 --> 00:54:19,185
What matters is the fact that my holdings are now worthless.

922
00:54:19,185 --> 00:54:23,185
I don't care whether or not the government went and siphoned off the money or
didn't.

923
00:54:23,185 --> 00:54:32,185
If you've ever worked with stock ownership, you realize that creating new stock,
that you can create new stock that is non-dilutive.

924
00:54:32,185 --> 00:54:34,185
That's what we're talking about.

925
00:54:34,185 --> 00:54:42,185
So you're basically saying that even if it cost $100 to print a $100 bill, you could
still print trillions of them and then there would definitely be the people who…

926
00:54:42,185 --> 00:54:43,185
There would be more of them.

927
00:54:43,185 --> 00:54:48,185
People who had money under the mattress would be very upset about that and they
would have less purchasing power.

928
00:54:48,185 --> 00:54:49,185
No, that's…

929
00:54:49,185 --> 00:54:51,185
Because the cost is taken out of my ability to buy things.

930
00:54:51,185 --> 00:54:56,185
That's the fact that I'm trying to… that is the idea that I'm trying to correct.

931
00:54:56,185 --> 00:55:02,185
People have conflated supply with the amount in existence.

932
00:55:02,185 --> 00:55:10,185
And they're not… in economics, they are not the same thing and they… the amount in
existence is not what is referred to when we talk about supply and demand.

933
00:55:10,185 --> 00:55:14,185
Supply is the other… is another demand of another person.

934
00:55:14,185 --> 00:55:15,185
It's willingness…

935
00:55:15,185 --> 00:55:16,185
That one sounds unpopular.

936
00:55:16,185 --> 00:55:17,185
No.

937
00:55:17,185 --> 00:55:18,185
Yeah.

938
00:55:18,185 --> 00:55:19,185
This is very unpopular.

939
00:55:19,185 --> 00:55:20,185
This is very unpopular.

940
00:55:20,185 --> 00:55:21,185
This is not unpopular.

941
00:55:21,185 --> 00:55:22,185
Just nerdish.

942
00:55:22,185 --> 00:55:23,185
Let's get back to popular.

943
00:55:23,185 --> 00:55:24,185
This is not an opinion.

944
00:55:24,185 --> 00:55:25,185
I stated what is in every economics textbook.

945
00:55:25,185 --> 00:55:27,185
Supply is not the amount in existence.

946
00:55:28,185 --> 00:55:30,185
Karl Menger who developed these ideas.

947
00:55:30,185 --> 00:55:33,185
Supply and demand are not mentioned in his work.

948
00:55:33,185 --> 00:55:34,185
There's two demanders.

949
00:55:34,185 --> 00:55:43,185
It wasn't until Marshall comes along years later and talks about this and then
people start conflating a demand function with an amount in existence.

950
00:55:43,185 --> 00:55:47,185
Look, I don't care what a bunch of economics and ivory towers would ever say.

951
00:55:47,185 --> 00:55:52,185
The fact of the matter is I have one bitcoin and you are diluting my bitcoin.

952
00:55:52,185 --> 00:55:53,185
It might go to power but…

953
00:55:53,185 --> 00:56:00,185
The fact that you're using the term diluting means you are referring to these
economic concepts which you just claim to not care about.

954
00:56:00,185 --> 00:56:05,185
Dilution of that kind does not exist when you issue new stock at a higher price.

955
00:56:05,185 --> 00:56:11,185
But from a practical point of view, I know the fact that my bitcoin will now go buy
less because there's more of them out there.

956
00:56:11,185 --> 00:56:12,185
I don't care whether…

957
00:56:12,185 --> 00:56:14,185
That's an assumption which is incorrect.

958
00:56:14,185 --> 00:56:17,185
I don't think we'll get through this one.

959
00:56:17,185 --> 00:56:21,185
I mean we've been talking about this for like two days.

960
00:56:21,185 --> 00:56:28,185
But I do think that is a very interesting thing that I never thought about when
people conflate supply with units.

961
00:56:28,185 --> 00:56:31,185
It's unique to bitcoin which is why people have a hard time getting their heads
around it.

962
00:56:31,185 --> 00:56:32,185
And supply is actually a verb.

963
00:56:32,185 --> 00:56:33,185
You're supplying something.

964
00:56:33,185 --> 00:56:36,185
It's not the amount of that thing.

965
00:56:36,185 --> 00:56:38,185
I think that definitely made me think.

966
00:56:38,185 --> 00:56:42,185
Matthew Mazincus is aggressively texting me from the crowd.

967
00:56:42,185 --> 00:56:47,185
He wants me to let everybody know there's $30 trillion in base money.

968
00:56:47,185 --> 00:56:49,185
You subtract $2 trillion in gold reserves.

969
00:56:49,185 --> 00:56:53,185
So that's $28 trillion worth of scenery on top of the gold reserves.

970
00:56:53,185 --> 00:56:59,185
So this equals the cost to the state to attack divided by the attack cost.

971
00:56:59,185 --> 00:57:04,185
So he's trying to provide us with an equation to try to calculate this cost.

972
00:57:04,185 --> 00:57:06,185
He's saying that's how much money they'd be willing to spend?

973
00:57:06,185 --> 00:57:07,185
Yes.

974
00:57:07,185 --> 00:57:09,185
Well, I think there's some truth to that.

975
00:57:09,185 --> 00:57:16,185
I think there's some truth to that that it is about the revenues they would lose
because of bitcoin's existence.

976
00:57:16,185 --> 00:57:18,185
So I think that…

977
00:57:18,185 --> 00:57:20,185
If you do nothing, it goes away.

978
00:57:20,185 --> 00:57:22,185
We're probably not going to settle any of these.

979
00:57:22,185 --> 00:57:24,185
Next unpopular opinion.

980
00:57:24,185 --> 00:57:26,185
Alternative implementations.

981
00:57:26,185 --> 00:57:28,185
Crazy.

982
00:57:28,185 --> 00:57:29,185
What do you mean?

983
00:57:29,185 --> 00:57:30,185
Like bitcoin knots?

984
00:57:30,185 --> 00:57:31,185
What are we talking about?

985
00:57:31,185 --> 00:57:32,185
Bitcoin knots.

986
00:57:32,185 --> 00:57:35,185
Bitcoin first version, second version, third version, fourth version?

987
00:57:35,185 --> 00:57:39,185
People think we just need one team building.

988
00:57:39,185 --> 00:57:42,185
There's a new one that comes out every six months or so, right?

989
00:57:42,185 --> 00:57:43,185
Yeah.

990
00:57:43,185 --> 00:57:46,185
Well, Satoshi had a well-known opinion that it was not a good idea.

991
00:57:46,185 --> 00:57:49,185
Second implementation would be a menace to the network, he said.

992
00:57:49,185 --> 00:57:51,185
But I think we do have new versions.

993
00:57:51,185 --> 00:57:55,185
The idea of the soft fork was that they would all be mutually compatible with each
other.

994
00:57:55,185 --> 00:57:56,185
But they haven't been.

995
00:57:56,185 --> 00:58:01,185
To the extent that that is true remains to be 100% proven, but it's sort of working
so far.

996
00:58:01,185 --> 00:58:06,185
It is a necessary evil that we ever release new versions of Bitcoin Core.

997
00:58:06,185 --> 00:58:14,185
And by releasing code that comes from a completely different code base, you're just
making that into an unnecessary evil.

998
00:58:14,185 --> 00:58:22,185
But there have been, I think the people behind Bitcoin have argued, which is my way
of saying I don't want to really defend this too much.

999
00:58:22,185 --> 00:58:27,185
But I think they said something like when they redid Bitcoin in JavaScript, they
found something.

1000
00:58:27,185 --> 00:58:28,185
They fucked up multiple times.

1001
00:58:28,185 --> 00:58:31,185
Yeah, but didn't they find something that was useful to someone else at some point?

1002
00:58:31,185 --> 00:58:33,185
Well, the thing is, so hang on.

1003
00:58:33,185 --> 00:58:35,185
Now we're dealing in opinions.

1004
00:58:35,185 --> 00:58:39,185
Multiple implementations are a good thing.

1005
00:58:39,185 --> 00:58:42,185
Using multiple implementations is a terrible idea.

1006
00:58:42,185 --> 00:58:43,185
And that's the distinction.

1007
00:58:43,185 --> 00:58:48,185
I personally have written an alternative implementation of the Bitcoin Core
scripting.

1008
00:58:48,185 --> 00:58:50,185
And in the process I learned a bunch about it.

1009
00:58:50,185 --> 00:58:54,185
That code base at the very top says you are fucking crazy if you ever actually use
this.

1010
00:58:54,185 --> 00:58:58,185
Because there's no way it will ever agree with what Bitcoin Core does.

1011
00:58:58,185 --> 00:59:00,185
It's just not going to happen.

1012
00:59:00,185 --> 00:59:04,185
And I know that because I found so many issues with it trying to go develop it.

1013
00:59:04,185 --> 00:59:11,185
Computer science isn't ready for consensus systems where you have two separate code
bases and want them to agree.

1014
00:59:12,185 --> 00:59:13,185
We're not going to get there.

1015
00:59:13,185 --> 00:59:15,185
Most of the core guys would say that.

1016
00:59:15,185 --> 00:59:18,185
Unpopular applause.

1017
00:59:18,185 --> 00:59:21,185
But brave. I respect that.

1018
00:59:21,185 --> 00:59:22,185
I mean it should be unpopular.

1019
00:59:22,185 --> 00:59:24,185
It's a failure of computer science.

1020
00:59:24,185 --> 00:59:25,185
But that's just the reality.

1021
00:59:25,185 --> 00:59:29,185
We're much better off having one code base and trying to change as little as
possible.

1022
00:59:29,185 --> 00:59:31,185
You can't go wrong in this room because if you're popular, you're popular.

1023
00:59:31,185 --> 00:59:34,185
If you're unpopular, then you nailed it.

1024
00:59:34,185 --> 00:59:36,185
So anything goes, I guess.

1025
00:59:36,185 --> 00:59:44,185
So your point would be that different versions of Bitcoin Core upgrades are
different implementations but only slightly diverging.

1026
00:59:44,185 --> 00:59:47,185
So it's less likely to have consensus failure.

1027
00:59:47,185 --> 00:59:48,185
Exactly.

1028
00:59:48,185 --> 00:59:53,185
I've made this argument myself in pull request view.

1029
00:59:53,185 --> 00:59:55,185
Such and such, why are you changing this?

1030
00:59:55,185 --> 00:59:57,185
This is risky for very little gain.

1031
00:59:57,185 --> 01:00:03,185
And changing the entire code base is really risky for very little gain.

1032
01:00:03,185 --> 01:00:07,185
But keep in mind where this really matters is what miners use.

1033
01:00:07,185 --> 01:00:11,185
Because miners are the ones who go produce blocks and miners are the ones who can go
fork the chain.

1034
01:00:11,185 --> 01:00:19,185
Now if it's only, say, you're a merchant and you're running your own copy of Bitcoin
Core that you modified a bunch of ways.

1035
01:00:19,185 --> 01:00:23,185
And it's behind a more stock copy of Bitcoin Core, that's fine.

1036
01:00:23,185 --> 01:00:25,185
All that matters is you might go down a bit.

1037
01:00:25,185 --> 01:00:26,185
That's really not a big deal.

1038
01:00:27,185 --> 01:00:29,185
Are you referring to the July 2015?

1039
01:00:29,185 --> 01:00:32,185
What did you say when you said miners fork the chain?

1040
01:00:32,185 --> 01:00:33,185
Is that what you mean?

1041
01:00:33,185 --> 01:00:36,185
If miners fork the chain, they just lose money.

1042
01:00:36,185 --> 01:00:38,185
No, they screw over all of us.

1043
01:00:38,185 --> 01:00:41,185
No, the block just gets rejected by everybody and they lost money.

1044
01:00:41,185 --> 01:00:43,185
Yeah, but see this is the thing.

1045
01:00:43,185 --> 01:00:44,185
But that's not how this works in practice.

1046
01:00:44,185 --> 01:00:52,185
No, but you said this before, Peter, because you're like everyone has to run their
own node and your full node only protects you and doesn't protect anyone else.

1047
01:00:52,185 --> 01:00:57,185
And then the July 2015 thing happened and then you're like, oh, this is so bad.

1048
01:00:57,185 --> 01:01:01,185
But really it didn't affect anyone running their own full node in the slightest.

1049
01:01:01,185 --> 01:01:02,185
Well, that's not true, though.

1050
01:01:02,185 --> 01:01:03,185
Hang on, hang on.

1051
01:01:03,185 --> 01:01:04,185
That's not true, though.

1052
01:01:04,185 --> 01:01:06,185
So then I was like, eh, I was kind of looking at that.

1053
01:01:06,185 --> 01:01:07,185
That is not true.

1054
01:01:07,185 --> 01:01:16,185
People who are running a full node making the same type of mistake that miners did
to get the chain to fork, they were equally as screwed.

1055
01:01:16,185 --> 01:01:18,185
But that's not miners being most important.

1056
01:01:18,185 --> 01:01:20,185
That's everybody else being most important.

1057
01:01:20,185 --> 01:01:24,185
Miners are just going to lose money if they run some random broken node.

1058
01:01:24,185 --> 01:01:25,185
So there's a nuance here.

1059
01:01:25,185 --> 01:01:31,185
If only miners are the ones crazy enough to run an alternate implementation, then
yes, it doesn't really matter.

1060
01:01:31,185 --> 01:01:32,185
But that's not how the world works.

1061
01:01:32,185 --> 01:01:37,185
The way the world really works is both miners and users will go make this mistake.

1062
01:01:37,185 --> 01:01:39,185
And then that becomes very dangerous.

1063
01:01:39,185 --> 01:01:42,185
And we would like users to be able to make this mistake.

1064
01:01:42,185 --> 01:01:46,185
So is an insulated Libitcoin possible?

1065
01:01:46,185 --> 01:01:48,185
So a consensus library possible?

1066
01:01:48,185 --> 01:01:49,185
Is that possible?

1067
01:01:49,185 --> 01:01:50,185
Sure.

1068
01:01:50,185 --> 01:01:51,185
Yeah.

1069
01:01:51,185 --> 01:01:52,185
It's doable.

1070
01:01:52,185 --> 01:01:53,185
It's kind of annoying and hard to do.

1071
01:01:53,185 --> 01:01:54,185
And also, like, why bother?

1072
01:01:54,185 --> 01:01:57,185
Why not just run Bitcoin Core and then put something behind it?

1073
01:01:57,185 --> 01:01:59,185
It works fine.

1074
01:01:59,185 --> 01:02:03,185
There's so much engineering effort going into this concept.

1075
01:02:03,185 --> 01:02:08,185
The difference of performance between Eric's implementation and Core are staggering,
right?

1076
01:02:08,185 --> 01:02:12,185
The things you tried in the sub-other were, the metrics were...

1077
01:02:12,185 --> 01:02:17,185
You've got to understand that most of the node and all the stuff that goes around
it, which you can't, the node does nothing of its own.

1078
01:02:18,185 --> 01:02:21,185
You need a communication mechanism to the store and all that stuff.

1079
01:02:21,185 --> 01:02:24,185
That has nothing to do with consensus, right?

1080
01:02:24,185 --> 01:02:27,185
P2P protocol has nothing to do with consensus.

1081
01:02:27,185 --> 01:02:29,185
Well, P2P protocol is consensus related.

1082
01:02:29,185 --> 01:02:30,185
The querying interface...

1083
01:02:30,185 --> 01:02:32,185
2015 beta was a...

1084
01:02:32,185 --> 01:02:35,185
You can use carrier pigeons to deliver transactions.

1085
01:02:35,185 --> 01:02:36,185
They can take any form you want.

1086
01:02:36,185 --> 01:02:38,185
It has nothing to do with consensus.

1087
01:02:38,185 --> 01:02:43,185
Consensus is much smaller than the node.

1088
01:02:43,185 --> 01:02:47,185
Peer-to-peer, in reality, ends up being consensus because...

1089
01:02:47,185 --> 01:02:48,185
Yes, it does.

1090
01:02:48,185 --> 01:02:54,185
Because what happens is if you can't get the consensus related data to the other
peers, consensus will fail anyway.

1091
01:02:54,185 --> 01:03:00,185
But what's strange with the peer-to-peer code is that, unlike the Core consensus,
you can layer it.

1092
01:03:00,185 --> 01:03:04,185
Say, for example, we have two completely different networking protocols in Bitcoin.

1093
01:03:04,185 --> 01:03:07,185
If one of the two fails, nothing happens.

1094
01:03:07,185 --> 01:03:10,185
I would argue we've got hundreds of different networking protocols in Bitcoin.

1095
01:03:10,185 --> 01:03:12,185
There are a lot of different things out there.

1096
01:03:12,185 --> 01:03:15,185
I look at the P2P traffic and there's not one.

1097
01:03:15,185 --> 01:03:24,185
But, again, this is why there's a nuance there where P2P is related to consensus,
but in an additive way, unlike the Core consensus.

1098
01:03:24,185 --> 01:03:28,185
In other words, not consensus critical, which can easily be seen on the network.

1099
01:03:28,185 --> 01:03:32,185
No, no. The wallet is not consensus critical because nothing the wallet does is ever
going to affect consensus.

1100
01:03:32,185 --> 01:03:35,185
So the vast majority of what's in a node is not consensus critical code.

1101
01:03:35,185 --> 01:03:47,185
There is very identifiable specific code that's consensus critical, which goes
beyond script, which is the lib consensus subset of Core, which is isolated to just
script.

1102
01:03:47,185 --> 01:03:55,185
There are other checks that happen at the transaction block and chain level, but
they can be identified, can be well-documented, can be well-tested.

1103
01:03:55,185 --> 01:03:58,185
The database is consensus critical, too.

1104
01:03:58,185 --> 01:03:59,185
Sorry?

1105
01:03:59,185 --> 01:04:01,185
Remember, the database is consensus critical, too.

1106
01:04:02,185 --> 01:04:10,185
If your database does not faithfully reproduce what you put in it, you have a
consensus bug.

1107
01:04:10,185 --> 01:04:14,185
And when that actually did happen with Core, they fixed the database.

1108
01:04:14,185 --> 01:04:18,185
They did not accept the consensus criticality of that error.

1109
01:04:18,185 --> 01:04:22,185
They fixed it so that it faithfully reproduced the data that went into it.

1110
01:04:22,185 --> 01:04:27,185
Yes, they did. They took several stages and hard forks to get it fixed, but it got
fixed.

1111
01:04:28,185 --> 01:04:30,185
That did happen.

1112
01:04:30,185 --> 01:04:31,185
It did happen.

1113
01:04:31,185 --> 01:04:36,185
But at first they went back, though, and they said, we do accept this as a…

1114
01:04:36,185 --> 01:04:47,185
We artificially put a constraint on the number of hashes that could exist in a block
so that they wouldn't trigger the database limit, and then they went and fixed the
database limit, and then they went and removed that restriction.

1115
01:04:47,185 --> 01:04:49,185
In the end, the database was fixed.

1116
01:04:49,185 --> 01:04:54,185
It did not keep the consensus that actually existed up to that point.

1117
01:04:54,185 --> 01:04:58,185
How much time do we have? Because I want to see the cage about Drivechain between
Peter and Paul.

1118
01:04:58,185 --> 01:04:59,185
I'll do that any time.

1119
01:04:59,185 --> 01:05:02,185
I was thinking, why don't we do something about lightning?

1120
01:05:02,185 --> 01:05:05,185
Well, I was going to say, we have at least 15 minutes. We have plenty of time.

1121
01:05:07,185 --> 01:05:12,185
Peter and I were talking about an uncomfortable truth that people really don't like
to acknowledge.

1122
01:05:12,185 --> 01:05:13,185
We can do that.

1123
01:05:16,185 --> 01:05:21,185
We need to hard fork Bitcoin at one point because of this Unix timestamping bug.

1124
01:05:21,185 --> 01:05:22,185
Nobody likes to talk about it.

1125
01:05:22,185 --> 01:05:23,185
2032, right?

1126
01:05:24,185 --> 01:05:25,185
2032 bug, but…

1127
01:05:25,185 --> 01:05:27,185
No, no, no, not 2032. It's much later.

1128
01:05:27,185 --> 01:05:28,185
Assigned or unsigned?

1129
01:05:28,185 --> 01:05:30,185
The bug is called 2032.

1130
01:05:30,185 --> 01:05:32,185
There's a signed and unsigned variant of this bug.

1131
01:05:32,185 --> 01:05:46,185
Well, yeah, but so the code, the Bitcoin core code base, so it has a 32-bit unsigned
integer in the block header, and that is used for time, and that will roll over in
2106, and at that point, Bitcoin will cease to exist.

1132
01:05:46,185 --> 01:05:50,185
There are a couple of that, right? There is a block height, which may have…

1133
01:05:50,185 --> 01:05:52,185
But block height is thousands of years in the future.

1134
01:05:53,185 --> 01:05:58,185
If it's treated as signed, it rolls over sooner. There's another bit there if it's
treated as unsigned.

1135
01:05:58,185 --> 01:06:01,185
So I think the unsigned bug is 21-something.

1136
01:06:01,185 --> 01:06:02,185
There is one which is closer.

1137
01:06:02,185 --> 01:06:03,185
And I'm not sure…

1138
01:06:03,185 --> 01:06:04,185
No, there isn't one that's closer.

1139
01:06:04,185 --> 01:06:05,185
I don't remember whether the code actually…

1140
01:06:05,185 --> 01:06:08,185
Nothing that will stop Bitcoin from operating.

1141
01:06:09,185 --> 01:06:12,185
I just don't remember whether the code treats it as signed.

1142
01:06:12,185 --> 01:06:16,185
Satoshi used an awful lot of signed integers where they shouldn't have been signed
integers.

1143
01:06:16,185 --> 01:06:20,185
But the one that actually will make consensus stop is 2106, or whatever the heck
the…

1144
01:06:20,185 --> 01:06:21,185
That's the later one.

1145
01:06:21,185 --> 01:06:24,185
Yeah, but there is an earlier one that will make consensus stop.

1146
01:06:24,185 --> 01:06:28,185
So that would be the signed… That would have to be the signed one, which is 2032.

1147
01:06:28,185 --> 01:06:30,185
But the consensus doesn't use it as signed integer.

1148
01:06:30,185 --> 01:06:34,185
Now, Bitcoiners are mocking us because we don't know about critical consensus bugs.

1149
01:06:34,185 --> 01:06:37,185
We are joking. We know what we're talking about.

1150
01:06:37,185 --> 01:06:40,185
This is not limited to Bitcoin. This is like a Y2K bug.

1151
01:06:40,185 --> 01:06:47,185
It's very much like a Y2K bug, and it exists because the Unix timestamp is limited,
and that's it, right?

1152
01:06:47,185 --> 01:06:51,185
And we've been fixing this Y2K bug since well before Bitcoin ever existed.

1153
01:06:51,185 --> 01:06:53,185
Satoshi doing this was just a complete fuck-up.

1154
01:06:53,185 --> 01:06:56,185
Like, there's no reason this should have ever been done this way.

1155
01:06:56,185 --> 01:07:00,185
But he saved 32 bits. Otherwise, it would have lasted for the next 10 million years.

1156
01:07:00,185 --> 01:07:05,185
I mean, on top of this, the fact the header is hashed directly in the proof-of-work
is also crazy.

1157
01:07:05,185 --> 01:07:06,185
Why do you say that?

1158
01:07:06,185 --> 01:07:11,185
Well, it's crazy because it enables very clever optimization, so we don't want that.

1159
01:07:11,185 --> 01:07:14,185
The ASICs boost attack.

1160
01:07:15,185 --> 01:07:19,185
You know, Bapu likes miners, so he doesn't consider it an attack.

1161
01:07:19,185 --> 01:07:23,185
Yes, I have a very—my view, I've learned, is very different than a lot of people's.

1162
01:07:23,185 --> 01:07:27,185
So maybe it's an unpopular view, but I have this very like Nietzschean kind of view
towards mining,

1163
01:07:27,185 --> 01:07:33,185
that the strong will win and will be protected by the champions of history.

1164
01:07:33,185 --> 01:07:36,185
That's crazy. The strong are centralized.

1165
01:07:36,185 --> 01:07:43,185
Some people have this thing where it's like we have to make sure that the weakest
miners are still in the pool.

1166
01:07:43,185 --> 01:07:45,185
Yes, we want all miners to be equal.

1167
01:07:45,185 --> 01:07:48,185
Well, what do you define as a strong miner?

1168
01:07:48,185 --> 01:07:50,185
We say the strong are centralized.

1169
01:07:50,185 --> 01:07:57,185
If they go against the Nietzschean mining thing, then it's got to be against clever
mining ideas like reusing natural gas.

1170
01:07:57,185 --> 01:08:05,185
I think of my upstream data hash that's mining off-grid on natural gas on a farm
that nobody can ever find as the strongest miner in the world.

1171
01:08:05,185 --> 01:08:09,185
But hang on a minute. But is that strongest in terms of profitability?

1172
01:08:09,185 --> 01:08:16,185
See, the problem we have is, first of all, energy and cooling is inherently
decentralized.

1173
01:08:16,185 --> 01:08:20,185
The way that physics works, especially with cooling—

1174
01:08:20,185 --> 01:08:24,185
But I have to stop you to point out that you're using the word decentralized in a
completely different way.

1175
01:08:24,185 --> 01:08:31,185
It may also be valid, but it's a completely different way than the way that Satoshi
used the concept of Bitcoin as being peer-to-peer,

1176
01:08:31,185 --> 01:08:33,185
where each software node has equal status.

1177
01:08:33,185 --> 01:08:35,185
Who cares what crazy ideas he had in 2009?

1178
01:08:35,185 --> 01:08:37,185
Because he's using this word, and—

1179
01:08:38,185 --> 01:08:41,185
But I'm just talking about the fact that you are—

1180
01:08:41,185 --> 01:08:45,185
What you mean is that in Euclidean space, the thermodynamics is distributed—

1181
01:08:45,185 --> 01:08:47,185
Yes, I mean physically decentralized.

1182
01:08:47,185 --> 01:08:49,185
But I'm just pointing out that this is like sometimes—

1183
01:08:49,185 --> 01:08:51,185
Because this is sometimes like a trick that people do.

1184
01:08:51,185 --> 01:08:52,185
But peer-to-peer—

1185
01:08:52,185 --> 01:08:53,185
Hang on.

1186
01:08:53,185 --> 01:08:55,185
Peer-to-peer being decentralized.

1187
01:08:55,185 --> 01:08:58,185
There are economies of scale in physical mining, not economies of scale.

1188
01:08:58,185 --> 01:09:01,185
We were kicking around ASIC boost, and I just want to make a comment about this.

1189
01:09:01,185 --> 01:09:04,185
And it kind of, I think, leans towards what Paul was saying.

1190
01:09:04,185 --> 01:09:10,185
Somebody's smart enough to figure out some faster way to hash, using the existing
rules, by the way, and then patents it.

1191
01:09:10,185 --> 01:09:13,185
Why does Bitcoin care about patents?

1192
01:09:13,185 --> 01:09:15,185
The whole point is you do what you want.

1193
01:09:15,185 --> 01:09:20,185
This is not kowtowing to monopoly controls, which is what patents are.

1194
01:09:20,185 --> 01:09:25,185
Bitcoin cares about patents because it's hardware, and that itself is its own form
of centralization.

1195
01:09:25,185 --> 01:09:32,185
Again, if you want to mine efficiently and it requires you to violate a patent, just
like mining illegal transactions, that's what you do.

1196
01:09:32,185 --> 01:09:34,185
And you become the strong, and you survive.

1197
01:09:34,185 --> 01:09:37,185
But that's very nice, but the world doesn't work that way.

1198
01:09:37,185 --> 01:09:41,185
It's not an attack to use the most efficient legal—you know, there's no illegality
in Bitcoin.

1199
01:09:41,185 --> 01:09:43,185
It's not an attack to use the most efficient—

1200
01:09:43,185 --> 01:09:44,185
Yes, it is an attack.

1201
01:09:44,185 --> 01:09:45,185
It's an attack against—

1202
01:09:45,185 --> 01:09:47,185
To use the most efficient algorithm you can find.

1203
01:09:47,185 --> 01:09:48,185
I agree.

1204
01:09:48,185 --> 01:09:55,185
And I also have something else to say about ASIC boost that I think is very
unpopular, or at least something that I've never been able to understand myself,
which is that I thought—

1205
01:09:55,185 --> 01:10:02,185
The way it was explained to me was that there was a conflict between SegWit and ASIC
boost, and that activating SegWit would turn ASIC boost on.

1206
01:10:02,185 --> 01:10:06,185
Well, covert ASIC boost, because overt one was possible without SegWit.

1207
01:10:06,185 --> 01:10:14,185
But I don't understand exactly why it was the fact that when SegWit actually
activated, nothing happened to the difficulty immediately afterwards.

1208
01:10:14,185 --> 01:10:15,185
It's because the covert—

1209
01:10:15,185 --> 01:10:17,185
It seems to prove that it had no impact.

1210
01:10:17,185 --> 01:10:30,185
But the covert ASIC boost, so the way that worked was the people doing this in
secret could go and get money from their mining that they weren't telling other
people they could go do, which comes down like commercial contracts between those—

1211
01:10:30,185 --> 01:10:33,185
Yeah, but who the heck was making ASIC boost?

1212
01:10:33,185 --> 01:10:36,185
And, you know, the people they were selling hashing equipment to.

1213
01:10:36,185 --> 01:10:46,185
And, of course, the reason why this didn't change was, if I remember correctly, that
they had switched to non-covert ASIC boost at that point.

1214
01:10:46,185 --> 01:10:49,185
Yeah, the overt one is still going on in blocks right now.

1215
01:10:49,185 --> 01:10:51,185
The covert one was impossible without it.

1216
01:10:51,185 --> 01:10:54,185
Why is there such a big controversy about it, you know?

1217
01:10:54,185 --> 01:10:58,185
Well, it was controversial because in secret they had an incentive to go screw up
SegWit.

1218
01:10:58,185 --> 01:11:00,185
They were stealing hash.

1219
01:11:00,185 --> 01:11:04,185
They could just switch it from the SegWit screwing up version to the non-SegWit.

1220
01:11:04,185 --> 01:11:10,185
When I make my really fast mining code, I'm going to keep it a secret and mine the
crap out of it until everybody else figures it out. That's good business.

1221
01:11:10,185 --> 01:11:13,185
Yeah, we don't want that. In general, in Bitcoin, we do not want people to be
clever.

1222
01:11:13,185 --> 01:11:16,185
Bitcoin does not want that, but that's how it works.

1223
01:11:16,185 --> 01:11:18,185
Who cares how it works? I want to have a good system.

1224
01:11:18,185 --> 01:11:26,185
But we have to establish – we actually want to have this conversation because this
is a very – this is a good conversation, but it's liable to go off on many different
tangents or whatever.

1225
01:11:26,185 --> 01:11:40,185
If you're not going to use the criterion of just the most profitable miner is the
healthiest miner and, therefore, we're actually raising the difficulty to its
highest value and we're making the 51% attack as expensive as possible.

1226
01:11:40,185 --> 01:11:47,185
If you're not using that as your criterion, you should also explain what exactly
you're saying that it should –

1227
01:11:47,185 --> 01:11:49,185
And if code is not law, right?

1228
01:11:49,185 --> 01:11:50,185
Right.

1229
01:11:50,185 --> 01:11:51,185
The code allows this.

1230
01:11:51,185 --> 01:11:54,185
He's saying something like it has to be in a different continent or something.

1231
01:11:54,185 --> 01:12:04,185
Let's do a thought experiment. Let's suppose ASIC boost, rather than being like a 30
or 50 or whatever the heck the percentage was, let's suppose it was a 10 million
percent difference.

1232
01:12:04,185 --> 01:12:08,185
Quantum mining with Grover algorithms, which is quadratic, quadratic.

1233
01:12:08,185 --> 01:12:11,185
Sure, but you know perfectly well that the difficulty would reset and then –

1234
01:12:11,185 --> 01:12:16,185
Let's suppose it was some very, very clever thing that cut difficulty by like 10
million.

1235
01:12:16,185 --> 01:12:19,185
That would imply some consensus rule that allowed that.

1236
01:12:19,185 --> 01:12:20,185
Yeah, but afterwards –

1237
01:12:20,185 --> 01:12:28,185
But again, listen, from the point of view of Bitcoiners, would we want to have a
system which had a secret 10 million advantage?

1238
01:12:28,185 --> 01:12:35,185
We want symmetry. So, every asymmetry in mining is not illegal or immoral. It's just
dangerous for the future of attacks.

1239
01:12:35,185 --> 01:12:40,185
So, we want to – if we can, we want to minimize asymmetries between miners.

1240
01:12:40,185 --> 01:12:46,185
So, I mean, I don't think we're really disagreeing here. We would all like to see
people mining on an equal footing.

1241
01:12:46,185 --> 01:12:55,185
We would like to see the strong survive for various reasons, but the fact that the
code allows these things and people take advantage of them does not make them evil
people.

1242
01:12:55,185 --> 01:12:58,185
Who cares who are evil? I just don't want them to exist.

1243
01:12:59,185 --> 01:13:05,185
Probably we have to distinguish between something which is bad for Bitcoin
functionally and something which is morally bad.

1244
01:13:05,185 --> 01:13:12,185
I mean, I think we can agree that we would fix this issue, like SegWit took care of
the ASIC boost, essentially, and that's a good thing.

1245
01:13:12,185 --> 01:13:19,185
But people taking advantage of what the code allows them to do, what the rules allow
them to do, is not somebody being evil or keeping it a secret.

1246
01:13:19,185 --> 01:13:21,185
It's not somebody being evil, right? That's just good business.

1247
01:13:22,185 --> 01:13:27,185
That's a very huge discussion, like Satoshi Dice bloating the chain. Is that evil?

1248
01:13:27,185 --> 01:13:32,185
I mean, you are damaging the experience for everybody else, but if somebody can do
it, then you can do it.

1249
01:13:32,185 --> 01:13:39,185
Or the DAO hacker. Was he a hacker? No, he was just a lawyer because the code is low
and the code allows him to take money.

1250
01:13:39,185 --> 01:13:40,185
He was a good Samaritan.

1251
01:13:40,185 --> 01:13:41,185
He was a good lawyer.

1252
01:13:41,185 --> 01:13:42,185
He was a good Samaritan.

1253
01:13:43,185 --> 01:13:47,185
I mean, I historically have argued quite strongly things like Satoshi Dice.

1254
01:13:47,185 --> 01:13:52,185
We should stop caring about whether or not spamming and think about how do we design
a system where this doesn't matter.

1255
01:13:52,185 --> 01:13:58,185
But at the same time, we should be clear that we do want mining to be a level
playing field.

1256
01:13:58,185 --> 01:14:06,185
This is good for Bitcoin because it means mining is more decentralized and there's
fewer people that we can go and do things to, to go and make mining...

1257
01:14:07,185 --> 01:14:13,185
Yeah, but I don't think this is actually... I think this is not conceived the right
way.

1258
01:14:13,185 --> 01:14:21,185
I think it's overgeneralization. You want the node to be very easy for everyone to
run, and I don't think anyone disagrees with that.

1259
01:14:21,185 --> 01:14:28,185
That is the cost of joining the network. That is the cost of getting the view into
the... So everyone agrees about that.

1260
01:14:28,185 --> 01:14:34,185
But then this view is often reapplied that it should be very, very easy to join the
mining world.

1261
01:14:34,185 --> 01:14:41,185
I don't know if that makes any sense. I think, first of all, if the miners are at
different scales, that still doesn't necessarily mean that small people can't join.

1262
01:14:41,185 --> 01:14:44,185
They can just form a little corporation and become part of something.

1263
01:14:44,185 --> 01:14:50,185
I mean, what is the ultimate objective? We just want mining to be difficult to take
over by the state, which is the whole point of Bitcoin.

1264
01:14:50,185 --> 01:14:58,185
Yes, if you want that, though, then every time you sacrifice the efficiency of
mining, you're leaving money on the table.

1265
01:14:58,185 --> 01:15:02,185
You're just saying it's making the state easier to attack Bitcoin.

1266
01:15:02,969 --> 01:15:06,969
And you're doing that because, you know, you feel bad that some person feels left
out of the mining or something.

1267
01:15:06,969 --> 01:15:09,969
And that just sounds like common sense to me.

1268
01:15:09,969 --> 01:15:15,969
Would any of us disagree that we would like reward to be proportional to work? I
mean, assuming equal efficiency.

1269
01:15:15,969 --> 01:15:19,969
Well, but that's the thing, is people are constantly improving the efficiency.

1270
01:15:19,969 --> 01:15:25,969
Right, but aside from somebody's efficiency versus somebody else's because their
operations or whatever,

1271
01:15:25,969 --> 01:15:31,969
we would like Bitcoin to produce a proportional reward given the amount of work,
right?

1272
01:15:31,969 --> 01:15:32,969
Let me put it like this.

1273
01:15:32,969 --> 01:15:33,969
And it doesn't do that.

1274
01:15:33,969 --> 01:15:37,969
Can I put it like this? Like the state, when it attacks, it will choose the maximum
efficiency.

1275
01:15:37,969 --> 01:15:41,969
Which means the largest mine possible, right? It's going to put it all in one place.

1276
01:15:41,969 --> 01:15:43,969
Largest mining is not most efficient.

1277
01:15:43,969 --> 01:15:44,969
Sorry?

1278
01:15:44,969 --> 01:15:46,969
The largest mining is not most efficient.

1279
01:15:46,969 --> 01:15:48,969
The set of mines that are closest together have an advantage.

1280
01:15:48,969 --> 01:15:51,969
No, it's not. The state will be the most efficient.

1281
01:15:51,969 --> 01:15:55,969
That's why we should also have the defense pick the most efficient.

1282
01:15:55,969 --> 01:16:00,969
So first of all, the statement that the largest miner is most efficient, it is not.

1283
01:16:00,969 --> 01:16:01,969
I think it's not necessarily.

1284
01:16:01,969 --> 01:16:04,969
It fundamentally is not because there are disincentives to scale in mining.

1285
01:16:04,969 --> 01:16:05,969
There are many disincentives.

1286
01:16:05,969 --> 01:16:07,969
Fundamental physics disincentives.

1287
01:16:07,969 --> 01:16:13,969
Largest as in a hash rate collusion, not largest as in the physical facility of a
mining block.

1288
01:16:13,969 --> 01:16:17,969
Distance between seen blocks, right, makes a difference.

1289
01:16:17,969 --> 01:16:21,969
You see as a larger, meaning you're physically closer together.

1290
01:16:21,969 --> 01:16:24,969
All other efficiencies you may have versus somebody else's side.

1291
01:16:24,969 --> 01:16:28,969
If you're physically closer together, information takes time to travel.

1292
01:16:28,969 --> 01:16:29,969
You see your own...

1293
01:16:29,969 --> 01:16:30,969
But that's a very small...

1294
01:16:30,969 --> 01:16:32,969
Yeah, that's a very, very small...

1295
01:16:32,969 --> 01:16:33,969
The cooling disadvantages you get are...

1296
01:16:33,969 --> 01:16:35,969
I'm not arguing in orders of magnitude.

1297
01:16:35,969 --> 01:16:39,969
I'm just saying that if you're closer together, you see your own blocks before other
people see them.

1298
01:16:39,969 --> 01:16:41,969
Therefore, you're mining on the new blocks faster.

1299
01:16:41,969 --> 01:16:42,969
I agree with that.

1300
01:16:42,969 --> 01:16:43,969
This is a well-established principle in Bitcoin.

1301
01:16:43,969 --> 01:16:45,969
It is, but I think they now do it...

1302
01:16:45,969 --> 01:16:48,969
But it's tiny because we have 10-minute block controls.

1303
01:16:48,969 --> 01:16:50,969
That's why it makes very little difference.

1304
01:16:50,969 --> 01:16:51,969
I agree with that.

1305
01:16:51,969 --> 01:16:53,969
They do the spy mining and SPV mining.

1306
01:16:53,969 --> 01:16:56,969
They've done that for many years, so they can find very quickly...

1307
01:16:56,969 --> 01:16:59,969
You're talking like less than one second into the 600 seconds...

1308
01:16:59,969 --> 01:17:05,969
But in a decentralized world where everybody's more spread out, right, it starts to
make a bigger difference

1309
01:17:05,969 --> 01:17:07,969
when one miner can put them all together.

1310
01:17:07,969 --> 01:17:09,969
That's where it starts to matter.

1311
01:17:09,969 --> 01:17:12,969
We don't see it now because you basically have one big...

1312
01:17:12,969 --> 01:17:13,969
Everybody's spy mining.

1313
01:17:13,969 --> 01:17:14,969
Everybody else giant mines everywhere.

1314
01:17:14,969 --> 01:17:16,969
And people don't see orphans.

1315
01:17:16,969 --> 01:17:21,969
But when you spread this out and everybody's hiding and mining illegally, you will
see a lot of orphans.

1316
01:17:21,969 --> 01:17:24,969
No, I don't think we will because they do the spy mining.

1317
01:17:24,969 --> 01:17:26,969
Yeah, we still won't because 10 minutes is very big.

1318
01:17:26,969 --> 01:17:29,969
Do we think the state is competent to do all this, though?

1319
01:17:29,969 --> 01:17:33,969
Well, they will hire people who are, and I wouldn't assume they can't.

1320
01:17:33,969 --> 01:17:38,969
I consider it a fatal flaw in anybody's security model to assume the enemy is
stupid.

1321
01:17:38,969 --> 01:17:43,969
Also, if it's very, very cheap, then we could get any, you know, whatever.

1322
01:17:43,969 --> 01:17:47,969
Someone with a visa or someone could do it just to, you know, I don't know why they
would...

1323
01:17:47,969 --> 01:17:49,969
Also, they don't need to be clever.

1324
01:17:49,969 --> 01:17:51,969
They can just outspend us.

1325
01:17:51,969 --> 01:17:58,969
Like, you know, if being clever means you're 10% more efficient or you just go and
spend, you know, 50% more,

1326
01:17:58,969 --> 01:18:00,969
the 50% more guy will go win.

1327
01:18:00,969 --> 01:18:04,969
Well, governments around the world are throwing trillions of dollars at this energy
crisis,

1328
01:18:04,969 --> 01:18:06,969
and it's not going to fix that anytime soon, so...

1329
01:18:06,969 --> 01:18:08,969
How much time do we still have? I'm just curious.

1330
01:18:08,969 --> 01:18:11,969
Oh, we have like 10 minutes. We should probably... or 5 minutes.

1331
01:18:12,969 --> 01:18:15,969
Why is not Carballo on the panel? I mean...

1332
01:18:15,969 --> 01:18:24,969
I have a question about how many people here think that the beliefs of the experts
in the lightning network world

1333
01:18:24,969 --> 01:18:31,969
accurately match the ones, beliefs held by maybe let's call them influencers or the
audience or the layperson.

1334
01:18:31,969 --> 01:18:38,969
To a certain extent, which is on a scale of 0% to negative 100% or a scale of 100%
to whatever you like.

1335
01:18:38,969 --> 01:18:39,969
So, if the lightning network...

1336
01:18:39,969 --> 01:18:44,969
We're talking like blue mat versus like maybe random guy over there or something,
you know.

1337
01:18:44,969 --> 01:18:47,969
So, if it was overhyped by experts?

1338
01:18:47,969 --> 01:18:51,969
I'm just saying if you think it's... No, I think the experts have a more sober view.

1339
01:18:51,969 --> 01:18:53,969
That's actually what I think is the case.

1340
01:18:53,969 --> 01:18:56,969
See, lightning is written in C. That's crazy.

1341
01:18:56,969 --> 01:19:03,969
Right there, there's a major disconnect between most programmers and like what those
experts decided to go do.

1342
01:19:03,969 --> 01:19:06,969
Yeah, but that's not what I mean. I mean like, you know, like...

1343
01:19:07,969 --> 01:19:09,969
But C is faster than C++.

1344
01:19:09,969 --> 01:19:10,969
Rust.

1345
01:19:11,969 --> 01:19:13,969
C is faster than everything, isn't it?

1346
01:19:15,969 --> 01:19:17,969
What do you think, Giacomo? You got anything for that or no?

1347
01:19:17,969 --> 01:19:19,969
You think like everyone understands lightning?

1348
01:19:19,969 --> 01:19:30,969
So, I think that the scaling, the block size debate made clear some fundamental
limitations of the blockchain design that was not clear for anybody.

1349
01:19:30,969 --> 01:19:36,969
It was clear for like Alphine in 2009 immediately, but for most people not until the
war, basically.

1350
01:19:36,969 --> 01:19:39,969
So, blockchain sucks. They cannot scale.

1351
01:19:39,969 --> 01:19:43,969
And I think that this created a lot of awareness about the problem.

1352
01:19:43,969 --> 01:19:54,969
The solution, which is lightning, which is removing stuff for the global consensus
as a theoretical solution, received a lot of hype and attention because it's the
only reasonable solution.

1353
01:19:54,969 --> 01:19:57,969
And I think that that kind of hype was justified.

1354
01:19:57,969 --> 01:20:04,969
The actual implementation of the most popular off-chain protocol, which is
lightning, received a little bit of overhype.

1355
01:20:04,969 --> 01:20:14,969
I remember people telling me, we will do one million transactions per second, which
will be possible in theory, but it will not be possible for database management of
the channel updates.

1356
01:20:14,969 --> 01:20:16,969
So, there has been overhype of lightning.

1357
01:20:16,969 --> 01:20:27,969
I think it's already self-corrected because we were thinking like lightning will
destroy altcoins, lightning will make privacy perfect, lightning will scale.

1358
01:20:27,969 --> 01:20:31,969
We already have a self-correction of this over-expectation.

1359
01:20:31,969 --> 01:20:33,969
Now, I think the perception is realistic.

1360
01:20:33,969 --> 01:20:40,969
Lightning is great for some things, like fast payments, small payments, and network
and chain privacy.

1361
01:20:40,969 --> 01:20:51,969
Lightning sucks for other things, which is basically large payments, security
long-term, and network privacy because you're basically reusing your address all the
time.

1362
01:20:51,969 --> 01:20:55,969
So, there are trade-offs, and having both is way better than having one.

1363
01:20:55,969 --> 01:21:02,969
So, experts were not so wrong in the idea, but we exaggerated with the hype.

1364
01:21:02,969 --> 01:21:04,969
We did it with Taproot as well.

1365
01:21:04,969 --> 01:21:09,969
I mean, some people were actually expecting something magical happening after
Taproot adoption.

1366
01:21:09,969 --> 01:21:13,969
Literally nothing happened because it's something we need for the future.

1367
01:21:13,969 --> 01:21:16,969
So, there is a risk of overhyping stuff, yes.

1368
01:21:18,969 --> 01:21:20,969
Peter, you got anything on that or not?

1369
01:21:20,969 --> 01:21:21,969
I'm probably guilty of it sometimes.

1370
01:21:21,969 --> 01:21:22,969
Do we have time?

1371
01:21:22,969 --> 01:21:23,969
How much time do we have?

1372
01:21:23,969 --> 01:21:27,969
I don't know. Max, we have time.

1373
01:21:27,969 --> 01:21:39,969
So, Giacomo, the scaling strategy for sidechains is also, as I presented in your
very presence in Milan, it has two aspects, one of which I presented in 2016.

1374
01:21:39,969 --> 01:21:45,969
It also involves, though, hiding transactions from people and getting them out of
the global consensus.

1375
01:21:45,969 --> 01:21:50,969
So, I'm just wondering what everyone's giant problem with that is since it's this
exact same thing as the Lightning Network.

1376
01:21:50,969 --> 01:21:56,969
Because I don't think it gets them out of the global consensus because miners will
go find ways to make more money by taking advantage of that.

1377
01:21:56,969 --> 01:21:58,969
But, again, it has nothing to do with the miners.

1378
01:21:58,969 --> 01:22:02,969
The miners can maybe find an advantage by digging up natural gases.

1379
01:22:02,969 --> 01:22:07,969
But miners are the things that we need to be decentralized.

1380
01:22:07,969 --> 01:22:09,969
Like, yes, absolutely, it gets out of the global consensus.

1381
01:22:09,969 --> 01:22:11,969
But you could make the same argument about SPV.

1382
01:22:11,969 --> 01:22:14,969
SPV means all this blockchain stuff gets out of the global consensus.

1383
01:22:14,969 --> 01:22:16,969
You don't have to pay attention to it.

1384
01:22:16,969 --> 01:22:21,969
Yeah, but I don't agree that miners' costs have anything to do with really anything
at all.

1385
01:22:22,969 --> 01:22:27,969
Because, first of all, if you care about miners' costs, then what you should do is
you should get rid of all the upward difficulty adjustments.

1386
01:22:27,969 --> 01:22:30,969
Sorry, now we're talking mostly about blind.

1387
01:22:30,969 --> 01:22:34,969
No, this is where there's a big difference between marginal costs and overhead
costs.

1388
01:22:34,969 --> 01:22:37,969
We're okay with marginal costs, but overhead costs are very dangerous.

1389
01:22:37,969 --> 01:22:39,969
Because it makes it much harder to go set up a new pool.

1390
01:22:39,969 --> 01:22:41,969
Yeah, but that doesn't make any sense either, Peter.

1391
01:22:41,969 --> 01:22:47,969
Because the overhead costs of running these nodes, these sidechain nodes, that
theoretically they could have any cost.

1392
01:22:47,969 --> 01:22:52,969
But in practice, the network won't exist unless regular full-node people can run
them.

1393
01:22:52,969 --> 01:22:55,969
Like a 0% miner can run them, basically.

1394
01:22:55,969 --> 01:23:02,969
And that means it's many, many orders of magnitude smaller than what it would even
cost to set up a normal mining operation.

1395
01:23:02,969 --> 01:23:08,969
So let's move from merge mining to actually hashrate escrow, which is the main
contention about Drivechains.

1396
01:23:08,969 --> 01:23:14,969
So I think that Drivechains are an example of off-chain evolution that I think is
positive overall.

1397
01:23:14,969 --> 01:23:20,969
But the main concern people have about Drivechains is that miners right now have one
kind of political power.

1398
01:23:20,969 --> 01:23:24,969
Which is censoring and colluding with double spends.

1399
01:23:24,969 --> 01:23:29,969
If you include a successful Drivechain in the model, now they have an additional
political power.

1400
01:23:29,969 --> 01:23:33,969
And you cannot stop people from running Drivechain code.

1401
01:23:33,969 --> 01:23:38,969
But if I have to choose a world where they have an additional political power, I may
be less concerned if they haven't.

1402
01:23:38,969 --> 01:23:44,969
But this is a case where the political power that they have is provisional and given
to them by the user.

1403
01:23:44,969 --> 01:23:48,969
So it's no different than if the user sold their Bitcoin and bought Solana or
something.

1404
01:23:48,969 --> 01:23:55,969
If you're not using any of the BIP300 sidechains, then the miners do not have any
additional power whatsoever of any kind.

1405
01:23:55,969 --> 01:24:01,969
And if you do opt in, then that's a risk the user has chosen to take.

1406
01:24:01,969 --> 01:24:03,969
And it's no different than if they joined a blockchain.

1407
01:24:03,969 --> 01:24:07,969
I get screwed over because now being a miner is harder to go do.

1408
01:24:07,969 --> 01:24:09,969
And thus we're centralizing mining.

1409
01:24:09,969 --> 01:24:12,969
Just adding choices to mining makes it harder to go do.

1410
01:24:12,969 --> 01:24:16,969
Wait, I have a question I wanted to ask yesterday when you and John were up here
talking about this.

1411
01:24:16,969 --> 01:24:20,969
Does this add an MEV aspect to mining?

1412
01:24:20,969 --> 01:24:21,969
No.

1413
01:24:21,969 --> 01:24:23,969
Does that perturb incentives?

1414
01:24:23,969 --> 01:24:27,969
It's possible to make any mistake you like on the sidechain.

1415
01:24:27,969 --> 01:24:29,969
So the sidechain could have MEV.

1416
01:24:29,969 --> 01:24:33,969
The way it's designed is that all the drama goes in one direction.

1417
01:24:33,969 --> 01:24:37,969
So if the mainchain reorgs, then everything above it would have to reorg.

1418
01:24:37,969 --> 01:24:39,969
But the reverse is not the case.

1419
01:24:39,969 --> 01:24:46,969
So if a sidechain has MEV and it's being reorged all the time because it's badly
designed stuff up for grabs up there,

1420
01:24:46,969 --> 01:24:51,969
then someone who's just running Bitcoin Core today will not even notice.

1421
01:24:51,969 --> 01:24:53,969
So the answer is no.

1422
01:24:53,969 --> 01:24:55,969
Yeah, that's another unpopular opinion.

1423
01:24:55,969 --> 01:24:57,969
MEV is actually good for chain security.

1424
01:24:58,969 --> 01:24:59,969
Complexity is bad.

1425
01:24:59,969 --> 01:25:01,969
Complexity is very bad.

1426
01:25:01,969 --> 01:25:05,969
Ideally, in a hypothetical world where everything was perfect,

1427
01:25:05,969 --> 01:25:07,969
miners would not see transactions.

1428
01:25:07,969 --> 01:25:10,969
They would just see a bunch of encrypted data and they'd have no clue what it was,

1429
01:25:10,969 --> 01:25:12,969
they'd have no way to extract value out of it.

1430
01:25:12,969 --> 01:25:14,969
They would just magically get money, and that would be that.

1431
01:25:14,969 --> 01:25:17,969
Yeah, but I don't know why you say that, though.

1432
01:25:17,969 --> 01:25:22,969
Because part of what the blockchain is doing is purging the contradictory double
spends.

1433
01:25:23,969 --> 01:25:29,969
But I'm saying in a perfect world where we could have the magical crypto to make
this possible...

1434
01:25:29,969 --> 01:25:31,969
Yeah, but if I can do it, then the miners can also do it, right?

1435
01:25:31,969 --> 01:25:34,969
If I'm going to look into the blockchain and see if there's double spends there,

1436
01:25:34,969 --> 01:25:36,969
then the miners have to be able to do it.

1437
01:25:36,969 --> 01:25:37,969
I think I missed my point.

1438
01:25:37,969 --> 01:25:42,969
If we had the cryptography to encrypt transactions and somehow prevent double spends
even there,

1439
01:25:42,969 --> 01:25:44,969
I'm not saying it's possible.

1440
01:25:44,969 --> 01:25:49,969
I'm just saying if we had that world, that would be categorically better than what
we have right now,

1441
01:25:49,969 --> 01:25:51,969
which is where miners can go see what transactions do.

1442
01:25:51,969 --> 01:25:54,969
We do not want miners to have any influence in transactions other than,

1443
01:25:54,969 --> 01:25:57,969
yeah, I created a block, and I got a bit of money, and that was that.

1444
01:25:57,969 --> 01:26:00,969
In the client-side validation model, they will not see the content,

1445
01:26:00,969 --> 01:26:02,969
assuming they're not colluded with it.

1446
01:26:02,969 --> 01:26:06,969
So if I pass you off-chain the content of the transaction, they only see the
single-use seals.

1447
01:26:06,969 --> 01:26:07,969
Yeah, yeah.

1448
01:26:07,969 --> 01:26:10,969
I mean, this is why I came up with stuff like single-use seals.

1449
01:26:10,969 --> 01:26:12,969
In RGV, they have no idea what they're doing.

1450
01:26:12,969 --> 01:26:15,969
You still need to pay the transaction fee to the miner in something that they
understand.

1451
01:26:15,969 --> 01:26:16,969
That's the only thing.

1452
01:26:16,969 --> 01:26:17,969
You pay a fee to publish some data.

1453
01:26:18,969 --> 01:26:24,969
There's this idea that, again, if the user—like you believe, both of you over there,
you two jokers over there,

1454
01:26:24,969 --> 01:26:29,969
you believe that people have the right to sell their Bitcoin for fiat if they
choose, right?

1455
01:26:29,969 --> 01:26:31,969
And they can spend their Bitcoin on goods and services?

1456
01:26:31,969 --> 01:26:33,969
No, they're not allowed to sell their Bitcoin for fiat.

1457
01:26:33,969 --> 01:26:35,969
They're only allowed to hodl.

1458
01:26:35,969 --> 01:26:36,969
Okay.

1459
01:26:36,969 --> 01:26:38,969
But, you know, they're the owner.

1460
01:26:38,969 --> 01:26:41,969
So what I'm saying is they should be allowed to pay into this script

1461
01:26:41,969 --> 01:26:45,969
where they have an additional vulnerability, if that's true.

1462
01:26:45,969 --> 01:26:49,969
Nobody's arguing against any user doing anything with code.

1463
01:26:49,969 --> 01:26:52,969
They can use shitcoin, change Bitcoin.

1464
01:26:52,969 --> 01:26:58,969
The problem is what I hope will be widespread or not,

1465
01:26:58,969 --> 01:27:02,969
and how I can influence that by speaking up about one option and the other.

1466
01:27:02,969 --> 01:27:07,969
We will never initiate violence about shitcoiners or drive-chainers.

1467
01:27:07,969 --> 01:27:09,969
But you understand the point, though.

1468
01:27:09,969 --> 01:27:16,969
The analogy is that they can deposit coins into this BIP300 Drivechain or whatever
you want to call it.

1469
01:27:16,969 --> 01:27:19,969
The user—they're the owner.

1470
01:27:19,969 --> 01:27:20,969
They can do whatever.

1471
01:27:20,969 --> 01:27:22,969
They can destroy the coin so they can do it.

1472
01:27:22,969 --> 01:27:26,969
Only the people who have opted in have this additional—

1473
01:27:26,969 --> 01:27:29,969
The other thing we have to point out is that this same vulnerability—

1474
01:27:29,969 --> 01:27:31,969
you could make the same argument about the Lightning Network.

1475
01:27:31,969 --> 01:27:33,969
You could say this gives miners an extra power.

1476
01:27:33,969 --> 01:27:34,969
Miners can—

1477
01:27:34,969 --> 01:27:36,969
But I would make that argument.

1478
01:27:37,969 --> 01:27:42,969
I think the fact that Lightning creates a system where miners could potentially go
screw over

1479
01:27:42,969 --> 01:27:44,969
a huge number of users is dangerous.

1480
01:27:44,969 --> 01:27:47,969
I would love to see somehow for Lightning to get rid of that.

1481
01:27:47,969 --> 01:27:48,969
And it's also way different because it's local.

1482
01:27:48,969 --> 01:27:50,969
Miners have to collude with a single peer.

1483
01:27:50,969 --> 01:27:52,969
But I think it's much worse for that reason.

1484
01:27:52,969 --> 01:27:53,969
Don't you agree or no?

1485
01:27:53,969 --> 01:27:57,969
In a Drivechain, it's public, so they can just do something very visible.

1486
01:27:57,969 --> 01:28:00,969
In Lightning, they have to collude with one peer to screw up the other.

1487
01:28:00,969 --> 01:28:01,969
So it's way more localized.

1488
01:28:01,969 --> 01:28:04,969
Well, there's a little bit of prep, but the fact that it's localized,

1489
01:28:04,969 --> 01:28:06,969
it's kind of like, I don't know, maybe—

1490
01:28:06,969 --> 01:28:10,969
The fact that it's localized means that it's very difficult for anyone to get

1491
01:28:10,969 --> 01:28:12,969
some maybe what you might call sympathy.

1492
01:28:12,969 --> 01:28:15,969
Like you'd go to Twitter and you could say,

1493
01:28:15,969 --> 01:28:18,969
miners stole all the money out of my Lightning channel.

1494
01:28:18,969 --> 01:28:21,969
But people would be like, I don't know if they're telling the truth.

1495
01:28:21,969 --> 01:28:23,969
So you want something more like—

1496
01:28:23,969 --> 01:28:25,969
Remember the strength in numbers?

1497
01:28:25,969 --> 01:28:29,969
You want something more where it's like if you attack this thing,

1498
01:28:29,969 --> 01:28:31,969
it attacks the whole group.

1499
01:28:31,969 --> 01:28:34,969
In the case of miners screwed over my Lightning channel,

1500
01:28:34,969 --> 01:28:36,969
that's a relatively easy thing to go prove,

1501
01:28:36,969 --> 01:28:38,969
because you can go show that there was a transaction

1502
01:28:38,969 --> 01:28:40,969
that should have achieved some states in the Lightning channel

1503
01:28:40,969 --> 01:28:41,969
that never went through entirely.

1504
01:28:41,969 --> 01:28:44,969
But not necessarily, though, because you can show the transaction,

1505
01:28:44,969 --> 01:28:46,969
but you don't know when it was revealed in the past.

1506
01:28:46,969 --> 01:28:49,969
You have to ironically hash it a different way, timestamp it,

1507
01:28:49,969 --> 01:28:51,969
and then hope that people understand that.

1508
01:28:51,969 --> 01:28:54,969
Realistically, the fact that it showed up in a whole bunch of block explorers

1509
01:28:54,969 --> 01:28:58,969
and everyone could have seen it, and then finally the timeline gets reached

1510
01:28:58,969 --> 01:28:59,969
even though it wasn't mined,

1511
01:28:59,969 --> 01:29:01,969
that's more than enough evidence in practice.

1512
01:29:01,969 --> 01:29:05,969
But I don't know if it would be persuasive to a huge group of people.

1513
01:29:05,969 --> 01:29:08,969
If this happens, you will see it on Twitter in real time,

1514
01:29:08,969 --> 01:29:11,969
and millions of people will know this is happening as it's happening.

1515
01:29:11,969 --> 01:29:14,969
There will be plenty of people seeing this

1516
01:29:14,969 --> 01:29:17,969
and plenty of people angry that, oh shit, Bitcoin just broke.

1517
01:29:17,969 --> 01:29:19,969
This is our Bitcoin conversation.

1518
01:29:19,969 --> 01:29:22,969
We're arguing how to solve these problems over Twitter.

1519
01:29:22,969 --> 01:29:27,969
But I just want to make one observation in terms of privacy good.

1520
01:29:28,969 --> 01:29:32,969
No information being exposed to miners or others on the chain is good.

1521
01:29:32,969 --> 01:29:40,969
But keep in mind the threat is presumably able to compel transparency.

1522
01:29:40,969 --> 01:29:44,969
So if you want to regulate all your white market miners and demand

1523
01:29:44,969 --> 01:29:49,969
that they get KYC information for any transaction that's mined

1524
01:29:49,969 --> 01:29:52,969
or they go to jail, that's not hard to do.

1525
01:29:52,969 --> 01:29:55,969
And that's already kind of being done and will probably continue.

1526
01:29:56,969 --> 01:30:00,969
So you can't really use technology to solve that problem.

1527
01:30:00,969 --> 01:30:02,969
That's a great point.

1528
01:30:02,969 --> 01:30:06,969
You cannot fix blacklists fundamentally because you just turned them into
whitelists.

1529
01:30:06,969 --> 01:30:09,969
Even if you use Monero and you say Monero has ring signatures

1530
01:30:09,969 --> 01:30:11,969
and confidential transactions still,

1531
01:30:11,969 --> 01:30:17,969
somebody could tell you that they're not allowed to accept your Monero transaction

1532
01:30:17,969 --> 01:30:22,969
unless you provide them off-band your membership in the ring signature or whatever.

1533
01:30:22,969 --> 01:30:24,969
I can't agree with that.

1534
01:30:24,969 --> 01:30:26,969
In practice, politics isn't that simple.

1535
01:30:26,969 --> 01:30:30,969
In theory, you can always just ban Bitcoin altogether.

1536
01:30:30,969 --> 01:30:32,969
In practice, politics isn't that simple.

1537
01:30:32,969 --> 01:30:34,969
Politics is a complex thing.

1538
01:30:34,969 --> 01:30:37,969
We're being able to do certain – a good example being,

1539
01:30:37,969 --> 01:30:41,969
in theory, governments would just go to the ASIC manufacturers of the world

1540
01:30:41,969 --> 01:30:44,969
and say, all right, no more Bitcoin mining equipment, end of story.

1541
01:30:44,969 --> 01:30:48,969
We will now do a 50% attack and we'll just kill Bitcoin, that's that.

1542
01:30:48,969 --> 01:30:51,969
Arguments that politics is hard is the status quo.

1543
01:30:51,969 --> 01:30:54,969
If politics could secure Bitcoin, we wouldn't need Bitcoin.

1544
01:30:54,969 --> 01:30:56,969
We'd just use the dollar. We'd vote for what we wanted.

1545
01:30:56,969 --> 01:30:58,969
We'd get what we wanted.

1546
01:30:58,969 --> 01:31:01,969
So Twitter-based arguments and politics is hard,

1547
01:31:01,969 --> 01:31:05,969
these are to me antithetical to Bitcoin.

1548
01:31:05,969 --> 01:31:11,969
When anybody argues Bitcoin is secure because it would be very hard for politicians
to do this,

1549
01:31:11,969 --> 01:31:13,969
people would rise up.

1550
01:31:13,969 --> 01:31:15,969
I go, well, why didn't we rise up against the dollar and the euro?

1551
01:31:15,969 --> 01:31:16,969
We don't.

1552
01:31:16,969 --> 01:31:18,969
So Bitcoin is a way for individuals to –

1553
01:31:18,969 --> 01:31:20,969
But we do. We created Bitcoin.

1554
01:31:20,969 --> 01:31:22,969
We're rising up against the dollar and the euro.

1555
01:31:22,969 --> 01:31:23,969
Exactly.

1556
01:31:23,969 --> 01:31:28,969
And reverting Bitcoin's security model to, it's politically hard so they won't do
it,

1557
01:31:28,969 --> 01:31:30,969
is a break in that design.

1558
01:31:30,969 --> 01:31:34,969
But that is what Bitcoin's security model is at some level.

1559
01:31:34,969 --> 01:31:37,969
Yes, because if you're in North Korea,

1560
01:31:37,969 --> 01:31:40,969
you're not allowed to stand in front of a computer without being arrested.

1561
01:31:40,969 --> 01:31:42,969
There is no cypherpunk ideal that can save you.

1562
01:31:42,969 --> 01:31:46,969
So some level of political evaluation,

1563
01:31:46,969 --> 01:31:50,969
political possibility evaluation is needed even in Bitcoin.

1564
01:31:50,969 --> 01:31:52,969
You can minimize it a lot.

1565
01:31:52,969 --> 01:31:55,969
I would call hiding from the state political.

1566
01:31:55,969 --> 01:31:59,969
In a different world, we wouldn't even have access to general purpose computing.

1567
01:31:59,969 --> 01:32:05,969
It would have been very easy for the evolution of computing to be that you could not
get a compiler.

1568
01:32:05,969 --> 01:32:09,969
Let's say we should not count on the fact that politicians will not attack Bitcoin,

1569
01:32:09,969 --> 01:32:11,969
like my current policy.

1570
01:32:11,969 --> 01:32:15,969
Or that voters won't vote for it because we apparently have voted for the dollar
that we have

1571
01:32:15,969 --> 01:32:16,969
or the euro that we have.

1572
01:32:16,969 --> 01:32:21,969
So there's some smaller group of people out there that don't want to rely on the
political system

1573
01:32:21,969 --> 01:32:23,969
to secure their money because it hasn't worked yet.

1574
01:32:23,969 --> 01:32:27,969
And so Bitcoin is allowing those people to opt out if they have the ability to hide
from it.

1575
01:32:27,969 --> 01:32:28,969
That's it.

1576
01:32:28,969 --> 01:32:31,969
And anytime there's an argument that reverts back to,

1577
01:32:31,969 --> 01:32:35,969
well, we need to vote for this, we need to stop this politically,

1578
01:32:35,969 --> 01:32:37,969
I'm like, why bother with that?

1579
01:32:37,969 --> 01:32:39,969
That is unrelated to Bitcoin.

1580
01:32:39,969 --> 01:32:41,969
This is like Satoshi's unpopular opinion.

1581
01:32:41,969 --> 01:32:44,969
When everybody was hyped about WikiLeaks accepting Bitcoin,

1582
01:32:44,969 --> 01:32:46,969
Satoshi went all unpopular and said,

1583
01:32:46,969 --> 01:32:49,969
no, please don't because you're kicking the hornet's nest, it's too soon.

1584
01:32:49,969 --> 01:32:54,969
So there is some level in which maybe you don't want political acceptance

1585
01:32:54,969 --> 01:32:57,969
as the ultimate security because it's not,

1586
01:32:57,969 --> 01:33:00,969
but you want it maybe temporarily in order to grow enough.

1587
01:33:00,969 --> 01:33:04,969
Take whatever you can get, but don't base your security model on it.

1588
01:33:04,969 --> 01:33:05,969
Yeah, sure.

1589
01:33:05,969 --> 01:33:07,969
Well, I mean, you would like not to base your security model on it,

1590
01:33:07,969 --> 01:33:09,969
but you don't necessarily have a choice.

1591
01:33:09,969 --> 01:33:12,969
And again, we do not have a choice to rely on

1592
01:33:12,969 --> 01:33:14,969
general purpose computing availability

1593
01:33:14,969 --> 01:33:17,969
and also currently with proof of work function ASICs.

1594
01:33:17,969 --> 01:33:19,969
And we have some ways to play games around this,

1595
01:33:19,969 --> 01:33:22,969
like if they went to the intels of the world and said,

1596
01:33:22,969 --> 01:33:23,969
all right, no more mining ASICs,

1597
01:33:23,969 --> 01:33:26,969
we'd probably fall back on the general purpose computing capability.

1598
01:33:26,969 --> 01:33:28,969
And these are the problems we work on.

1599
01:33:28,969 --> 01:33:31,969
In general, the stronger computing technology is

1600
01:33:31,969 --> 01:33:34,969
and the better bandwidth is around the world,

1601
01:33:34,969 --> 01:33:36,969
that's all better for us.

1602
01:33:36,969 --> 01:33:39,969
Gentlemen, we could keep going for hours,

1603
01:33:39,969 --> 01:33:40,969
but this has been fascinating.

1604
01:33:40,969 --> 01:33:41,969
So you know why?

1605
01:33:41,969 --> 01:33:44,969
Because we cannot block each other on stage, only on Twitter.

1606
01:33:46,969 --> 01:33:47,969
I'm blocked.

1607
01:33:47,969 --> 01:33:48,969
All right.

1608
01:33:48,969 --> 01:33:51,969
I mean, I'm sure we will continue this conversation after.

1609
01:33:51,969 --> 01:33:53,969
This is, I think, a rare occurrence.

1610
01:33:53,969 --> 01:33:54,969
I think this is the only time I've been on a panel

1611
01:33:54,969 --> 01:33:56,969
where somebody on the panel hasn't had me blocked.

1612
01:33:58,969 --> 01:34:00,969
I don't have you blocked.

1613
01:34:00,969 --> 01:34:02,969
I'm not blocked by any of you either, I don't think.

1614
01:34:02,969 --> 01:34:04,969
He blocked me and I blocked him back.

1615
01:34:04,969 --> 01:34:05,969
I don't even know why.

1616
01:34:05,969 --> 01:34:07,969
I was going to message you about Peter Todd's article

1617
01:34:07,969 --> 01:34:08,969
and then I was like, he blocked me,

1618
01:34:08,969 --> 01:34:09,969
and then I was like, I'm blocking him.

1619
01:34:09,969 --> 01:34:10,969
That's the second time I've blocked you.

1620
01:34:10,969 --> 01:34:12,969
Are we going to unblock each other after this or no?

1621
01:34:12,969 --> 01:34:13,969
I guess so.

1622
01:34:13,969 --> 01:34:14,969
Maybe, yes.

1623
01:34:14,969 --> 01:34:17,969
I mean, the panels are, we are sweet people at the end.

1624
01:34:17,969 --> 01:34:20,969
It's much better to have these conversations in person.

1625
01:34:20,969 --> 01:34:21,969
Much more.

1626
01:34:21,969 --> 01:34:22,969
I don't know if this was productive at all,

1627
01:34:22,969 --> 01:34:24,969
but it was definitely fun.

1628
01:34:25,969 --> 01:34:27,969
Guys, Max is going to come up here,

1629
01:34:27,969 --> 01:34:28,969
give some closing remarks.

1630
01:34:28,969 --> 01:34:30,969
Thank you for listening to our debate.

1631
01:34:30,969 --> 01:34:31,969
Thank you.

1632
01:34:35,969 --> 01:34:36,969
Thank you.

1633
01:34:49,969 --> 01:34:54,969
Guys, I wanted to thank all of you for being here

1634
01:34:54,969 --> 01:34:57,969
and I hope you enjoyed this year's Honey Badger.

1635
01:34:58,969 --> 01:35:02,969
I want to specifically thank our amazing MCs,

1636
01:35:03,969 --> 01:35:04,969
Bitcoin Stage.

1637
01:35:10,969 --> 01:35:12,969
And also Luna for Sat Stage.

1638
01:35:13,969 --> 01:35:14,969
There he is.

1639
01:35:18,969 --> 01:35:22,969
I want to thank all the speakers for coming here

1640
01:35:22,969 --> 01:35:25,969
and devoting their time, sharing ideas and thoughts.

1641
01:35:25,969 --> 01:35:30,969
I want to thank our security team, catering team,

1642
01:35:30,969 --> 01:35:32,969
video, sound team and venue team.

1643
01:35:32,969 --> 01:35:34,969
And of course, I want to thank hodl hodl team

1644
01:35:34,969 --> 01:35:38,969
and volunteers for doing this amazing weekend for you.

1645
01:35:39,969 --> 01:35:41,969
They deserve that, honestly.

1646
01:35:46,969 --> 01:35:50,969
And I actually would love to thank,

1647
01:35:50,969 --> 01:35:56,969
a huge thank for a person who actually did 99% of the job.

1648
01:35:57,969 --> 01:36:01,969
She hasn't been sleeping for a week, Anna.

1649
01:36:02,969 --> 01:36:07,969
Yeah, she's like, she's somewhere there, but yeah.

1650
01:36:09,969 --> 01:36:14,969
Basically, this year's Honey Badger happened because of her.

1651
01:36:14,969 --> 01:36:17,969
And one more time, please, applause for her.

1652
01:36:17,969 --> 01:36:18,969
Thank you.

1653
01:36:24,969 --> 01:36:25,969
Guys, thank you.

1654
01:36:25,969 --> 01:36:26,969
You've been amazing.

1655
01:36:26,969 --> 01:36:29,969
We did our best to match your expectations.

1656
01:36:29,969 --> 01:36:30,969
We're all Satoshi.

1657
01:36:30,969 --> 01:36:31,969
Thank you.

1658
01:36:47,969 --> 01:36:49,969
Thank you.