1
00:00:00,000 --> 00:00:07,440
Thousands of diligent, informed Bitcoiners lost their entire stack.

2
00:00:08,220 --> 00:00:16,260
Years of savings, recently due to a randomness bug in a popular Bitcoin hardware wallet.

3
00:00:17,280 --> 00:00:20,000
The most careful lost everything.

4
00:00:20,700 --> 00:00:22,820
So does this mean that self-custody is dead?

5
00:00:23,560 --> 00:00:25,640
That don't trust verify is meaningless?

6
00:00:25,640 --> 00:00:30,840
that hold your own keys and Bitcoin only are failed heuristics?

7
00:00:31,340 --> 00:00:33,220
I've been seeing a lot of people say that.

8
00:00:34,320 --> 00:00:36,000
And I'm going to have to disagree.

9
00:00:36,760 --> 00:00:38,740
And I think it's time for a Guy's Take episode.

10
00:00:40,960 --> 00:00:42,960
So if everyone did the right thing,

11
00:00:43,260 --> 00:00:46,560
and they went with a popular, trusted Bitcoin hardware wallet,

12
00:00:46,640 --> 00:00:48,060
Bitcoin only hardware wallet,

13
00:00:48,340 --> 00:00:51,340
and created their own keys, withdrew to self-custody,

14
00:00:51,420 --> 00:00:53,760
and they still lost their coins,

15
00:00:54,500 --> 00:01:00,320
Doesn't that mean that this is all meaningless and you never should have done any of that in the first place?

16
00:01:01,040 --> 00:01:10,920
Yes, in exactly the same way that building your house out of a steel frame and making a stone exterior so it can survive difficult storms,

17
00:01:11,100 --> 00:01:16,660
but it still gets destroyed by a tornado means that you should have just built your house out of cardboard.

18
00:01:23,760 --> 00:01:26,540
Or in other words, yes, if you're an idiot.

19
00:01:27,200 --> 00:01:28,900
These things are lessons.

20
00:01:29,420 --> 00:01:30,840
They are heuristics.

21
00:01:31,100 --> 00:01:35,340
They don't have magical powers to stop any and all bad things from happening.

22
00:01:35,700 --> 00:01:41,220
And importantly, they cannot be fully and completely realized in any sense.

23
00:01:41,300 --> 00:01:46,120
The idea of don't trust verify is about how to think about a problem.

24
00:01:47,240 --> 00:01:52,360
Obviously, you can't do anything without some degree of trust somewhere.

25
00:01:53,000 --> 00:01:57,840
If you ever thought that that's what it would meant, that you could just never trust anything,

26
00:01:58,160 --> 00:02:02,300
then I'd argue that you don't have a realistic opinion about how anything works.

27
00:02:02,980 --> 00:02:06,420
Literally, can you build the cold card personally?

28
00:02:07,180 --> 00:02:10,680
Do you go to the grocery store and buy food ever?

29
00:02:11,280 --> 00:02:14,620
Do you run and understand the entire firmware stack?

30
00:02:14,620 --> 00:02:18,140
Do you understand the computational stack that goes into your Bitcoin node?

31
00:02:18,220 --> 00:02:19,780
Do you not understand how to design a chip?

32
00:02:19,780 --> 00:02:23,280
Do you understand how chip manufacturing and fabrication even works?

33
00:02:23,740 --> 00:02:30,780
Every single thing that goes into the computational stack that runs your Bitcoin node, can you

34
00:02:30,780 --> 00:02:34,620
write and understand the entirety of the Bitcoin software?

35
00:02:35,600 --> 00:02:41,100
Then obviously you have to trust certain things and there's no escaping it.

36
00:02:41,500 --> 00:02:47,760
But this does not mean that the only option is to blindly trust everything and to never

37
00:02:47,760 --> 00:02:54,640
actually attempt to verify those things that we can. We should attempt to trust as little as

38
00:02:54,640 --> 00:03:01,400
possible. We should design things to verify as much as we can. We should take as much responsibility

39
00:03:01,400 --> 00:03:08,580
and control and our own personal education as makes economic sense. As Thomas Sowell said,

40
00:03:08,580 --> 00:03:14,700
there are no solutions. There are only trade-offs. So in the context of coal card and

41
00:03:14,700 --> 00:03:21,180
is self-custody dead. I think the really important thing to understand is that there's no situation

42
00:03:21,180 --> 00:03:26,880
where you own Bitcoin without keys. Somebody is always managing the keys. Someone is always

43
00:03:26,880 --> 00:03:32,500
generating the seed. The only question involved is whether it's you or somebody else.

44
00:03:33,320 --> 00:03:38,720
A service provider or an exchange could get hit with the exact same bug just as easily,

45
00:03:39,220 --> 00:03:44,140
along with a huge list of additional attack vectors and problems that they could have,

46
00:03:44,140 --> 00:03:49,640
because they just let a whole bunch of arbitrary people connect to them and they have to issue withdrawals.

47
00:03:49,760 --> 00:03:56,700
They have to know who and where people are and how to prove that this is actually the user in order to

48
00:03:57,260 --> 00:04:00,900
execute approval for moving Bitcoin for

49
00:04:01,340 --> 00:04:13,458
thousands hundreds of thousands even millions of people The number of attack vectors that they are vulnerable to pales in comparison to the amount that self has to deal with

50
00:04:14,158 --> 00:04:19,998
Now, I understand why the cold card entropy bug hurts as bad as it does.

51
00:04:20,298 --> 00:04:31,378
It is by far the biggest hit to the people who did the right thing and who actually knew largely what they were doing for protecting their Bitcoin keys.

52
00:04:31,498 --> 00:04:39,658
But a huge part of the reason why this event is actually so significant is because it almost never happens.

53
00:04:40,538 --> 00:04:54,518
I think it's worth a little bit of perspective in order to make sure we don't have amnesia about the last 17 years in Bitcoin and the extremely hard-won lessons that we have learned.

54
00:04:55,198 --> 00:04:59,898
Now, I won't be able to go through all of these because there's just too many,

55
00:05:00,058 --> 00:05:03,758
but we will at least hit the highlights so that the lesson is firmly in place

56
00:05:03,758 --> 00:05:09,818
for anybody who may be traumatized into retardation by this cold card event.

57
00:05:10,598 --> 00:05:12,578
Don't worry, you can be saved.

58
00:05:15,938 --> 00:05:22,298
We will start with Mt. Gox in 2011 with a hack of 2,609 Bitcoin,

59
00:05:22,298 --> 00:05:27,518
Then Bitcoinica in 2012 at 43,500 Bitcoin.

60
00:05:28,018 --> 00:05:31,978
Then the Linode hosted wallet service at 46,000 Bitcoin.

61
00:05:32,398 --> 00:05:35,858
Bitflor also in 2012 at 24,000 Bitcoin.

62
00:05:36,318 --> 00:05:40,218
Then Trade Fortress in 2013 at 4,100 Bitcoin.

63
00:05:40,458 --> 00:05:44,418
Then Pico Stocks 2013, 6,000 Bitcoin.

64
00:05:44,778 --> 00:05:49,678
Then Sheep Marketplace in 2013, 96,000 Bitcoin.

65
00:05:49,678 --> 00:06:00,178
Then, of course, the big Mt. Gox, the apex of failures in 2013 at 850,000 Bitcoin.

66
00:06:01,038 --> 00:06:07,198
Then Global Bond, Limited, Flexcoin, Poloniex, MintPal, all 2014 at another 5,000 Bitcoin.

67
00:06:07,318 --> 00:06:13,118
And Bitstamp, 2015, a major and reputable exchange, one of the biggest at the time, 19,000 Bitcoin.

68
00:06:13,458 --> 00:06:17,818
Kipcoin, Beter, Kripsy, all 2015, 23,000 Bitcoin.

69
00:06:17,818 --> 00:06:21,418
Gatecoin, Shapeshift, Bitfinex at 121,000 Bitcoin to that.

70
00:06:21,798 --> 00:06:25,878
UBit, NiceHash, CoinSecure, Zave from AprilChange at another 12,000.

71
00:06:25,998 --> 00:06:31,078
Up to 2019 now, we got CordrigaCX collapsed at 26,000 Bitcoin loss.

72
00:06:31,378 --> 00:06:35,558
Binance in 2019, another major reputable exchange.

73
00:06:35,718 --> 00:06:37,418
Plenty of budget for security problems.

74
00:06:37,858 --> 00:06:39,178
7,000 Bitcoin loss.

75
00:06:39,318 --> 00:06:43,738
Bitpoint, DragonX, AltsBit, Kasia, Qcoin, LiquidGlobal, AfroCrypt,

76
00:06:43,738 --> 00:06:52,358
All 2019 to 2021 zone at about 71,000 Bitcoin, then DMM Bitcoin in 2024 with 4,500 Bitcoin.

77
00:06:52,838 --> 00:07:07,878
And we are at a cool total, just the highlights of this problem, at 1,360,709 Bitcoin lost due to centralized custodians.

78
00:07:08,278 --> 00:07:10,858
But wait, there is more.

79
00:07:11,578 --> 00:07:19,938
We're just talking about the Bitcoin, not the money in custody owed to customers and the investment capital owed back to customers.

80
00:07:19,938 --> 00:07:35,558
With that, we actually have to include FTX at $10 billion, Celsius at $4.7 billion, Voyager Digital at $1.3 billion, BlockFi, Genesis Capital, Gemini, Three Arrows Capital, together about $7 billion.

81
00:07:35,558 --> 00:07:44,618
So we're going to just tack another cool $23 billion on top of that, just in the custody of Bitcoiners money.

82
00:07:45,158 --> 00:07:50,558
And we would actually have to add a lot more to this list if we wanted to be very precise about our comparison.

83
00:07:50,558 --> 00:07:54,558
Stop! Stop! He's already dead!

84
00:07:54,558 --> 00:08:03,238
These were all massive hacks, failures, fraud, or just internal irresponsibility that hit

85
00:08:03,238 --> 00:08:19,056
sometimes thousands and up to even millions of users who trusted the largest exchange the most reputable institution the one with the most money for security audits the one with the most connections

86
00:08:19,436 --> 00:08:23,056
the ones that had all the crypto stuff and worked with everyone

87
00:08:23,056 --> 00:08:26,256
because they were the real solid guys.

88
00:08:26,976 --> 00:08:29,536
This is the record of custodial Bitcoin.

89
00:08:30,296 --> 00:08:35,556
And this is the record of big reputable companies in the Bitcoin space.

90
00:08:36,056 --> 00:08:42,896
This is exactly where the not your keys, not your coins, where don't trust verify, where these things have come from.

91
00:08:43,116 --> 00:08:44,236
It is not arbitrary.

92
00:08:44,756 --> 00:08:48,536
It has been fucking earned a hundred times over.

93
00:08:48,536 --> 00:08:54,676
Now, a lot of people have been trying to track the rough amount stolen from all of our cold cards.

94
00:08:55,376 --> 00:09:01,636
The best estimate right now on the low end is somewhere around 1,778 Bitcoin.

95
00:09:01,636 --> 00:09:06,376
and maybe on the high end of 2,400 Bitcoin.

96
00:09:06,796 --> 00:09:09,776
Despite the significance of this event,

97
00:09:11,256 --> 00:09:15,476
the perspective, the ability to even compare these

98
00:09:15,476 --> 00:09:18,296
is almost nonsensical.

99
00:09:19,076 --> 00:09:22,456
Now, of course, this isn't the only loss of self-custody

100
00:09:22,456 --> 00:09:25,676
as those major hacks and failures of institutions

101
00:09:25,676 --> 00:09:30,096
weren't the only losses of centralized entities either.

102
00:09:30,096 --> 00:09:34,936
So people lose their keys all the time or they drop their hardware wallet off a boat, right?

103
00:09:35,336 --> 00:09:42,536
But this is similar to someone's email getting hacked and then them withdrawing from a custodial exchange all of their balance.

104
00:09:42,956 --> 00:09:57,516
The cold card bug is more akin to the exchange hack in the sense that it is a systemic failure of the mechanism by which thousands or tens of thousands of people were securing their Bitcoin.

105
00:09:57,516 --> 00:10:01,216
or maybe even millions in the context of a custodian.

106
00:10:01,576 --> 00:10:06,236
That even those who seemed to or attempted to follow the rules

107
00:10:06,236 --> 00:10:11,816
still ended up the victim of shoddy, overcomplicated firmware

108
00:10:11,816 --> 00:10:16,816
and the device that they trusted, that we all trusted.

109
00:10:17,936 --> 00:10:19,696
Now, there's a lot to learn from this.

110
00:10:20,416 --> 00:10:25,476
We luckily actually have a lot of tools to look far more seriously at the code base,

111
00:10:25,476 --> 00:10:28,876
even for people who can't write code now, thanks to AI.

112
00:10:29,336 --> 00:10:33,556
And a lot more people, for better or worse, understand what entropy is

113
00:10:33,556 --> 00:10:36,496
and how it plays a role with the creation of their Bitcoin keys.

114
00:10:36,916 --> 00:10:40,796
And unfortunately, almost every major lesson in Bitcoin's history

115
00:10:40,796 --> 00:10:43,416
came alongside some giant disaster.

116
00:10:44,236 --> 00:10:46,516
You know, that's the nature of anti-fragile things.

117
00:10:46,596 --> 00:10:48,236
It's not that bad things don't happen.

118
00:10:48,356 --> 00:10:54,556
It's that when bad things happen, the system learns and it adjusts and it gets stronger.

119
00:10:55,016 --> 00:11:03,136
The thing is, is if we learn lessons from this, it has to be forward looking and it has to work if applied to the past.

120
00:11:03,276 --> 00:11:12,556
If we just kind of like arbitrarily point at some things like, oh, NBK was snotty on Twitter and then think that that rule is going to work.

121
00:11:12,836 --> 00:11:14,336
Apply that to Bitcoin's history.

122
00:11:14,496 --> 00:11:21,256
If you just didn't listen to everybody who was a little bit snotty or arrogant, you wouldn't have had a very good time.

123
00:11:21,876 --> 00:11:25,156
What we need to learn from this, what I want to learn from this,

124
00:11:25,416 --> 00:11:30,916
is something meaningful that can work in the future and works when applied to the past.

125
00:11:31,516 --> 00:11:37,476
One of the few that I think I can pull from this are that source available and open source aren't the same thing.

126
00:11:37,476 --> 00:11:45,596
There's no incentive for people to look at code that isn't open source if they can't actually do anything with it.

127
00:11:45,596 --> 00:11:50,056
Source Available is trying to get, and I did not appreciate this in hindsight,

128
00:11:50,296 --> 00:11:53,716
that Source Available is trying to get the apparent benefit of open source

129
00:11:53,716 --> 00:11:57,076
without actually getting any of the scrutiny that makes it worthwhile.

130
00:11:57,556 --> 00:12:03,656
Other people building on it and needing to verify is exactly what makes open source stronger.

131
00:12:04,176 --> 00:12:05,656
Now, it's not a silver bullet.

132
00:12:05,796 --> 00:12:10,536
Just because you went with open source, there's plenty of open source things that have had detrimental bugs.

133
00:12:10,536 --> 00:12:28,534
But I do think it a meaningful part of how all this played out and why this failure wasn called Another one independent audits We can just say don trust verify slogans or slogans and they don mean jack if you don actually apply them

134
00:12:28,934 --> 00:12:34,414
And for those who can't actually check the firmware, we need independent audits. We need

135
00:12:34,414 --> 00:12:41,914
people who are contentious, checking the work of other people. We want foundation arguing with cold

136
00:12:41,914 --> 00:12:49,374
card and pointing out bugs on purpose. We want Bitbox and Ledger fighting each other. I think we

137
00:12:49,374 --> 00:12:55,894
need more disagreement or more cooperation for the exact same reason. Doesn't really matter how

138
00:12:55,894 --> 00:13:02,514
we get it. It's great. It's nice to be nice, but I don't really care. Hate each other. If it means

139
00:13:02,514 --> 00:13:06,594
that you're actually going to check and we're going to get independent audits and we're going

140
00:13:06,594 --> 00:13:12,354
to get people aggressively pushing against each other's software, that's great news to me.

141
00:13:12,854 --> 00:13:17,454
And then the third thing that as much as people say it's ridiculous, and I understand,

142
00:13:17,774 --> 00:13:23,774
part of me thinks this too, I think we should walk away with this thinking roll your own

143
00:13:23,774 --> 00:13:26,334
randomness as another slogan in Bitcoin.

144
00:13:26,754 --> 00:13:28,914
We need to understand what entropy is.

145
00:13:28,914 --> 00:13:32,274
Entropy is at the heart of everything that happens here.

146
00:13:32,274 --> 00:13:58,474
And the fact that we did just kind of glaze over it is a failure on us as educators, as people who are trying to point out what is important here. And honestly, just adding that one concept to the stack makes a massive difference when you apply it to the past and you apply it to this cold card situation. And it ain't a hard thing to understand.

147
00:13:58,474 --> 00:14:00,254
And that's not too difficult.

148
00:14:00,474 --> 00:14:07,854
That's not asking too much to just recognize what randomness is and how it is being applied

149
00:14:07,854 --> 00:14:09,834
to the creation of your Bitcoin keys.

150
00:14:09,954 --> 00:14:14,074
If we're going to worry so much about holding your keys, we should at least talk a little

151
00:14:14,074 --> 00:14:16,434
bit about what it means to create keys.

152
00:14:16,954 --> 00:14:22,294
Right now, those are the only three really good ones that I think I can pull from this.

153
00:14:22,834 --> 00:14:28,454
But if we just ignore everything that we've learned so far in Bitcoin, if we're just

154
00:14:28,454 --> 00:14:33,794
going to go with the big company because they have more resources? If that ever works, why is

155
00:14:33,794 --> 00:14:40,654
Bitcoin history just riddled with failures of the absolute biggest institutions? Why are all the

156
00:14:40,654 --> 00:14:46,834
crypto people who have the most money to do the most security audits and checks, why are they the

157
00:14:46,834 --> 00:14:54,334
ones with the most spectacular failures? If self-custody is somehow dead, then why is it by

158
00:14:54,334 --> 00:15:01,714
multiple orders of magnitude still the safest way to own your Bitcoin? I think the answer is quite

159
00:15:01,714 --> 00:15:08,794
simple. Those are the wrong lessons to learn. Now I'll be doing a longer and more in-depth Q&A on

160
00:15:08,794 --> 00:15:12,894
the cold card situation. We'll probably be recording it tomorrow morning, so hopefully it'll be out

161
00:15:12,894 --> 00:15:17,614
soon. But I just wanted to knock this one out first because this one was just getting to me.

162
00:15:18,254 --> 00:15:23,574
If you have any questions though, let me know. Check out the Bitbox hardware wallet. Obviously,

163
00:15:23,574 --> 00:15:27,694
this is fully open source. I've also been keeping a very close eye on the red teaming,

164
00:15:27,994 --> 00:15:31,674
not only against all the other hardware wallets, but Bitbox more specifically,

165
00:15:31,674 --> 00:15:38,314
since I am promoting them and I use them quite a bit. So if there are any uncovered vulnerabilities

166
00:15:38,314 --> 00:15:44,294
or anything worth reporting or sharing with you, I will share it here on the show as fast as I can.

167
00:15:45,274 --> 00:15:51,534
But so far, everything looks really good. There is some interesting little like cork edge cases

168
00:15:51,534 --> 00:15:53,374
that I think I might actually be fun to dig into

169
00:15:53,374 --> 00:15:55,374
and we might get at it on the show.

170
00:15:56,614 --> 00:15:58,254
But stay tuned for that.

171
00:15:58,334 --> 00:16:01,734
Don't forget to subscribe to the Guys Take channel on YouTube.

172
00:16:01,934 --> 00:16:02,854
That is our new channel.

173
00:16:03,174 --> 00:16:05,874
So if you're seeing this, please, it's a huge help

174
00:16:05,874 --> 00:16:08,614
because we're still trying to bootstrap that network over there.

175
00:16:09,114 --> 00:16:12,234
And I will catch you guys on the next episode.

176
00:16:12,234 --> 00:16:15,474
Until then, everybody, that's my two sats.

177
00:16:21,534 --> 00:16:22,034
you