Breaking down the Coldcard exploit & ramifications for self-custody

1
00:00:06,880 --> 00:00:08,479
Welcome back to The Starting Block.

2
00:00:08,480 --> 00:00:10,199
Happy Monday to you wherever you are

3
00:00:10,200 --> 00:00:11,200
around the world.

4
00:00:12,180 --> 00:00:13,439
It's week number five, I think we're

5
00:00:13,440 --> 00:00:15,139
number six. No, I we're week number

6
00:00:15,140 --> 00:00:17,059
six, so we're a

7
00:00:17,060 --> 00:00:18,719
good six weeks into the start of the

8
00:00:18,720 --> 00:00:20,899
show. It's been an absolute pleasure

9
00:00:20,900 --> 00:00:22,919
broadcasting every day.

10
00:00:22,920 --> 00:00:24,479
This is The Block's new daily show,

11
00:00:24,480 --> 00:00:25,439
and if you're joining us for the

12
00:00:25,440 --> 00:00:26,399
first time, my name is Gareth

13
00:00:26,400 --> 00:00:28,259
Jenkinson. I'm your host and head of

14
00:00:28,260 --> 00:00:30,519
multimedia here at The Block.

15
00:00:30,520 --> 00:00:32,039
If you wanna catch up on some of the

16
00:00:32,040 --> 00:00:33,339
content we've been putting out over

17
00:00:33,340 --> 00:00:34,699
the last few weeks, you can go and

18
00:00:34,700 --> 00:00:36,709
check out The Block YouTube channel.

19
00:00:36,710 --> 00:00:38,649
Give us a subscribe

20
00:00:38,650 --> 00:00:40,589
if you're there and watching us

21
00:00:40,590 --> 00:00:42,389
today. And if you watching us live

22
00:00:42,390 --> 00:00:44,249
on X this afternoon or this

23
00:00:44,250 --> 00:00:45,209
morning, if you in the United

24
00:00:45,210 --> 00:00:46,429
States, please like and read

25
00:00:46,430 --> 00:00:47,809
broadcast so we can get as many

26
00:00:47,810 --> 00:00:49,549
people joining us today for what

27
00:00:49,550 --> 00:00:51,389
promises to be a jam-packed

28
00:00:51,390 --> 00:00:52,849
show. As you can see by the name of

29
00:00:52,850 --> 00:00:54,669
the show, we are gonna be dedicating

30
00:00:54,670 --> 00:00:56,649
a lot of time today to talk about

31
00:00:56,650 --> 00:00:58,169
the cold card exploit.

32
00:00:58,170 --> 00:00:59,909
And if don't know what that is, I

33
00:00:59,910 --> 00:01:01,809
will give you a full rundown once

34
00:01:01,810 --> 00:01:03,689
we have our guests on the show

35
00:01:03,690 --> 00:01:05,149
a little bit later today.

36
00:01:05,150 --> 00:01:07,049
If you are here for the first time,

37
00:01:07,050 --> 00:01:08,169
the show is very simple.

38
00:01:08,170 --> 00:01:09,609
We start off with a news bulletin

39
00:01:09,610 --> 00:01:11,189
that's called the Block Briefing.

40
00:01:11,190 --> 00:01:13,329
We'll run down six or

41
00:01:13,330 --> 00:01:15,629
seven of the top stories

42
00:01:15,630 --> 00:01:17,369
in the news over the last 24 hours.

43
00:01:17,370 --> 00:01:18,449
Obviously it's a Monday, so we'll

44
00:01:18,450 --> 00:01:19,489
look at some of the stuff from the

45
00:01:19,490 --> 00:01:21,509
weekend and then we'll jump into

46
00:01:21,510 --> 00:01:22,329
the hot seat today.

47
00:01:22,330 --> 00:01:24,009
We won't have a market pulse because

48
00:01:24,010 --> 00:01:25,689
we're gonna be deep diving into that

49
00:01:25,690 --> 00:01:27,669
cold exploit and what it means for

50
00:01:27,670 --> 00:01:29,529
Bitcoin self-custody going

51
00:01:29,530 --> 00:01:31,029
forward. Really, really excited to

52
00:01:31,030 --> 00:01:32,609
have Zach Herbert from Foundation

53
00:01:32,610 --> 00:01:34,678
and Jameson Lopp.

54
00:01:34,679 --> 00:01:36,579
From Casa to talk

55
00:01:36,580 --> 00:01:38,659
about Bitcoin self-custody and

56
00:01:38,660 --> 00:01:40,059
the cold card exploit.

57
00:01:40,060 --> 00:01:41,959
But before we do any of that,

58
00:01:41,960 --> 00:01:43,659
I'm gonna give you a rundown of the

59
00:01:43,660 --> 00:01:46,099
top stories you need to know this

60
00:01:46,100 --> 00:01:47,399
Monday.

61
00:01:47,400 --> 00:01:48,859
Starting off with the big headline,

62
00:01:48,860 --> 00:01:50,819
and this news dropped less

63
00:01:50,820 --> 00:01:53,039
than 15 minutes ago,

64
00:01:53,040 --> 00:01:55,159
Michael Saylor's strategy continues

65
00:01:55,160 --> 00:01:56,939
to sell Bitcoin.

66
00:01:56,940 --> 00:01:58,939
A fresh SEC filing shows the

67
00:01:58,940 --> 00:02:01,999
firm offloaded 1,638

68
00:02:02,000 --> 00:02:03,419
Bitcoin last week.

69
00:02:03,420 --> 00:02:06,099
For around $105 million at

70
00:02:06,100 --> 00:02:08,420
an average price near $64,000.

71
00:02:09,460 --> 00:02:11,399
That trims its total stack to

72
00:02:11,400 --> 00:02:13,339
roughly 842,000

73
00:02:13,340 --> 00:02:15,819
Bitcoin, still around 4%

74
00:02:15,820 --> 00:02:17,599
of all the Bitcoin that will ever

75
00:02:17,600 --> 00:02:19,539
exist, but now carrying close

76
00:02:19,540 --> 00:02:21,559
to an $11 billion paper

77
00:02:21,560 --> 00:02:23,999
loss at current prices.

78
00:02:24,000 --> 00:02:25,879
My producer Jordan was teasing

79
00:02:25,880 --> 00:02:27,339
us all fair a little bit earlier

80
00:02:27,340 --> 00:02:29,359
once he saw this news come out, but

81
00:02:29,360 --> 00:02:31,379
just to put this in perspective,

82
00:02:31,380 --> 00:02:32,729
this is about 0.

83
00:02:32,730 --> 00:02:34,709
1% of their total Bitcoin

84
00:02:34,710 --> 00:02:36,889
stack that they've sold and

85
00:02:36,890 --> 00:02:38,309
strategy says the proceeds went

86
00:02:38,310 --> 00:02:40,229
towards preferred preferred

87
00:02:40,230 --> 00:02:42,149
stock Distributions and buying

88
00:02:42,150 --> 00:02:44,549
back its STR shares a

89
00:02:44,550 --> 00:02:46,229
clear shift away from sailors long

90
00:02:46,230 --> 00:02:48,129
running the buy only Playbook

91
00:02:48,130 --> 00:02:49,989
and if you go and look at X

92
00:02:49,990 --> 00:02:51,829
and strategies announcement, there's

93
00:02:51,830 --> 00:02:53,229
plenty of negative comments there

94
00:02:53,230 --> 00:02:55,369
But I am seeing a few Bitcoin

95
00:02:55,370 --> 00:02:56,749
is that I follow quite closely

96
00:02:56,750 --> 00:02:58,909
giving a dose of perspective here

97
00:02:58,910 --> 00:03:00,919
and kind of calling

98
00:03:00,920 --> 00:03:02,959
out strategies,

99
00:03:02,960 --> 00:03:05,039
building up of their credit standing

100
00:03:05,040 --> 00:03:07,059
as they build up a much

101
00:03:07,060 --> 00:03:09,079
bigger US dollar reserve

102
00:03:09,080 --> 00:03:11,039
to give themselves more than

103
00:03:11,040 --> 00:03:13,039
a two year runway to meet

104
00:03:13,040 --> 00:03:14,919
dividend obligations.

105
00:03:14,920 --> 00:03:16,180
Moving on to our next story,

106
00:03:17,320 --> 00:03:19,439
a big on-chain move raised eyebrows

107
00:03:19,440 --> 00:03:21,359
while it's tied to Trump Media, the

108
00:03:21,360 --> 00:03:23,459
company behind True Social shifted

109
00:03:23,460 --> 00:03:25,899
2,628 Bitcoin

110
00:03:25,900 --> 00:03:27,819
with around $165 million

111
00:03:27,820 --> 00:03:30,449
to crypto.com over the weekend.

112
00:03:30,450 --> 00:03:32,649
Spread across two transactions,

113
00:03:32,650 --> 00:03:34,549
but the company says don't read

114
00:03:34,550 --> 00:03:35,909
too much into it.

115
00:03:35,910 --> 00:03:37,729
A Trump media spokesperson told the

116
00:03:37,730 --> 00:03:39,509
block the Bitcoin was transferred

117
00:03:39,510 --> 00:03:41,629
and not sold, the same explanation

118
00:03:41,630 --> 00:03:43,429
they gave back in May.

119
00:03:43,430 --> 00:03:45,269
The tag wallet still hold roughly

120
00:03:45,270 --> 00:03:47,889
4,261 Bitcoin,

121
00:03:47,890 --> 00:03:50,229
about $260 million worth

122
00:03:50,230 --> 00:03:51,949
closely matching the collateral

123
00:03:51,950 --> 00:03:54,289
backing its convertible notes as

124
00:03:54,290 --> 00:03:56,509
of the first quarter filing.

125
00:03:56,510 --> 00:03:58,609
Over in Washington, a regulatory

126
00:03:58,610 --> 00:04:00,419
turf war is heating up.

127
00:04:00,420 --> 00:04:02,419
The SEC has hit pause on

128
00:04:02,420 --> 00:04:04,659
NASDAQ's PHLX's

129
00:04:04,660 --> 00:04:06,679
plans to list cash settled Bitcoin

130
00:04:06,680 --> 00:04:08,919
index options after CME

131
00:04:08,920 --> 00:04:10,619
group petitioned the commissioners

132
00:04:10,620 --> 00:04:13,139
to overturn a staff approval.

133
00:04:13,140 --> 00:04:15,179
CME's argument here, these

134
00:04:15,180 --> 00:04:17,379
contracts are commodity option swaps

135
00:04:17,380 --> 00:04:19,179
that belong squarely under the

136
00:04:19,180 --> 00:04:21,259
CFTC's oversight,

137
00:04:21,260 --> 00:04:22,779
not the SEC.

138
00:04:22,780 --> 00:04:24,639
The commission granted CMEs petition

139
00:04:24,640 --> 00:04:26,519
for review and kept the

140
00:04:26,520 --> 00:04:27,529
approval state.

141
00:04:27,530 --> 00:04:29,309
Written statements are due on the

142
00:04:29,310 --> 00:04:31,249
24th of August, later this month.

143
00:04:31,250 --> 00:04:33,169
It's a jurisdictional

144
00:04:33,170 --> 00:04:35,189
tug of war that could help shape

145
00:04:35,190 --> 00:04:37,369
who regulates the next wave of

146
00:04:37,370 --> 00:04:39,369
crypto derivative products in

147
00:04:39,370 --> 00:04:41,329
the United States.

148
00:04:41,330 --> 00:04:42,889
Then a sobering story out of Hong

149
00:04:42,890 --> 00:04:44,949
Kong. Police say a local insurance

150
00:04:44,950 --> 00:04:46,949
agent lost more than $3.3 million

151
00:04:46,950 --> 00:04:48,789
after being lured into

152
00:04:48,790 --> 00:04:50,389
a fake cryptocurrency investment

153
00:04:50,390 --> 00:04:52,409
scheme through a romance scam.

154
00:04:52,410 --> 00:04:53,819
It wasn't isolated.

155
00:04:53,820 --> 00:04:55,479
It was the single largest case among

156
00:04:55,480 --> 00:04:57,399
25 romance scams reported

157
00:04:57,400 --> 00:04:59,679
in just one week, from July 24th

158
00:04:59,680 --> 00:05:01,599
to the 30th, adding up to nearly

159
00:05:01,600 --> 00:05:04,119
$9 million in total losses.

160
00:05:04,120 --> 00:05:05,639
It's a stark reminder as crypto goes

161
00:05:05,640 --> 00:05:07,759
mainstream, so do the scams

162
00:05:07,760 --> 00:05:10,379
targeting everyday people.

163
00:05:10,380 --> 00:05:12,259
Then prediction markets continue on

164
00:05:12,260 --> 00:05:13,159
a tear.

165
00:05:13,160 --> 00:05:14,319
Combined trading volume across

166
00:05:14,320 --> 00:05:16,119
Polymarket, Polymarket U.S.

167
00:05:16,120 --> 00:05:18,779
And Kalshi hit a record.

168
00:05:18,780 --> 00:05:20,619
$50 billion in volume

169
00:05:20,620 --> 00:05:23,039
was up 7.7%.

170
00:05:23,040 --> 00:05:24,999
From June, according to the block's

171
00:05:25,000 --> 00:05:26,219
data dashboard.

172
00:05:26,220 --> 00:05:27,719
Continued momentum around the World

173
00:05:27,720 --> 00:05:29,539
Cup helped drive that, but there's

174
00:05:29,540 --> 00:05:30,859
a twist under the hood.

175
00:05:30,860 --> 00:05:32,879
Polymarket US's volume surged

176
00:05:32,880 --> 00:05:34,719
54 percent, while the main

177
00:05:34,720 --> 00:05:36,359
polymarket platform's volume

178
00:05:36,360 --> 00:05:38,239
actually fell 26 percent

179
00:05:38,240 --> 00:05:40,039
compared to June, a sign that

180
00:05:40,040 --> 00:05:41,879
regulated US-facing prediction

181
00:05:41,880 --> 00:05:43,699
markets are pulling in more of

182
00:05:43,700 --> 00:05:45,719
the action as the sector keeps

183
00:05:45,720 --> 00:05:46,759
gaining steam.

184
00:05:46,760 --> 00:05:48,999
And finally, South Korea,

185
00:05:49,000 --> 00:05:50,839
but the country's second largest

186
00:05:50,840 --> 00:05:52,039
crypto exchange.

187
00:05:52,040 --> 00:05:53,699
Has officially confirmed it's

188
00:05:53,700 --> 00:05:55,639
targeting a public listing in

189
00:05:55,640 --> 00:05:57,039
2028.

190
00:05:57,040 --> 00:05:58,919
In a notice posted on its

191
00:05:58,920 --> 00:06:00,759
website, the exchange laid out a

192
00:06:00,760 --> 00:06:02,519
multi-year preparation plan,

193
00:06:02,520 --> 00:06:03,679
including a request for a

194
00:06:03,680 --> 00:06:06,279
preliminary listing review in 2027

195
00:06:06,280 --> 00:06:08,619
before pushing for the IPO the

196
00:06:08,620 --> 00:06:09,499
following year.

197
00:06:09,500 --> 00:06:11,019
It's a notable step for a major

198
00:06:11,020 --> 00:06:13,159
Asian exchange moving towards public

199
00:06:13,160 --> 00:06:14,979
markets, and another sign that

200
00:06:14,980 --> 00:06:16,839
South Korea's crypto industry is

201
00:06:16,840 --> 00:06:18,299
maturing fast.

202
00:06:18,300 --> 00:06:19,499
We'll be watching how that roadmap

203
00:06:19,500 --> 00:06:20,459
unfolds.

204
00:06:20,460 --> 00:06:22,679
Over the next couple of

205
00:06:22,680 --> 00:06:24,219
years. And that's it for the block

206
00:06:24,220 --> 00:06:25,159
briefing.

207
00:06:25,160 --> 00:06:26,539
I think we are gonna take a very

208
00:06:26,540 --> 00:06:28,499
short break before we bring Zach

209
00:06:28,500 --> 00:06:30,279
Herbert on and Jameson Locke a

210
00:06:30,280 --> 00:06:32,079
little bit later on the show.

211
00:06:32,080 --> 00:06:34,179
Stay with us as we talk about the

212
00:06:34,180 --> 00:06:35,939
cold card exploit and everything you

213
00:06:35,940 --> 00:06:37,819
need to know about what it means

214
00:06:37,820 --> 00:06:39,759
for Bitcoin self custody going

215
00:06:39,760 --> 00:06:40,799
forward. Don't go away.

216
00:06:40,800 --> 00:06:41,800
We'll be right back.

217
00:09:35,740 --> 00:09:37,099
Welcome back to the starting block.

218
00:09:37,100 --> 00:09:38,179
We're jumping into the hot seat and

219
00:09:38,180 --> 00:09:39,779
we're gonna dedicate two interviews

220
00:09:39,780 --> 00:09:42,599
today to the hotseat segment.

221
00:09:42,600 --> 00:09:43,719
It's a great pleasure to have Zach

222
00:09:43,720 --> 00:09:45,599
Herbert, who's the co-founder and

223
00:09:45,600 --> 00:09:47,679
CEO of Foundation, joining

224
00:09:47,680 --> 00:09:49,399
us today to talk about the Cold Card

225
00:09:49,400 --> 00:09:50,379
exploit.

226
00:09:50,380 --> 00:09:51,559
And if you're joining us for the

227
00:09:51,560 --> 00:09:52,759
first time and you don't know about

228
00:09:52,760 --> 00:09:54,799
the cold card exploit, Cold Card is

229
00:09:54,800 --> 00:09:57,019
of course a Bitcoin-only

230
00:09:57,020 --> 00:09:58,939
hardware manufacturer that

231
00:09:58,940 --> 00:10:00,859
was hit last week with an

232
00:10:00,860 --> 00:10:02,639
exploit that...

233
00:10:02,640 --> 00:10:04,259
As far as we would know right now,

234
00:10:04,260 --> 00:10:06,859
could impact all of their devices.

235
00:10:06,860 --> 00:10:08,699
And we've seen more

236
00:10:08,700 --> 00:10:10,699
than a thousand Bitcoin drained

237
00:10:10,700 --> 00:10:12,579
from certain models

238
00:10:12,580 --> 00:10:14,519
of cold card wallets

239
00:10:14,520 --> 00:10:16,359
in the past 36

240
00:10:16,360 --> 00:10:18,379
hours or more.

241
00:10:18,380 --> 00:10:20,139
It's been a concerted effort across

242
00:10:20,140 --> 00:10:21,959
the Bitcoin ecosystem to,

243
00:10:21,960 --> 00:10:23,599
first of all, make sure that

244
00:10:23,600 --> 00:10:24,719
everyone knows that this is

245
00:10:24,720 --> 00:10:26,459
happening, especially for cold card

246
00:10:26,460 --> 00:10:28,219
users that have not been online the

247
00:10:28,220 --> 00:10:29,159
last few days.

248
00:10:29,160 --> 00:10:30,379
They might not even know, and I've

249
00:10:30,380 --> 00:10:32,039
seen some harrowing stories as well

250
00:10:32,040 --> 00:10:33,799
that. Some people went away on

251
00:10:33,800 --> 00:10:35,479
holiday for a couple of days, saw

252
00:10:35,480 --> 00:10:37,079
the news, they weren't able to get

253
00:10:37,080 --> 00:10:38,339
to their devices. And when they got

254
00:10:38,340 --> 00:10:39,919
home to their device or got someone

255
00:10:39,920 --> 00:10:42,519
to actually open

256
00:10:42,520 --> 00:10:44,599
up the devices and check their

257
00:10:44,600 --> 00:10:46,499
balances, Bitcoin had

258
00:10:46,500 --> 00:10:48,119
been drained. So a really dire

259
00:10:48,120 --> 00:10:49,479
situation.

260
00:10:49,480 --> 00:10:50,659
And we're gonna be talking about

261
00:10:50,660 --> 00:10:52,639
that at length today with

262
00:10:52,640 --> 00:10:54,139
Zach and Jameson Lopp, who's joining

263
00:10:54,140 --> 00:10:55,379
us a little bit later on the show.

264
00:10:55,380 --> 00:10:56,539
Zach, first things first, thank you

265
00:10:56,540 --> 00:10:58,159
so much for taking the time to chat

266
00:10:58,160 --> 00:10:58,999
to us today.

267
00:10:59,000 --> 00:11:01,219
It's the first time we've spoken in

268
00:11:01,220 --> 00:11:02,469
person, virtually.

269
00:11:02,470 --> 00:11:04,529
So it's good to meet you.

270
00:11:04,530 --> 00:11:05,749
I can imagine it's been a rough few

271
00:11:05,750 --> 00:11:06,759
days, how are you doing?

272
00:11:06,760 --> 00:11:08,699
I'm doing okay a

273
00:11:08,700 --> 00:11:10,619
little tired and some of our

274
00:11:10,620 --> 00:11:12,839
team is a little tire as well

275
00:11:12,840 --> 00:11:14,679
But I mean compared to those folks

276
00:11:14,680 --> 00:11:16,879
who lost money right and it's just

277
00:11:16,880 --> 00:11:18,359
yeah and comparable and it has been

278
00:11:18,360 --> 00:11:20,619
a really difficult day for

279
00:11:20,620 --> 00:11:21,620
Bitcoin

280
00:11:22,600 --> 00:11:24,619
Bitcoiners and Most hardcore

281
00:11:24,620 --> 00:11:26,419
Bitcoiners right that I've been told

282
00:11:26,420 --> 00:11:27,739
self-custody and that have helped

283
00:11:27,740 --> 00:11:29,299
other self- custody So it's

284
00:11:29,300 --> 00:11:31,359
definitely everyone's reeling from

285
00:11:31,360 --> 00:11:33,259
the events of the last few days.

286
00:11:33,260 --> 00:11:35,339
I don't think it's gonna

287
00:11:35,340 --> 00:11:37,299
uh, stop either anytime soon,

288
00:11:37,300 --> 00:11:38,439
we're going to be feeling the

289
00:11:38,440 --> 00:11:40,399
effects of this potentially for

290
00:11:40,400 --> 00:11:42,659
months, maybe even years.

291
00:11:42,660 --> 00:11:44,179
And so, um, yeah, thank you for

292
00:11:44,180 --> 00:11:46,019
having me on to, uh, talk about it.

293
00:11:47,330 --> 00:11:49,469
Yeah, sure. And I think just for

294
00:11:49,470 --> 00:11:50,949
our viewers out there that actually

295
00:11:50,950 --> 00:11:52,889
might not be too familiar

296
00:11:52,890 --> 00:11:54,189
with the story, maybe you can give

297
00:11:54,190 --> 00:11:56,229
us your kind of TLDR

298
00:11:56,230 --> 00:11:57,769
of what happened here,

299
00:11:59,490 --> 00:12:01,049
a very high level overview before

300
00:12:01,050 --> 00:12:02,389
perhaps we get into some of the

301
00:12:02,390 --> 00:12:03,449
specifics.

302
00:12:03,450 --> 00:12:05,269
For sure. And I'll try to keep it

303
00:12:05,270 --> 00:12:06,329
brief so that you can ask me

304
00:12:06,330 --> 00:12:07,849
questions, you know, because I don't

305
00:12:07,850 --> 00:12:09,249
know how to in the weeds, right?

306
00:12:09,250 --> 00:12:10,529
We want to get here.

307
00:12:10,530 --> 00:12:12,349
But essentially a cold card,

308
00:12:12,350 --> 00:12:14,509
right, is a, is a popular Bitcoin

309
00:12:14,510 --> 00:12:16,309
only hardware wallet made by a

310
00:12:16,310 --> 00:12:17,909
company called coin kite.

311
00:12:17,910 --> 00:12:19,869
That's MDK or Rodolfo Novak's

312
00:12:19,870 --> 00:12:21,389
company. You know, he's a pretty

313
00:12:21,390 --> 00:12:23,269
huge name in the, in the

314
00:12:23,270 --> 00:12:24,569
Bitcoin space.

315
00:12:24,570 --> 00:12:26,909
And on Thursday afternoon reports

316
00:12:26,910 --> 00:12:28,449
start to come out that funds were

317
00:12:28,450 --> 00:12:29,549
being stolen.

318
00:12:29,550 --> 00:12:31,829
It seemed like from folks

319
00:12:31,830 --> 00:12:33,549
with cold card devices.

320
00:12:33,550 --> 00:12:35,469
And so you can imagine the

321
00:12:35,470 --> 00:12:37,349
entire industry scrambled to

322
00:12:37,350 --> 00:12:38,829
figure out what was happening, what

323
00:12:38,830 --> 00:12:39,929
was going on.

324
00:12:39,930 --> 00:12:41,929
And the root cause that was

325
00:12:41,930 --> 00:12:43,609
first disclosed, I think probably

326
00:12:43,610 --> 00:12:45,429
blocked it, the best job, getting

327
00:12:45,430 --> 00:12:47,209
on top of this and disclosing it

328
00:12:47,210 --> 00:12:48,889
fast and writing up a report.

329
00:12:48,890 --> 00:12:50,069
So many of us were kind of figuring

330
00:12:50,070 --> 00:12:52,249
out in real time and

331
00:12:52,250 --> 00:12:53,989
obviously checking that for all the

332
00:12:53,990 --> 00:12:55,029
other hardware wallet companies,

333
00:12:55,030 --> 00:12:56,329
checking that our customers were

334
00:12:56,330 --> 00:12:57,989
also not affected.

335
00:12:57,990 --> 00:12:59,729
But there was a bug introduced many

336
00:12:59,730 --> 00:13:01,589
years ago into the cold card code

337
00:13:01,590 --> 00:13:04,389
base in early 2021

338
00:13:04,390 --> 00:13:07,189
that effectively reduced

339
00:13:07,190 --> 00:13:09,049
by a significant factor

340
00:13:09,050 --> 00:13:11,289
the amount of randomness

341
00:13:11,290 --> 00:13:13,509
or we call entropy that goes into

342
00:13:13,510 --> 00:13:15,829
the seed generation process.

343
00:13:15,830 --> 00:13:18,529
And so what that meant is that

344
00:13:18,530 --> 00:13:20,509
cold card seeds starting from early

345
00:13:20,510 --> 00:13:22,589
2021 and even basically

346
00:13:22,590 --> 00:13:24,699
through today or I should say maybe

347
00:13:24,700 --> 00:13:25,820
through over the weekend, because

348
00:13:25,821 --> 00:13:27,699
they released an update a

349
00:13:27,700 --> 00:13:28,740
day or two ago,

350
00:13:29,900 --> 00:13:31,899
which you could install and would

351
00:13:31,900 --> 00:13:33,939
fix that ongoing on new devices.

352
00:13:33,940 --> 00:13:36,279
But from 2021, basically,

353
00:13:36,280 --> 00:13:37,280
until now,

354
00:13:38,600 --> 00:13:40,239
all seeds generated on cold card

355
00:13:40,240 --> 00:13:42,419
devices are insecure.

356
00:13:42,420 --> 00:13:44,829
Now the level of

357
00:13:46,590 --> 00:13:48,669
the security for the seed varies by

358
00:13:48,670 --> 00:13:49,709
model, right?

359
00:13:49,710 --> 00:13:51,789
So in this case, the mark three

360
00:13:51,790 --> 00:13:54,449
model, which was the model that was

361
00:13:54,450 --> 00:13:56,329
introduced around that time

362
00:13:56,330 --> 00:13:57,729
a little bit earlier is the one

363
00:13:57,730 --> 00:13:58,769
that's most affected.

364
00:13:58,770 --> 00:14:00,789
And the thought is that

365
00:14:00,790 --> 00:14:02,809
most of the funds have already

366
00:14:02,810 --> 00:14:04,769
been taken from

367
00:14:04,770 --> 00:14:06,529
those wallets and devices.

368
00:14:06,530 --> 00:14:08,389
When we say taken, we don't

369
00:14:08,390 --> 00:14:09,829
actually mean that you need any kind

370
00:14:09,830 --> 00:14:11,249
of physical access to the hardware

371
00:14:11,250 --> 00:14:12,229
wallet, right. And so that's what

372
00:14:12,230 --> 00:14:14,349
makes this really difficult,

373
00:14:14,350 --> 00:14:15,529
really acute.

374
00:14:15,530 --> 00:14:17,609
It means that an attacker can

375
00:14:17,610 --> 00:14:19,529
sit down at the computer, they can

376
00:14:19,530 --> 00:14:21,609
write a script that can run on,

377
00:14:21,610 --> 00:14:22,949
very often it could run on your

378
00:14:22,950 --> 00:14:25,389
laptop, it could run on Nvidia GPUs,

379
00:14:25,390 --> 00:14:26,749
they could be renting, you know,

380
00:14:26,750 --> 00:14:28,369
compute somewhere, and they're just

381
00:14:28,370 --> 00:14:29,769
cracking seeds.

382
00:14:29,770 --> 00:14:31,209
And so they're going through seed by

383
00:14:31,210 --> 00:14:32,749
seed, they're generating potential

384
00:14:32,750 --> 00:14:34,609
seeds, and then they have scripts to

385
00:14:34,610 --> 00:14:36,449
check, is there is

386
00:14:36,450 --> 00:14:38,469
there money at this Bitcoin address,

387
00:14:38,470 --> 00:14:40,129
right, that's that they're able to

388
00:14:40,130 --> 00:14:41,869
derive that based on the seed.

389
00:14:41,870 --> 00:14:42,779
And so.

390
00:14:42,780 --> 00:14:44,239
Our best guess now is that there's

391
00:14:44,240 --> 00:14:46,119
multiple attackers doing this right

392
00:14:46,120 --> 00:14:47,779
this copycat once you hear about it

393
00:14:47,780 --> 00:14:48,939
there's other bad actors that want

394
00:14:48,940 --> 00:14:51,099
to jump on And um,

395
00:14:51,100 --> 00:14:52,739
I know the counts the counts of our

396
00:14:52,740 --> 00:14:54,958
said over a thousand bitcoin

397
00:14:54,959 --> 00:14:56,859
Uh stolen I believe it's

398
00:14:56,860 --> 00:14:58,779
now something like over 1700 from

399
00:14:58,780 --> 00:15:00,599
what I can tell Um, it's very

400
00:15:00,600 --> 00:15:02,299
hard to know exactly though.

401
00:15:02,300 --> 00:15:04,159
So we're only going to figure it out

402
00:15:04,160 --> 00:15:06,699
in the coming weeks and months

403
00:15:06,700 --> 00:15:08,299
And even though it's not a huge

404
00:15:08,300 --> 00:15:09,859
number of Bitcoin compared to some

405
00:15:09,860 --> 00:15:11,960
of these mega exchange collapses

406
00:15:15,260 --> 00:15:17,359
So I'll stop there, yeah.

407
00:15:17,360 --> 00:15:18,999
Yeah, sorry, someone's ringing my

408
00:15:19,000 --> 00:15:20,059
doorbell. I'm sure you can hear that

409
00:15:20,060 --> 00:15:21,060
in the background.

410
00:15:22,180 --> 00:15:23,339
Someone will get that downstairs.

411
00:15:24,560 --> 00:15:26,079
Yeah, so I think let's take it from

412
00:15:26,080 --> 00:15:27,080
there. Obviously,

413
00:15:28,140 --> 00:15:30,779
once news of this had come out,

414
00:15:30,780 --> 00:15:32,059
there was a concerted effort across

415
00:15:32,060 --> 00:15:33,299
the industry to try and figure out

416
00:15:33,300 --> 00:15:34,179
what this happened.

417
00:15:34,180 --> 00:15:36,299
One of the questions I had

418
00:15:36,300 --> 00:15:38,019
posed to me online, just in

419
00:15:38,020 --> 00:15:39,959
discourse with other people was,

420
00:15:39,960 --> 00:15:41,919
how much do you think AI tools

421
00:15:41,920 --> 00:15:44,599
played in the eventual.

422
00:15:44,600 --> 00:15:46,639
Figuring out of this

423
00:15:46,640 --> 00:15:47,539
bug.

424
00:15:47,540 --> 00:15:49,079
Do you think these frontier models

425
00:15:49,080 --> 00:15:50,519
that we've seen come out in recent

426
00:15:50,520 --> 00:15:53,279
months might have been the

427
00:15:53,280 --> 00:15:54,999
key that has allowed some attackers

428
00:15:55,000 --> 00:15:57,119
to figure out this

429
00:15:57,120 --> 00:15:59,079
exploitable attack vector and go

430
00:15:59,080 --> 00:16:00,399
through it and then obviously now go

431
00:16:00,400 --> 00:16:03,159
through the process of generating

432
00:16:03,160 --> 00:16:04,399
seeds and then running them through

433
00:16:04,400 --> 00:16:05,819
all these addresses to see if they

434
00:16:05,820 --> 00:16:07,919
can steal this Bitcoin basically?

435
00:16:09,020 --> 00:16:10,549
My quick answer is yes.

436
00:16:10,550 --> 00:16:12,189
Absolutely.

437
00:16:12,190 --> 00:16:14,569
Some folks think that

438
00:16:14,570 --> 00:16:16,789
maybe the reason that this happened

439
00:16:16,790 --> 00:16:19,009
on Thursday was because

440
00:16:19,010 --> 00:16:20,709
the latest Kimi model had just come

441
00:16:20,710 --> 00:16:22,849
out some number of days before.

442
00:16:22,850 --> 00:16:23,729
I think it was just a few days

443
00:16:23,730 --> 00:16:26,429
before and these

444
00:16:26,430 --> 00:16:28,629
models that are open weight models

445
00:16:28,630 --> 00:16:30,549
mean that

446
00:16:30,550 --> 00:16:32,809
there's no cyber security safeguards

447
00:16:32,810 --> 00:16:34,829
and so anyone is

448
00:16:34,830 --> 00:16:36,709
allowed to use them and they

449
00:16:36,710 --> 00:16:38,629
will do a pretty good job

450
00:16:38,630 --> 00:16:39,649
finding bugs and code.

451
00:16:41,240 --> 00:16:43,419
We as foundation have access

452
00:16:43,420 --> 00:16:45,359
to like GPT's cyber

453
00:16:45,360 --> 00:16:46,439
program, right?

454
00:16:46,440 --> 00:16:48,379
And so we're also able to run the

455
00:16:48,380 --> 00:16:50,359
frontier models on our own code

456
00:16:50,360 --> 00:16:52,179
or on other code, but.

457
00:16:53,240 --> 00:16:55,059
As you know, new models are coming

458
00:16:55,060 --> 00:16:56,839
out every month or every two months.

459
00:16:58,120 --> 00:17:00,079
And in our experience,

460
00:17:00,080 --> 00:17:02,139
up until recently, the Frontier

461
00:17:02,140 --> 00:17:04,059
models have not been

462
00:17:04,060 --> 00:17:06,279
the best at embedded engineering

463
00:17:06,280 --> 00:17:07,578
work, meaning like firmware for

464
00:17:07,579 --> 00:17:08,579
physical devices.

465
00:17:09,400 --> 00:17:10,779
We've been running Frontier Models

466
00:17:10,780 --> 00:17:12,219
that are our own code for months

467
00:17:12,220 --> 00:17:14,239
now, and there's been so many

468
00:17:14,240 --> 00:17:15,240
false positives.

469
00:17:16,280 --> 00:17:18,519
If there's, a misunderstanding

470
00:17:18,520 --> 00:17:20,539
of how our code bases work,

471
00:17:20,540 --> 00:17:22,358
because typically you have

472
00:17:22,359 --> 00:17:24,059
this firmware and it's calling on

473
00:17:24,060 --> 00:17:24,979
other libraries.

474
00:17:24,980 --> 00:17:26,499
And then sometimes the bugs exist in

475
00:17:26,500 --> 00:17:27,899
the interaction between the hardware

476
00:17:27,900 --> 00:17:29,739
and the software on the phone

477
00:17:29,740 --> 00:17:31,159
or computer.

478
00:17:31,160 --> 00:17:32,739
And AI still has a really hard time

479
00:17:32,740 --> 00:17:35,159
holding all of that in context.

480
00:17:35,160 --> 00:17:36,859
So I think we may have crossed the

481
00:17:36,860 --> 00:17:38,939
threshold in the last few weeks

482
00:17:38,940 --> 00:17:39,999
with some of these new models,

483
00:17:40,000 --> 00:17:42,419
because I pointed GPT

484
00:17:42,420 --> 00:17:44,259
5.6 sole, which is

485
00:17:44,260 --> 00:17:46,309
the latest model from OpenAI.

486
00:17:46,310 --> 00:17:48,589
At the cold card code base

487
00:17:48,590 --> 00:17:50,769
that afternoon on Thursday.

488
00:17:50,770 --> 00:17:52,549
And it was able to find the bug

489
00:17:52,550 --> 00:17:53,550
without any issue.

490
00:17:54,830 --> 00:17:56,749
Well, there was some

491
00:17:56,750 --> 00:17:59,029
inflection point that took place

492
00:17:59,030 --> 00:18:00,949
and there's a few lessons learned

493
00:18:00,950 --> 00:18:02,889
but I do not think the takeaway

494
00:18:02,890 --> 00:18:05,289
is that open source is bad.

495
00:18:05,290 --> 00:18:07,189
My takeaway is actually that free

496
00:18:07,190 --> 00:18:08,469
and open source software is

497
00:18:08,470 --> 00:18:11,399
necessary because

498
00:18:11,400 --> 00:18:13,239
If this had been a different

499
00:18:13,240 --> 00:18:14,959
project in the space, like let's say

500
00:18:14,960 --> 00:18:17,279
there was a common library

501
00:18:17,280 --> 00:18:19,179
that means like some block of code

502
00:18:19,180 --> 00:18:21,199
that multiple companies are

503
00:18:21,200 --> 00:18:22,259
using and building off of.

504
00:18:22,260 --> 00:18:23,459
And there's a lot of examples in

505
00:18:23,460 --> 00:18:25,039
Bitcoin, like Block has Bitcoin

506
00:18:25,040 --> 00:18:26,899
Development Kit, EDK.

507
00:18:26,900 --> 00:18:28,439
There's like Rust Bitcoin libraries.

508
00:18:28,440 --> 00:18:29,479
There's all these Bitcoin libraries

509
00:18:29,480 --> 00:18:31,539
that are used by multiple companies.

510
00:18:31,540 --> 00:18:32,319
They're free and open source

511
00:18:32,320 --> 00:18:33,649
software.

512
00:18:33,650 --> 00:18:35,189
I think that the likelihood of

513
00:18:35,190 --> 00:18:37,409
someone catching that, if it's in a

514
00:18:37,410 --> 00:18:38,789
commonly used library, is much

515
00:18:38,790 --> 00:18:39,869
higher because of all these real

516
00:18:39,870 --> 00:18:41,629
companies that are using the code

517
00:18:41,630 --> 00:18:42,709
and sharing on it and building on

518
00:18:42,710 --> 00:18:43,489
it.

519
00:18:43,490 --> 00:18:45,329
In this case, this was a cold card

520
00:18:45,330 --> 00:18:47,189
library that was

521
00:18:47,190 --> 00:18:49,189
a set of code that was

522
00:18:49,190 --> 00:18:50,069
publicly available.

523
00:18:50,070 --> 00:18:51,929
So we say source available,

524
00:18:51,930 --> 00:18:53,969
but not fully open source.

525
00:18:53,970 --> 00:18:55,629
And my understanding, though I can't

526
00:18:55,630 --> 00:18:56,589
say with certainty, but my

527
00:18:56,590 --> 00:18:57,589
understanding is they were the only

528
00:18:57,590 --> 00:18:59,689
company that was

529
00:18:59,690 --> 00:19:01,309
using this specific library.

530
00:19:03,299 --> 00:19:05,079
Now, obviously, this was just a bug

531
00:19:05,080 --> 00:19:06,419
that was introduced back then.

532
00:19:06,420 --> 00:19:07,719
There was no, it was not

533
00:19:07,720 --> 00:19:08,919
intentional, right?

534
00:19:08,920 --> 00:19:10,859
It's, there's a lot of

535
00:19:10,860 --> 00:19:12,899
bugs that exist in any code base.

536
00:19:12,900 --> 00:19:14,619
Every time we turn a frontier model

537
00:19:14,620 --> 00:19:16,279
out of our code, it finds bugs.

538
00:19:16,280 --> 00:19:17,959
The question is, like, what level of

539
00:19:17,960 --> 00:19:19,639
criticality, right, is that bug?

540
00:19:19,640 --> 00:19:20,739
Is it like a critical bug?

541
00:19:20,740 --> 00:19:22,699
Is it a high, medium, low, and so

542
00:19:22,700 --> 00:19:24,559
on? I do think AI is

543
00:19:24,560 --> 00:19:25,999
part of the reason why this

544
00:19:26,000 --> 00:19:27,339
happened. I also think it's part of

545
00:19:27,340 --> 00:19:28,339
a solution.

546
00:19:28,340 --> 00:19:29,439
And so it's just like any new

547
00:19:29,440 --> 00:19:30,419
technology, right.

548
00:19:30,420 --> 00:19:32,239
You have the good effects, you have

549
00:19:32,240 --> 00:19:33,519
the bad effects.

550
00:19:33,520 --> 00:19:35,299
If we try to create a thought

551
00:19:35,300 --> 00:19:36,819
experiment about what happened if

552
00:19:36,820 --> 00:19:38,839
the cold card code base was closed

553
00:19:38,840 --> 00:19:40,639
source and this exact same thing

554
00:19:40,640 --> 00:19:42,519
happened, which is totally possible,

555
00:19:42,520 --> 00:19:44,239
by the way, his attackers, like they

556
00:19:44,240 --> 00:19:46,039
reverse engineer binaries, they hook

557
00:19:46,040 --> 00:19:47,319
stuff up in the lab.

558
00:19:47,320 --> 00:19:49,159
You could have like a bad actor

559
00:19:49,160 --> 00:19:50,739
within the company or at a different

560
00:19:50,740 --> 00:19:52,019
company that has or even at

561
00:19:52,020 --> 00:19:53,519
Microsoft who has access to all the

562
00:19:53,520 --> 00:19:54,839
GitHub private repos.

563
00:19:54,840 --> 00:19:56,159
So like, let's just say is the

564
00:19:56,160 --> 00:19:57,159
thought experiment. If this was

565
00:19:57,160 --> 00:19:59,099
closed sourced, we would not

566
00:19:59,100 --> 00:20:00,719
have seen this immediate response

567
00:20:00,720 --> 00:20:02,169
from the Bitcoin community.

568
00:20:02,170 --> 00:20:03,849
Figuring out exactly what happened

569
00:20:03,850 --> 00:20:05,589
and helping thousands of people get

570
00:20:05,590 --> 00:20:07,189
to safety, we probably would have

571
00:20:07,190 --> 00:20:08,989
seen like a fog of war for days or

572
00:20:08,990 --> 00:20:10,389
weeks because no one would know what

573
00:20:10,390 --> 00:20:11,769
happened or why it happened or what

574
00:20:11,770 --> 00:20:12,629
the bug was.

575
00:20:12,630 --> 00:20:14,429
So I'm still a huge open source

576
00:20:14,430 --> 00:20:16,489
advocate, but this was definitely,

577
00:20:16,490 --> 00:20:17,490
I think,

578
00:20:18,510 --> 00:20:19,509
definitely had something to do with

579
00:20:19,510 --> 00:20:20,709
the latest AI models.

580
00:20:23,310 --> 00:20:25,309
Obviously, Foundation does

581
00:20:25,310 --> 00:20:27,189
hardware wallets as well,

582
00:20:27,190 --> 00:20:28,869
and we've seen lots of reactions

583
00:20:28,870 --> 00:20:32,069
from different self-custodial

584
00:20:32,070 --> 00:20:33,969
hardware wallet service providers

585
00:20:33,970 --> 00:20:35,809
coming out into port,

586
00:20:35,810 --> 00:20:37,609
but I can imagine internally, there

587
00:20:37,610 --> 00:20:39,509
was a huge fear initially that,

588
00:20:39,510 --> 00:20:41,069
like, holy shit, what happens if

589
00:20:41,070 --> 00:20:42,949
this is something that is affecting

590
00:20:42,950 --> 00:20:44,149
all devices?

591
00:20:44,150 --> 00:20:45,949
How quick was it, or how hard was it

592
00:20:45,950 --> 00:20:47,229
to rule it out and figure out that

593
00:20:47,230 --> 00:20:49,509
this only the

594
00:20:49,510 --> 00:20:51,009
specific bug that that was

595
00:20:51,010 --> 00:20:53,979
impacting. Mainly the MK3s.

596
00:20:53,980 --> 00:20:55,899
And how important is

597
00:20:55,900 --> 00:20:58,119
it for people in the industry and

598
00:20:58,120 --> 00:21:00,179
wider crypto natives

599
00:21:00,180 --> 00:21:02,319
to understand that this is not

600
00:21:02,320 --> 00:21:03,939
impacting all hardware wallets?

601
00:21:03,940 --> 00:21:05,319
Because I've been seeing a lot of

602
00:21:05,320 --> 00:21:06,659
people saying online that

603
00:21:06,660 --> 00:21:07,919
self-custody is dead.

604
00:21:07,920 --> 00:21:09,819
And I fundamentally don't believe

605
00:21:09,820 --> 00:21:11,219
that, and I wanna push back against

606
00:21:11,220 --> 00:21:12,469
that narrative myself.

607
00:21:12,470 --> 00:21:13,709
Yeah, I'll answer that.

608
00:21:13,710 --> 00:21:14,729
But one thing you just said, I want

609
00:21:14,730 --> 00:21:16,709
to make really clear,

610
00:21:16,710 --> 00:21:18,669
um, MK threes have been,

611
00:21:18,670 --> 00:21:20,709
are, are most affected, but

612
00:21:20,710 --> 00:21:22,509
all MK foreign leaders are still at

613
00:21:22,510 --> 00:21:23,929
risk. They still have some better

614
00:21:23,930 --> 00:21:25,429
entropy. So just like, if you have a

615
00:21:25,430 --> 00:21:27,509
cold card, if it's MK three

616
00:21:27,510 --> 00:21:29,029
or newer, I would, I will get your

617
00:21:29,030 --> 00:21:30,029
funds out of there as fast as

618
00:21:30,030 --> 00:21:31,049
possible. So that's, that's the

619
00:21:31,050 --> 00:21:32,649
first thing I want to address, um

620
00:21:32,650 --> 00:21:33,670
with self custody.

621
00:21:35,150 --> 00:21:36,869
It's really, it's really dangerous

622
00:21:36,870 --> 00:21:38,549
to say self custody is dead because

623
00:21:38,550 --> 00:21:40,729
of this. I mean, I, I completely

624
00:21:40,730 --> 00:21:41,730
disagree.

625
00:21:42,840 --> 00:21:45,999
However, it's scary.

626
00:21:46,000 --> 00:21:47,879
Obviously, every hardware

627
00:21:47,880 --> 00:21:49,399
wallet maker, probably every wallet

628
00:21:49,400 --> 00:21:50,799
maker in the last few days across

629
00:21:50,800 --> 00:21:52,499
the entire industry has been, I

630
00:21:52,500 --> 00:21:54,439
hope, turning AI loose

631
00:21:54,440 --> 00:21:56,159
on their code base and reviewing

632
00:21:56,160 --> 00:21:57,399
some of the most critical parts of

633
00:21:57,400 --> 00:21:58,579
the code.

634
00:21:58,580 --> 00:21:59,899
There's been some amazing Bitcoiners

635
00:21:59,900 --> 00:22:01,999
that have been helping with that.

636
00:22:02,000 --> 00:22:03,779
Like Rob Hamilton at Anchor Watch

637
00:22:03,780 --> 00:22:05,659
has been using, I believe

638
00:22:05,660 --> 00:22:07,739
Kimmy to submit bug

639
00:22:07,740 --> 00:22:09,459
reports to virtually every hardware

640
00:22:09,460 --> 00:22:10,499
wallets, software wallet.

641
00:22:10,500 --> 00:22:12,549
We've been talking with him.

642
00:22:12,550 --> 00:22:14,589
He submitted stuff to us,

643
00:22:14,590 --> 00:22:17,069
there's one that scared us because

644
00:22:17,070 --> 00:22:19,029
Kimi thought that we had a potential

645
00:22:19,030 --> 00:22:20,869
bug. It was nothing like as

646
00:22:20,870 --> 00:22:22,869
bad as the cold card bug,

647
00:22:22,870 --> 00:22:24,869
but we had

648
00:22:24,870 --> 00:22:26,169
to really dive into it with human

649
00:22:26,170 --> 00:22:27,889
review and realize that it was

650
00:22:27,890 --> 00:22:29,709
wrong. And so there's another aspect

651
00:22:29,710 --> 00:22:31,569
here where there's

652
00:22:31,570 --> 00:22:32,869
going to be a lot of panic and

653
00:22:32,870 --> 00:22:34,129
uncertainty in the short term

654
00:22:34,130 --> 00:22:35,549
because customers are thinking, is

655
00:22:35,550 --> 00:22:37,099
my hardware wallet safe?

656
00:22:37,100 --> 00:22:38,639
They're running AI models now

657
00:22:38,640 --> 00:22:40,019
against their hardware wallets that

658
00:22:40,020 --> 00:22:41,479
they're using, and sometimes AI

659
00:22:41,480 --> 00:22:42,979
models are saying, oh, there's some

660
00:22:42,980 --> 00:22:44,799
critical bug, but the bug doesn't

661
00:22:44,800 --> 00:22:46,639
actually exist because the AI models

662
00:22:46,640 --> 00:22:48,559
are still not at that level

663
00:22:48,560 --> 00:22:50,199
and probably will never be at that

664
00:22:50,200 --> 00:22:51,759
levels of perfection, right?

665
00:22:51,760 --> 00:22:53,139
Nor are humans.

666
00:22:53,140 --> 00:22:54,959
And so there's a lot of complexity

667
00:22:54,960 --> 00:22:55,959
happening. So yeah, I mean,

668
00:22:55,960 --> 00:22:57,319
obviously, the first thing I did

669
00:22:57,320 --> 00:22:59,919
Thursday evening was

670
00:22:59,920 --> 00:23:01,739
run the

671
00:23:01,740 --> 00:23:03,659
Frontier models against our code

672
00:23:03,660 --> 00:23:05,549
base, and our CTO did the same.

673
00:23:05,550 --> 00:23:06,889
We breathed, you know, a sigh of

674
00:23:06,890 --> 00:23:07,849
relief, right? That there was no

675
00:23:07,850 --> 00:23:09,309
entropy bug. And I'm sure that exact

676
00:23:09,310 --> 00:23:11,189
same thing happened at all the other

677
00:23:11,190 --> 00:23:11,749
wallets.

678
00:23:11,750 --> 00:23:13,529
Um, yeah.

679
00:23:15,360 --> 00:23:17,739
Look, I think the

680
00:23:17,740 --> 00:23:19,379
next kind of question that comes

681
00:23:19,380 --> 00:23:22,279
into my mind now is the

682
00:23:22,280 --> 00:23:24,119
notion of trust in Bitcoin, because

683
00:23:24,120 --> 00:23:26,059
the old saying goes

684
00:23:26,060 --> 00:23:28,039
very hard, don't trust, verify, and

685
00:23:28,040 --> 00:23:29,679
a lot of people trusted here.

686
00:23:29,680 --> 00:23:30,959
And I think that's the great tragedy

687
00:23:30,960 --> 00:23:33,059
of this, is a lot people trusted

688
00:23:33,060 --> 00:23:35,059
CoinKite, the

689
00:23:35,060 --> 00:23:37,039
company, and NVK, and

690
00:23:37,040 --> 00:23:39,039
I'm not here to bash them, I'm just

691
00:23:39,040 --> 00:23:40,419
not that person. And no one could

692
00:23:40,420 --> 00:23:42,379
have known, and it was very hard to

693
00:23:42,380 --> 00:23:43,269
find this bike.

694
00:23:43,270 --> 00:23:45,149
And no one found it for a very long

695
00:23:45,150 --> 00:23:46,489
period of time, right?

696
00:23:46,490 --> 00:23:48,629
But there is that element of

697
00:23:48,630 --> 00:23:50,269
trust that most people don't even

698
00:23:50,270 --> 00:23:52,149
realize is still there when

699
00:23:52,150 --> 00:23:54,069
they're using a hardware wallet like

700
00:23:54,070 --> 00:23:55,989
this. And it does make

701
00:23:55,990 --> 00:23:57,729
me kind of wonder like how does the

702
00:23:57,730 --> 00:23:58,709
industry address that?

703
00:23:58,710 --> 00:24:00,509
Like this element of trusts because

704
00:24:00,510 --> 00:24:02,349
it was supposed to be trustless,

705
00:24:02,350 --> 00:24:03,289
right.

706
00:24:03,290 --> 00:24:05,169
Like here's a hardware wallets,

707
00:24:05,170 --> 00:24:06,989
now you can be the master

708
00:24:06,990 --> 00:24:08,209
of your own coins.

709
00:24:08,210 --> 00:24:11,109
And in this instance, that failed.

710
00:24:11,110 --> 00:24:13,009
How do you feel about that?

711
00:24:13,010 --> 00:24:14,489
Yeah, I think it's funny.

712
00:24:14,490 --> 00:24:16,109
I'm running a blog post right now

713
00:24:16,110 --> 00:24:17,929
that has to do with don't trust

714
00:24:17,930 --> 00:24:19,609
verify to talk about all this

715
00:24:19,610 --> 00:24:21,489
because I think what

716
00:24:21,490 --> 00:24:23,349
happened is that um

717
00:24:23,350 --> 00:24:25,609
The whole Bitcoin media and

718
00:24:25,610 --> 00:24:26,529
podcast. I don't want to say the

719
00:24:26,530 --> 00:24:27,829
whole but a majority of the Bitcoin

720
00:24:27,830 --> 00:24:29,490
media podcasting community

721
00:24:30,810 --> 00:24:31,810
has

722
00:24:32,990 --> 00:24:33,990
Trusted

723
00:24:34,929 --> 00:24:36,369
You know, I don't want to stay

724
00:24:36,370 --> 00:24:37,929
blindly trusted this company I think

725
00:24:37,930 --> 00:24:39,889
I think you know, they have good

726
00:24:39,890 --> 00:24:41,769
intentions I do not think that

727
00:24:41,770 --> 00:24:43,329
this bug was intentional.

728
00:24:43,330 --> 00:24:45,309
MBK has been a Bitcoiner for a long

729
00:24:45,310 --> 00:24:47,649
time, but

730
00:24:47,650 --> 00:24:49,449
this don't trust verify is really

731
00:24:49,450 --> 00:24:50,289
important.

732
00:24:50,290 --> 00:24:51,409
And I feel like the Bitcoin

733
00:24:51,410 --> 00:24:53,769
community should be leaning more on,

734
00:24:53,770 --> 00:24:55,609
if you're going to trust a wide

735
00:24:55,610 --> 00:24:57,609
array of technical, you

736
00:24:57,610 --> 00:24:58,789
know, folks in the Bitcoin

737
00:24:58,790 --> 00:25:00,809
Community, it shouldn't be

738
00:25:00,810 --> 00:25:03,229
based around who

739
00:25:03,230 --> 00:25:05,349
sponsors what podcast,

740
00:25:05,350 --> 00:25:07,269
who's friends with whom, it

741
00:25:07,270 --> 00:25:09,229
should be much more about

742
00:25:09,230 --> 00:25:11,189
trusting technical leaders within

743
00:25:11,190 --> 00:25:12,399
the community.

744
00:25:12,400 --> 00:25:14,539
And I do think that the

745
00:25:14,540 --> 00:25:16,519
media has to be careful, both within

746
00:25:16,520 --> 00:25:18,759
Bitcoin and within all of crypto,

747
00:25:18,760 --> 00:25:20,999
with recommending specific hardware

748
00:25:21,000 --> 00:25:22,299
wallets or software wallets, which

749
00:25:22,300 --> 00:25:23,419
is tough, right, because you want to

750
00:25:23,420 --> 00:25:24,979
be able to take sponsorship money.

751
00:25:24,980 --> 00:25:26,619
I mean, we've sponsored stuff before

752
00:25:26,620 --> 00:25:28,779
and sometimes still do, but

753
00:25:28,780 --> 00:25:30,339
it creates a really difficult

754
00:25:30,340 --> 00:25:31,339
cascading effect.

755
00:25:31,340 --> 00:25:33,039
And I know so many amazing.

756
00:25:35,360 --> 00:25:37,239
Know, Bitcoiners who have large

757
00:25:37,240 --> 00:25:39,299
audiences who are just torn up about

758
00:25:39,300 --> 00:25:41,279
this completely, right, because

759
00:25:41,280 --> 00:25:43,179
you recommend to so

760
00:25:43,180 --> 00:25:45,519
many people a

761
00:25:45,520 --> 00:25:47,199
specific device, you tell them to

762
00:25:47,200 --> 00:25:49,059
self-custody, not

763
00:25:49,060 --> 00:25:50,699
your keys, not your coins, don't

764
00:25:50,700 --> 00:25:52,619
trust verify, you help so many

765
00:25:52,620 --> 00:25:54,239
people, thousands of people, tons of

766
00:25:54,240 --> 00:25:55,499
thousands of people on board to self

767
00:25:55,500 --> 00:25:56,919
custody, and then this happens.

768
00:25:56,920 --> 00:25:58,819
And so I do think that

769
00:25:58,820 --> 00:26:00,219
there needs to be some introspection

770
00:26:00,220 --> 00:26:02,359
within the Bitcoin community.

771
00:26:02,360 --> 00:26:03,679
I don't know if that's a good enough

772
00:26:03,680 --> 00:26:04,529
answer.

773
00:26:04,530 --> 00:26:06,429
Like I wish I had a better answer to

774
00:26:06,430 --> 00:26:07,869
your question.

775
00:26:07,870 --> 00:26:09,109
I think the principles are still

776
00:26:09,110 --> 00:26:09,969
there.

777
00:26:09,970 --> 00:26:11,769
I just think that we as a community

778
00:26:11,770 --> 00:26:13,189
need to really figure out what that

779
00:26:13,190 --> 00:26:14,829
means. One thing I would say is that

780
00:26:14,830 --> 00:26:16,649
like, what

781
00:26:16,650 --> 00:26:17,989
are we looking for in terms of

782
00:26:17,990 --> 00:26:19,109
verification?

783
00:26:19,110 --> 00:26:21,269
Things like security audits are

784
00:26:21,270 --> 00:26:23,589
really important, engaging with

785
00:26:23,590 --> 00:26:25,449
the security research community.

786
00:26:25,450 --> 00:26:26,889
Like we'll get bug reports all the

787
00:26:26,890 --> 00:26:28,769
time. It's really important when you

788
00:26:28,770 --> 00:26:30,149
engage with security researchers

789
00:26:30,150 --> 00:26:31,969
that you do so with a great amount

790
00:26:31,970 --> 00:26:32,969
of respect.

791
00:26:32,970 --> 00:26:34,429
It can sometimes be an antagonistic

792
00:26:34,430 --> 00:26:35,649
relationship, right?

793
00:26:35,650 --> 00:26:37,009
Someone reaches out to us, they say

794
00:26:37,010 --> 00:26:38,469
they found a critical bug.

795
00:26:38,470 --> 00:26:39,849
Sometimes your first reaction as a

796
00:26:39,850 --> 00:26:41,789
wallet company is to get upset about

797
00:26:41,790 --> 00:26:43,589
that. And I think it needs to be the

798
00:26:43,590 --> 00:26:44,489
opposite.

799
00:26:44,490 --> 00:26:45,989
And then also just free and open

800
00:26:45,990 --> 00:26:47,869
source software, collaborating with

801
00:26:47,870 --> 00:26:49,409
each other, building on shared

802
00:26:49,410 --> 00:26:51,089
libraries.

803
00:26:51,090 --> 00:26:52,389
All those things I think are really

804
00:26:52,390 --> 00:26:54,189
helpful. And when we say don't

805
00:26:54,190 --> 00:26:55,509
trust, verify.

806
00:26:55,510 --> 00:26:56,529
I think that's the kind of

807
00:26:56,530 --> 00:26:57,809
verification that the space should

808
00:26:57,810 --> 00:26:59,109
be moving towards, you know, less

809
00:26:59,110 --> 00:27:03,129
about marketing

810
00:27:03,130 --> 00:27:05,029
a specific company, more

811
00:27:05,030 --> 00:27:07,129
about giving

812
00:27:07,130 --> 00:27:08,869
the technical leaders and the

813
00:27:08,870 --> 00:27:10,889
security experts a voice as much as

814
00:27:10,890 --> 00:27:12,149
possible, especially in the coming

815
00:27:12,150 --> 00:27:13,150
weeks.

816
00:27:13,550 --> 00:27:15,629
Yeah Zack we've got

817
00:27:15,630 --> 00:27:16,469
a lot backstage.

818
00:27:16,470 --> 00:27:18,309
So I'll ask my final question to

819
00:27:18,310 --> 00:27:20,269
you I mean just just in terms of

820
00:27:20,270 --> 00:27:22,029
the the hardware wallet space, you

821
00:27:22,030 --> 00:27:24,069
know We've seen kind of newcomers

822
00:27:24,070 --> 00:27:26,049
come in in recent years I think you

823
00:27:26,050 --> 00:27:28,409
know block streams Jade bit

824
00:27:28,410 --> 00:27:30,409
key from block a few

825
00:27:30,410 --> 00:27:32,609
newcomers that have definitely added

826
00:27:32,610 --> 00:27:34,009
More hardware wallets to the

827
00:27:34,010 --> 00:27:35,649
industry. I'm kind of curious.

828
00:27:35,650 --> 00:27:37,809
Do you think that this situation

829
00:27:37,810 --> 00:27:40,628
does call for for more?

830
00:27:40,629 --> 00:27:42,429
Options in the industry and will

831
00:27:42,430 --> 00:27:44,069
that kind of help drive things

832
00:27:44,070 --> 00:27:46,269
forward? And from a usability

833
00:27:46,270 --> 00:27:48,289
perspective, it does kind of beg the

834
00:27:48,290 --> 00:27:50,029
question. It's like, is it still too

835
00:27:50,030 --> 00:27:51,449
complicated? Does it need to be

836
00:27:51,450 --> 00:27:53,149
easier than it is?

837
00:27:53,150 --> 00:27:55,189
Because a lot of cold card users

838
00:27:55,190 --> 00:27:56,190
were told like,

839
00:27:57,990 --> 00:27:58,929
this will be safe.

840
00:27:58,930 --> 00:28:00,189
You're fine.

841
00:28:00,190 --> 00:28:01,349
But if you really want to go the

842
00:28:01,350 --> 00:28:05,509
extra step, you need to do the

843
00:28:05,510 --> 00:28:07,429
the coin toss dice

844
00:28:07,430 --> 00:28:08,779
roll.

845
00:28:08,780 --> 00:28:10,459
Option to go and generate a seed,

846
00:28:10,460 --> 00:28:12,219
that's hugely complicated for the

847
00:28:12,220 --> 00:28:13,959
layman who's looking to self-custody

848
00:28:13,960 --> 00:28:15,019
their Bitcoin.

849
00:28:15,020 --> 00:28:17,039
So A, do we need more

850
00:28:17,040 --> 00:28:18,439
variety of hardware wallet

851
00:28:18,440 --> 00:28:20,479
manufacturers in the industry?

852
00:28:20,480 --> 00:28:22,419
And do we have a user experience

853
00:28:22,420 --> 00:28:23,420
problem?

854
00:28:25,150 --> 00:28:27,269
I think that more is not

855
00:28:27,270 --> 00:28:28,889
necessarily the answer, though I'm

856
00:28:28,890 --> 00:28:30,090
conflicted, obviously.

857
00:28:31,690 --> 00:28:34,329
I think more free and open source

858
00:28:34,330 --> 00:28:36,249
software and hardware options

859
00:28:36,250 --> 00:28:38,089
is part of the

860
00:28:38,090 --> 00:28:39,349
answer.

861
00:28:39,350 --> 00:28:40,749
Some of those newcomers you

862
00:28:40,750 --> 00:28:42,889
mentioned are open source,

863
00:28:43,910 --> 00:28:45,129
BitKey is open source.

864
00:28:45,130 --> 00:28:46,649
It was not fully open source at

865
00:28:46,650 --> 00:28:47,769
first.

866
00:28:47,770 --> 00:28:49,549
Many fiercely criticized that,

867
00:28:49,550 --> 00:28:51,129
myself included, and they have now

868
00:28:51,130 --> 00:28:52,809
open sourced everything.

869
00:28:52,810 --> 00:28:54,729
Jade is open source, there's

870
00:28:54,730 --> 00:28:55,829
many that are open source.

871
00:28:55,830 --> 00:28:57,749
And so I think the open

872
00:28:57,750 --> 00:28:59,109
source aspect is very important

873
00:28:59,110 --> 00:29:01,209
here. And I think only

874
00:29:01,210 --> 00:29:03,129
newcomers to the hardware wallet

875
00:29:03,130 --> 00:29:04,749
world should be taken seriously if

876
00:29:04,750 --> 00:29:06,649
their stuff is open source, is a

877
00:29:06,650 --> 00:29:08,509
big takeaway that I want to

878
00:29:08,510 --> 00:29:10,489
push as part of this moment

879
00:29:10,490 --> 00:29:12,149
in conversation.

880
00:29:12,150 --> 00:29:14,229
With regard to the UX, absolutely a

881
00:29:14,230 --> 00:29:15,289
problem.

882
00:29:15,290 --> 00:29:16,149
I mean, 100%.

883
00:29:16,150 --> 00:29:17,589
And I've actually been trying to

884
00:29:17,590 --> 00:29:18,709
think about how to react to that,

885
00:29:18,710 --> 00:29:19,749
because I've been in a lot of these

886
00:29:19,750 --> 00:29:21,269
Twitter spaces.

887
00:29:21,270 --> 00:29:22,629
Trying to help some folks over the

888
00:29:22,630 --> 00:29:23,909
last few days.

889
00:29:23,910 --> 00:29:26,269
Um, there's been some incredible

890
00:29:26,270 --> 00:29:29,009
Twitter spaces, uh,

891
00:29:29,010 --> 00:29:30,929
one, one guy who goes by wicked

892
00:29:30,930 --> 00:29:32,489
has been hosting one basically for

893
00:29:32,490 --> 00:29:34,329
the last like four days straight is

894
00:29:34,330 --> 00:29:35,709
extraordinary that he's been able to

895
00:29:35,710 --> 00:29:37,529
do that, but everyone's coming

896
00:29:37,530 --> 00:29:39,029
up on stage asking about dice rolls

897
00:29:39,030 --> 00:29:40,489
and coin flips, and I don't think

898
00:29:40,490 --> 00:29:41,889
that should take away here.

899
00:29:41,890 --> 00:29:42,729
I think it makes sense.

900
00:29:42,730 --> 00:29:44,549
If you're a cold card customer and

901
00:29:44,550 --> 00:29:46,129
you, and you can't switch to another

902
00:29:46,130 --> 00:29:47,749
device and you're like, how do I

903
00:29:47,750 --> 00:29:50,399
generate a seed securely now?

904
00:29:50,400 --> 00:29:53,119
What could I have done differently?

905
00:29:53,120 --> 00:29:54,919
But for every wallet that we know

906
00:29:54,920 --> 00:29:56,099
of, every mainstream hardware wallet

907
00:29:56,100 --> 00:29:57,979
besides cold card, like the

908
00:29:57,980 --> 00:29:59,839
C generation code has been

909
00:29:59,840 --> 00:30:00,759
okay, right?

910
00:30:00,760 --> 00:30:02,239
It's been okay.

911
00:30:02,240 --> 00:30:04,159
And I'm sure Jameson who's gonna

912
00:30:04,160 --> 00:30:05,479
be up next will talk about

913
00:30:05,480 --> 00:30:07,299
multi-vendor, multi-SIG, which I

914
00:30:07,300 --> 00:30:08,899
think is a solution.

915
00:30:08,900 --> 00:30:10,799
There's some good solutions

916
00:30:10,800 --> 00:30:12,839
that offer a good user experience as

917
00:30:12,840 --> 00:30:13,779
part of that.

918
00:30:13,780 --> 00:30:15,999
From my perspective though,

919
00:30:16,000 --> 00:30:18,139
I think it really does come down to

920
00:30:18,140 --> 00:30:20,279
making sure that you're

921
00:30:20,280 --> 00:30:22,039
code is bulletproof for the C

922
00:30:22,040 --> 00:30:23,719
generation as a hardware wallet

923
00:30:23,720 --> 00:30:25,259
vendor, which is not that much good

924
00:30:25,260 --> 00:30:27,139
to, you know, inspect and

925
00:30:27,140 --> 00:30:28,499
trying to create a really great user

926
00:30:28,500 --> 00:30:30,119
experience, both for onboarding and

927
00:30:30,120 --> 00:30:31,299
backups.

928
00:30:31,300 --> 00:30:33,239
You know, we use like NFC key

929
00:30:33,240 --> 00:30:34,339
cards that we call them, where you

930
00:30:34,340 --> 00:30:35,799
can tap to the device as an option

931
00:30:35,800 --> 00:30:37,359
right now.

932
00:30:37,360 --> 00:30:38,519
BitKey does some really interesting

933
00:30:38,520 --> 00:30:39,959
stuff.

934
00:30:39,960 --> 00:30:41,679
We're going to be, I think, further

935
00:30:41,680 --> 00:30:43,679
improving our onboarding and

936
00:30:43,680 --> 00:30:45,079
backup process because you want to

937
00:30:45,080 --> 00:30:46,619
make it as easy as possible for

938
00:30:46,620 --> 00:30:48,059
folks. And so...

939
00:30:48,060 --> 00:30:49,919
I don't think the takeaway should

940
00:30:49,920 --> 00:30:52,019
be do not trust all

941
00:30:52,020 --> 00:30:54,199
random number generation after this.

942
00:30:54,200 --> 00:30:55,279
I think that's the takeaway I've

943
00:30:55,280 --> 00:30:57,199
been seeing on, you

944
00:30:57,200 --> 00:30:58,639
know, these spaces and everything

945
00:30:58,640 --> 00:31:00,399
because people are scared.

946
00:31:00,400 --> 00:31:02,099
But I do think that this was more

947
00:31:02,100 --> 00:31:04,359
about a horrible bug

948
00:31:04,360 --> 00:31:06,399
with a specific vendor.

949
00:31:06,400 --> 00:31:08,219
But if you are advanced enough and

950
00:31:08,220 --> 00:31:09,559
you do want to generate your own

951
00:31:09,560 --> 00:31:11,439
seed words, You can pick seabirds

952
00:31:11,440 --> 00:31:13,289
out of a hat, for example.

953
00:31:13,290 --> 00:31:14,589
Uh, you can enter them into the

954
00:31:14,590 --> 00:31:16,589
device, a device like passport can

955
00:31:16,590 --> 00:31:18,229
auto fill the last word for you,

956
00:31:18,230 --> 00:31:19,889
like the 11th, the 12th or the 24th

957
00:31:19,890 --> 00:31:21,669
word, where there's a bunch of other

958
00:31:21,670 --> 00:31:22,769
ways that folks have been covering,

959
00:31:22,770 --> 00:31:23,729
like you should do that.

960
00:31:23,730 --> 00:31:24,949
Right. Cause it's empowering to be

961
00:31:24,950 --> 00:31:26,369
able to do that yourself and not

962
00:31:26,370 --> 00:31:27,889
rely on anyone else.

963
00:31:27,890 --> 00:31:29,149
So I think that would be my major

964
00:31:29,150 --> 00:31:30,769
takeaways, but if I get just the one

965
00:31:30,770 --> 00:31:32,009
big takeaway, right.

966
00:31:32,010 --> 00:31:33,269
Free and open source software I

967
00:31:33,270 --> 00:31:35,549
think is, is our, uh,

968
00:31:35,550 --> 00:31:36,589
longterm solution.

969
00:31:38,300 --> 00:31:39,939
Zach, thanks so much for taking the

970
00:31:39,940 --> 00:31:41,139
time to chat to us today.

971
00:31:41,140 --> 00:31:42,339
And for anyone watching out there,

972
00:31:42,340 --> 00:31:44,159
you can follow Zach on X and go and

973
00:31:44,160 --> 00:31:45,339
check out Foundation as well if

974
00:31:45,340 --> 00:31:46,659
you're looking for alternative

975
00:31:46,660 --> 00:31:48,939
options. It's at Z Herbert on

976
00:31:48,940 --> 00:31:51,059
X and at Foundation

977
00:31:51,060 --> 00:31:53,179
HQ I believe is the X handle.

978
00:31:53,180 --> 00:31:54,199
Zach, thank you very much.

979
00:31:54,200 --> 00:31:55,459
We're gonna take a very short break

980
00:31:55,460 --> 00:31:57,279
just to check Lop's

981
00:31:57,280 --> 00:31:59,099
audio and visuals.

982
00:31:59,100 --> 00:32:00,260
We'll be right back. Don't go away.

983
00:34:28,179 --> 00:34:29,658
Welcome back to the starting block.

984
00:34:29,659 --> 00:34:31,138
We are continuing our hot seat

985
00:34:31,139 --> 00:34:32,619
conversation. We've got Jameson

986
00:34:32,620 --> 00:34:34,599
Lopp, who's of course a CTO at

987
00:34:34,600 --> 00:34:36,499
Casa, very well-known name in the

988
00:34:36,500 --> 00:34:37,779
Bitcoin industry.

989
00:34:37,780 --> 00:34:39,259
Jameson, it's great to have you on

990
00:34:39,260 --> 00:34:40,039
the show.

991
00:34:40,040 --> 00:34:41,859
First things first, I'd love to hear

992
00:34:41,860 --> 00:34:43,359
your kind of recollection of events

993
00:34:43,360 --> 00:34:45,218
last week, when you first saw

994
00:34:45,219 --> 00:34:47,459
the cold card exploit

995
00:34:47,460 --> 00:34:49,238
news and what your immediate

996
00:34:49,239 --> 00:34:50,839
thoughts were and what you did in

997
00:34:50,840 --> 00:34:51,840
the hours after that.

998
00:34:53,240 --> 00:34:55,238
Well, actually and

999
00:34:55,239 --> 00:34:57,399
oddly enough, I believe I

1000
00:34:57,400 --> 00:35:00,419
was one of the first people to

1001
00:35:00,420 --> 00:35:02,619
have a hint of it.

1002
00:35:02,620 --> 00:35:04,739
And that's because I had someone

1003
00:35:04,740 --> 00:35:07,059
email me privately through my

1004
00:35:07,060 --> 00:35:10,379
website in

1005
00:35:10,380 --> 00:35:12,339
the wee early hours of

1006
00:35:12,340 --> 00:35:13,619
the morning, because they were

1007
00:35:13,620 --> 00:35:15,639
actually over in Australia in

1008
00:35:15,640 --> 00:35:16,709
that area.

1009
00:35:16,710 --> 00:35:18,629
And they detailed that

1010
00:35:18,630 --> 00:35:20,509
they were a cold card user and they

1011
00:35:20,510 --> 00:35:22,609
had seen a partial loss

1012
00:35:22,610 --> 00:35:24,529
of funds and they described their

1013
00:35:24,530 --> 00:35:25,769
entire setup and their best

1014
00:35:25,770 --> 00:35:27,129
practices to me.

1015
00:35:27,130 --> 00:35:29,189
I was pretty baffled for several

1016
00:35:29,190 --> 00:35:30,389
reasons.

1017
00:35:30,390 --> 00:35:31,829
First of all, because it was only a

1018
00:35:31,830 --> 00:35:33,489
partial loss of fund and usually you

1019
00:35:33,490 --> 00:35:35,249
see full loss of funds when a seed

1020
00:35:35,250 --> 00:35:36,250
is compromised.

1021
00:35:37,730 --> 00:35:39,669
And at the moment,

1022
00:35:39,670 --> 00:35:41,489
I had just gotten back to them

1023
00:35:41,490 --> 00:35:43,729
and said, hey, we

1024
00:35:43,730 --> 00:35:45,510
need more information.

1025
00:35:46,700 --> 00:35:48,739
Need to better understand if

1026
00:35:48,740 --> 00:35:50,439
you have done something that could

1027
00:35:50,440 --> 00:35:52,419
have caused this.

1028
00:35:52,420 --> 00:35:54,239
And then of course, as the day went

1029
00:35:54,240 --> 00:35:55,979
on, we started seeing more and more

1030
00:35:55,980 --> 00:35:58,519
reports and the

1031
00:35:58,520 --> 00:36:00,459
only real commonality was that

1032
00:36:00,460 --> 00:36:02,479
these were all cold card users.

1033
00:36:02,480 --> 00:36:04,379
And so that's when the

1034
00:36:04,380 --> 00:36:07,559
security community really

1035
00:36:07,560 --> 00:36:09,599
woke up and said, okay, well,

1036
00:36:09,600 --> 00:36:11,419
we still don't know what it is,

1037
00:36:11,420 --> 00:36:13,299
but... Signs are pointing to

1038
00:36:13,300 --> 00:36:14,739
cold card and that's when a lot of

1039
00:36:14,740 --> 00:36:17,139
people started going and inspecting

1040
00:36:17,140 --> 00:36:19,339
the firmware more closely and

1041
00:36:19,340 --> 00:36:20,919
I think it was by the end of the day

1042
00:36:20,920 --> 00:36:22,759
that it was determined that

1043
00:36:22,760 --> 00:36:24,619
there was essentially

1044
00:36:24,620 --> 00:36:27,019
an entropy problem

1045
00:36:27,020 --> 00:36:28,619
you know with the random number

1046
00:36:28,620 --> 00:36:29,620
generator.

1047
00:36:30,320 --> 00:36:33,599
Once that was discovered then

1048
00:36:33,600 --> 00:36:35,419
really the goal was to

1049
00:36:35,420 --> 00:36:37,479
understand what the severity of it

1050
00:36:37,480 --> 00:36:39,379
was and you know

1051
00:36:39,380 --> 00:36:41,479
how Basically,

1052
00:36:41,480 --> 00:36:43,469
how long... People had before

1053
00:36:43,470 --> 00:36:45,349
they really could expect to get

1054
00:36:45,350 --> 00:36:47,109
popped, depending upon a number of

1055
00:36:47,110 --> 00:36:49,049
different variables of how they had

1056
00:36:49,050 --> 00:36:50,790
generated their seed phrase.

1057
00:36:52,110 --> 00:36:54,309
I mean, it's pretty crazy

1058
00:36:54,310 --> 00:36:56,269
how much Bitcoin has been stolen

1059
00:36:56,270 --> 00:36:58,189
since, but it's

1060
00:36:58,190 --> 00:36:59,229
kind of still ongoing, right?

1061
00:36:59,230 --> 00:37:00,729
And I saw a few anecdotes over the

1062
00:37:00,730 --> 00:37:02,809
weekend of people being

1063
00:37:02,810 --> 00:37:04,389
on holiday away from the cold card

1064
00:37:04,390 --> 00:37:05,829
devices, knowing that this has

1065
00:37:05,830 --> 00:37:06,849
played out and they can't do

1066
00:37:06,850 --> 00:37:08,069
anything, they can get to the

1067
00:37:08,070 --> 00:37:10,049
devices and move their funds

1068
00:37:10,050 --> 00:37:11,050
off.

1069
00:37:12,010 --> 00:37:13,889
Yeah, I guess my first kind

1070
00:37:13,890 --> 00:37:16,289
of question to you

1071
00:37:16,290 --> 00:37:18,519
is when you kind of saw that

1072
00:37:18,520 --> 00:37:20,399
the issue was with the random number

1073
00:37:20,400 --> 00:37:21,799
generator and the entropy involved

1074
00:37:21,800 --> 00:37:24,019
in mainly the MK3 devices,

1075
00:37:24,020 --> 00:37:25,839
but perhaps more of the cold card

1076
00:37:25,840 --> 00:37:26,840
devices.

1077
00:37:28,180 --> 00:37:29,879
You said to me, Orfea, that this is

1078
00:37:29,880 --> 00:37:31,739
something that one of the

1079
00:37:31,740 --> 00:37:32,939
sort of value propositions of

1080
00:37:32,940 --> 00:37:35,559
something like CASA is

1081
00:37:35,560 --> 00:37:37,439
ruling out this kind of

1082
00:37:37,440 --> 00:37:38,700
single point of failure.

1083
00:37:41,040 --> 00:37:42,779
I mean, I don't want to say, do you

1084
00:37:42,780 --> 00:37:43,979
think that this was kind of long

1085
00:37:43,980 --> 00:37:45,779
time coming that we would see an

1086
00:37:45,780 --> 00:37:47,139
issue like this with a hardware

1087
00:37:47,140 --> 00:37:48,179
wallet manufacturer?

1088
00:37:48,180 --> 00:37:50,019
But- Is it quite in your

1089
00:37:50,020 --> 00:37:51,419
mind rare that something like this

1090
00:37:51,420 --> 00:37:52,759
happened?

1091
00:37:52,760 --> 00:37:54,499
And do you think it sets a stake in

1092
00:37:54,500 --> 00:37:56,239
the ground where hopefully we don't

1093
00:37:56,240 --> 00:37:58,079
have this issue ever again with

1094
00:37:58,080 --> 00:37:59,699
any hardware wallet manufacturers?

1095
00:38:02,610 --> 00:38:04,589
It has happened, I want to say,

1096
00:38:04,590 --> 00:38:06,469
a dozen times before, and I

1097
00:38:06,470 --> 00:38:08,449
fully expect it to happen again

1098
00:38:08,450 --> 00:38:09,450
now.

1099
00:38:12,210 --> 00:38:14,529
The good part

1100
00:38:14,530 --> 00:38:16,929
about having critical

1101
00:38:16,930 --> 00:38:18,969
security failures like

1102
00:38:18,970 --> 00:38:21,109
this is that

1103
00:38:21,110 --> 00:38:23,029
the result is that

1104
00:38:23,030 --> 00:38:24,929
there is a sort of shared

1105
00:38:24,930 --> 00:38:26,089
learning.

1106
00:38:26,090 --> 00:38:29,109
And that it will result in

1107
00:38:29,110 --> 00:38:31,589
increased protections and

1108
00:38:31,590 --> 00:38:34,189
better practices taken,

1109
00:38:34,190 --> 00:38:36,129
but not by everyone.

1110
00:38:36,130 --> 00:38:37,929
And so, you know, the.

1111
00:38:38,940 --> 00:38:41,159
Various classes of security

1112
00:38:41,160 --> 00:38:43,659
failures tend to happen over

1113
00:38:43,660 --> 00:38:46,239
and over and over again because

1114
00:38:46,240 --> 00:38:48,759
even though our collective

1115
00:38:48,760 --> 00:38:50,539
understanding and consciousness

1116
00:38:50,540 --> 00:38:52,439
absorbs them and learns from

1117
00:38:52,440 --> 00:38:54,299
them, there will always be

1118
00:38:54,300 --> 00:38:56,259
new entrants in the future who come

1119
00:38:56,260 --> 00:38:58,619
in and don't learn the

1120
00:38:58,620 --> 00:39:00,599
hard lessons of history and end up

1121
00:39:00,600 --> 00:39:01,600
repeating them.

1122
00:39:02,640 --> 00:39:04,519
So, there's trade-offs

1123
00:39:04,520 --> 00:39:06,139
to all of this, good and bad.

1124
00:39:06,140 --> 00:39:07,340
And

1125
00:39:08,810 --> 00:39:10,789
I'm somewhat hopeful

1126
00:39:10,790 --> 00:39:13,589
that at least the

1127
00:39:13,590 --> 00:39:15,469
ecosystem-wide impact will

1128
00:39:15,470 --> 00:39:17,469
be less severe than what we've

1129
00:39:17,470 --> 00:39:20,149
seen from centralized exchange

1130
00:39:20,150 --> 00:39:22,169
providers that have gone bust

1131
00:39:22,170 --> 00:39:24,049
for various reasons.

1132
00:39:24,050 --> 00:39:25,869
I think

1133
00:39:25,870 --> 00:39:27,749
one nice

1134
00:39:27,750 --> 00:39:29,509
thing about this is that it is

1135
00:39:29,510 --> 00:39:30,890
somewhat distributed.

1136
00:39:32,590 --> 00:39:35,089
I don't know exactly the

1137
00:39:35,090 --> 00:39:37,229
size of Cold Cards

1138
00:39:37,230 --> 00:39:38,989
business and how many people bought

1139
00:39:38,990 --> 00:39:40,809
from them, but I think that

1140
00:39:40,810 --> 00:39:42,749
they are a lot smaller than

1141
00:39:42,750 --> 00:39:44,789
like Trezor, Ledger, whatever.

1142
00:39:44,790 --> 00:39:47,149
So on orders of magnitude impact,

1143
00:39:47,150 --> 00:39:48,229
it would have been far worse.

1144
00:39:48,230 --> 00:39:50,149
I think if this had occurred in

1145
00:39:50,150 --> 00:39:51,709
Trezr or Ledger.

1146
00:39:51,710 --> 00:39:52,829
But, and this is something that

1147
00:39:52,830 --> 00:39:54,789
we'll probably end up talking about

1148
00:39:54,790 --> 00:39:56,989
a lot in a variety of

1149
00:39:56,990 --> 00:39:58,899
ways, but

1150
00:39:58,900 --> 00:40:00,859
the size of the

1151
00:40:00,860 --> 00:40:02,959
organization and the project is

1152
00:40:02,960 --> 00:40:05,319
also somewhat

1153
00:40:05,320 --> 00:40:07,419
correlated to the security

1154
00:40:07,420 --> 00:40:08,459
of the project.

1155
00:40:08,460 --> 00:40:10,359
And so I think there's a strong

1156
00:40:10,360 --> 00:40:12,299
argument to be made that one

1157
00:40:12,300 --> 00:40:14,359
of the reasons why this

1158
00:40:14,360 --> 00:40:16,279
happened and went undetected for

1159
00:40:16,280 --> 00:40:18,499
so long was because

1160
00:40:18,500 --> 00:40:20,519
CoinKite is one of the smaller

1161
00:40:20,520 --> 00:40:21,919
hardware vendors.

1162
00:40:21,920 --> 00:40:23,639
And so, you know, didn't have the

1163
00:40:23,640 --> 00:40:24,640
same level of.

1164
00:40:25,560 --> 00:40:27,659
Eyeballs and other resources

1165
00:40:27,660 --> 00:40:30,019
focused on

1166
00:40:30,020 --> 00:40:31,499
auditing the code.

1167
00:40:31,500 --> 00:40:33,359
Now, we

1168
00:40:33,360 --> 00:40:35,399
know that people have looked

1169
00:40:35,400 --> 00:40:37,999
at the code over the years and

1170
00:40:38,000 --> 00:40:40,659
this particular bug was

1171
00:40:40,660 --> 00:40:42,159
fairly obscure.

1172
00:40:42,160 --> 00:40:44,219
It was not obvious I think just from

1173
00:40:44,220 --> 00:40:46,099
looking at the line

1174
00:40:46,100 --> 00:40:47,539
of code that was calling the random

1175
00:40:47,540 --> 00:40:48,939
number generator. And that's why it

1176
00:40:48,940 --> 00:40:51,019
was missed for so long.

1177
00:40:51,020 --> 00:40:52,639
And yes, I think AI

1178
00:40:54,129 --> 00:40:56,589
advancements have led

1179
00:40:56,590 --> 00:40:58,629
to this being found when

1180
00:40:58,630 --> 00:40:59,889
it was found.

1181
00:40:59,890 --> 00:41:01,429
Could it have been found earlier?

1182
00:41:01,430 --> 00:41:02,649
Sure.

1183
00:41:02,650 --> 00:41:04,529
You could probably play

1184
00:41:04,530 --> 00:41:06,669
coulda, shoulda, woulda

1185
00:41:06,670 --> 00:41:08,649
all day long,

1186
00:41:08,650 --> 00:41:10,589
but ultimately there is the

1187
00:41:10,590 --> 00:41:12,709
fact that advancements

1188
00:41:12,710 --> 00:41:14,729
in large language models are

1189
00:41:14,730 --> 00:41:16,369
drastically changing the security

1190
00:41:16,370 --> 00:41:18,610
landscape of everything.

1191
00:41:19,810 --> 00:41:21,729
This is something that

1192
00:41:21,730 --> 00:41:23,419
has been concerning.

1193
00:41:23,420 --> 00:41:25,079
To those of us in the security

1194
00:41:25,080 --> 00:41:26,979
community, really for the past

1195
00:41:26,980 --> 00:41:28,939
year, because we've been

1196
00:41:28,940 --> 00:41:31,779
seeing obscure

1197
00:41:31,780 --> 00:41:33,999
hidden vulnerabilities

1198
00:41:34,000 --> 00:41:36,159
in a variety of different popular

1199
00:41:36,160 --> 00:41:38,259
code bases, getting found,

1200
00:41:38,260 --> 00:41:39,339
some of which had been there for

1201
00:41:39,340 --> 00:41:41,319
five, 10,

1202
00:41:41,320 --> 00:41:44,199
even 15 or 20 years,

1203
00:41:44,200 --> 00:41:46,169
simply because...

1204
00:41:46,170 --> 00:41:48,289
You know, being open source is

1205
00:41:48,290 --> 00:41:49,429
not panacea.

1206
00:41:49,430 --> 00:41:51,549
It is not a guarantee that all

1207
00:41:51,550 --> 00:41:53,789
bugs will be found.

1208
00:41:53,790 --> 00:41:55,729
It's a combination

1209
00:41:55,730 --> 00:41:57,929
of being open-source

1210
00:41:57,930 --> 00:42:00,069
plus the level of scrutiny

1211
00:42:00,070 --> 00:42:02,429
and resources that are applied

1212
00:42:02,430 --> 00:42:04,129
upon that code.

1213
00:42:04,130 --> 00:42:06,289
And it's just that the LLM stuff

1214
00:42:06,290 --> 00:42:08,189
is it's

1215
00:42:08,190 --> 00:42:10,329
reducing the resource requirements

1216
00:42:10,330 --> 00:42:12,189
of applying level of screening

1217
00:42:12,190 --> 00:42:13,869
by orders of magnitude.

1218
00:42:13,870 --> 00:42:15,769
And, you know,

1219
00:42:15,770 --> 00:42:17,689
the the models that are

1220
00:42:17,690 --> 00:42:20,029
coming out on a monthly basis are

1221
00:42:20,030 --> 00:42:22,509
improving so

1222
00:42:22,510 --> 00:42:24,469
much that

1223
00:42:24,470 --> 00:42:27,189
it's

1224
00:42:27,190 --> 00:42:29,189
hard for the defenders to keep up

1225
00:42:29,190 --> 00:42:31,509
with the attackers because the

1226
00:42:31,510 --> 00:42:33,629
tool sets are improving for

1227
00:42:33,630 --> 00:42:35,409
both of them, but I think the

1228
00:42:35,410 --> 00:42:37,709
attackers are more motivated to

1229
00:42:37,710 --> 00:42:39,690
take that first mover advantage.

1230
00:42:41,010 --> 00:42:42,589
Yeah, I mean, it's been interesting

1231
00:42:42,590 --> 00:42:44,669
to see like cold kites,

1232
00:42:44,670 --> 00:42:46,469
I think, in one of the earlier sort

1233
00:42:46,470 --> 00:42:48,389
of announcements suggested

1234
00:42:48,390 --> 00:42:51,309
that it might be, you know,

1235
00:42:51,310 --> 00:42:53,329
AI related, obviously, we've

1236
00:42:53,330 --> 00:42:54,549
got these frontier models and

1237
00:42:54,550 --> 00:42:55,789
they're so powerful now that they

1238
00:42:55,790 --> 00:42:57,549
can go and search hard for these

1239
00:42:57,550 --> 00:42:59,469
bugs and then suggest ways to

1240
00:42:59,470 --> 00:43:00,469
exploit them.

1241
00:43:00,470 --> 00:43:02,209
But no one has really come out and

1242
00:43:02,210 --> 00:43:04,269
said like, this is probably how

1243
00:43:04,270 --> 00:43:05,809
it happened and this is who's doing

1244
00:43:05,810 --> 00:43:07,789
it. Our industry has obviously

1245
00:43:07,790 --> 00:43:09,650
been plagued by DPRK.

1246
00:43:10,690 --> 00:43:12,969
Hackers in the last few years,

1247
00:43:12,970 --> 00:43:14,849
would you expect sometime in the

1248
00:43:14,850 --> 00:43:16,069
next few months for us to actually

1249
00:43:16,070 --> 00:43:17,909
have a decent post-mortem that does

1250
00:43:17,910 --> 00:43:19,829
point a finger at someone with

1251
00:43:19,830 --> 00:43:21,709
the on-chain analysis that

1252
00:43:21,710 --> 00:43:22,710
we have today.

1253
00:43:25,480 --> 00:43:27,539
It's complicated because

1254
00:43:27,540 --> 00:43:29,819
obviously even though there was one

1255
00:43:29,820 --> 00:43:32,699
actor who was the first,

1256
00:43:32,700 --> 00:43:34,039
there's probably at least half a

1257
00:43:34,040 --> 00:43:35,919
dozen and like the number of

1258
00:43:35,920 --> 00:43:38,199
people who will be following

1259
00:43:38,200 --> 00:43:40,119
their footsteps and trying to

1260
00:43:40,120 --> 00:43:41,979
iteratively improve upon the

1261
00:43:41,980 --> 00:43:43,939
sophistication of the,

1262
00:43:43,940 --> 00:43:46,459
we'll call it search for

1263
00:43:46,460 --> 00:43:47,460
these keys,

1264
00:43:48,660 --> 00:43:49,879
that's ongoing.

1265
00:43:49,880 --> 00:43:52,779
And it

1266
00:43:52,780 --> 00:43:54,839
does sound like

1267
00:43:54,840 --> 00:43:57,879
the first attacker was

1268
00:43:57,880 --> 00:43:59,779
probably pretty naive

1269
00:43:59,780 --> 00:44:02,099
and didn't

1270
00:44:02,100 --> 00:44:03,979
fully understand the

1271
00:44:03,980 --> 00:44:06,099
intricacies of the protocol.

1272
00:44:06,100 --> 00:44:07,979
And we're probably doing all of

1273
00:44:07,980 --> 00:44:10,419
this by just like chaining LLM

1274
00:44:10,420 --> 00:44:11,499
prompts together.

1275
00:44:13,040 --> 00:44:14,999
And so that was actually, as

1276
00:44:15,000 --> 00:44:16,399
I mentioned at the very beginning,

1277
00:44:16,400 --> 00:44:17,599
one of the things that was baffling

1278
00:44:17,600 --> 00:44:19,499
to me was that the first person who

1279
00:44:19,500 --> 00:44:20,999
contacted me only had a partial

1280
00:44:21,000 --> 00:44:22,079
loss.

1281
00:44:22,080 --> 00:44:23,999
And it appears that the

1282
00:44:24,000 --> 00:44:26,419
reason for that is not because

1283
00:44:26,420 --> 00:44:28,599
the attacker

1284
00:44:28,600 --> 00:44:30,479
had only compromised some of the

1285
00:44:30,480 --> 00:44:32,859
keys, but rather the attacker was

1286
00:44:32,860 --> 00:44:35,359
unsophisticated in actually

1287
00:44:35,360 --> 00:44:37,659
doing their wallet

1288
00:44:37,660 --> 00:44:39,719
derivation and address

1289
00:44:39,720 --> 00:44:42,319
and funds search on Shane.

1290
00:44:42,320 --> 00:44:44,379
So they probably naively

1291
00:44:44,380 --> 00:44:46,339
had some LLM do

1292
00:44:46,340 --> 00:44:48,359
a fairly shallow search

1293
00:44:48,360 --> 00:44:50,239
for funds related to these

1294
00:44:50,240 --> 00:44:51,820
keys, and they were hitting various.

1295
00:44:53,410 --> 00:44:55,349
Roadblocks that were causing them to

1296
00:44:55,350 --> 00:44:57,169
stop searching the key

1297
00:44:57,170 --> 00:44:59,549
space of a given wallet, even after

1298
00:44:59,550 --> 00:45:00,889
all of the keys had theoretically

1299
00:45:00,890 --> 00:45:02,229
been compromised, you still have to

1300
00:45:02,230 --> 00:45:04,829
go generate

1301
00:45:04,830 --> 00:45:07,069
and do a search basically

1302
00:45:07,070 --> 00:45:08,849
in the UTXO set to figure out which

1303
00:45:08,850 --> 00:45:09,850
ones are available.

1304
00:45:11,290 --> 00:45:13,109
So yeah, I mean, that first one

1305
00:45:13,110 --> 00:45:14,969
or possibly few may have been more

1306
00:45:14,970 --> 00:45:17,849
naive and left enough of

1307
00:45:17,850 --> 00:45:19,929
a bread come trail that they might

1308
00:45:19,930 --> 00:45:22,259
get caught, but I expect that.

1309
00:45:22,260 --> 00:45:24,139
The later attackers are

1310
00:45:24,140 --> 00:45:25,579
just going to be more and more

1311
00:45:25,580 --> 00:45:26,580
sophisticated.

1312
00:45:28,620 --> 00:45:30,439
Lop, obviously, your website's a

1313
00:45:30,440 --> 00:45:31,739
wealth of information that I often

1314
00:45:31,740 --> 00:45:33,299
go to for reading and

1315
00:45:33,300 --> 00:45:34,699
recommendations on different things,

1316
00:45:34,700 --> 00:45:36,219
and obviously, you've got a whole

1317
00:45:36,220 --> 00:45:38,339
section dedicated to wallet

1318
00:45:38,340 --> 00:45:39,340
basics,

1319
00:45:40,520 --> 00:45:42,279
different hardware, manufacturers

1320
00:45:42,280 --> 00:45:44,319
that you would recommend,

1321
00:45:44,320 --> 00:45:45,819
et cetera.

1322
00:45:45,820 --> 00:45:47,139
The industry has always said that

1323
00:45:47,140 --> 00:45:48,779
self-custody is how you should do

1324
00:45:48,780 --> 00:45:49,999
Bitcoin, right? You know, you're

1325
00:45:50,000 --> 00:45:51,780
supposed to be banking yourself.

1326
00:45:53,090 --> 00:45:54,509
After this incident has happened,

1327
00:45:54,510 --> 00:45:55,609
you've been around a long time, so

1328
00:45:55,610 --> 00:45:56,449
you've seen worse.

1329
00:45:56,450 --> 00:45:57,929
I mean, I would say you've see worse

1330
00:45:57,930 --> 00:45:59,269
than this, but nothing like this has

1331
00:45:59,270 --> 00:46:02,029
ever really happened up until now.

1332
00:46:02,030 --> 00:46:03,529
How do you feel about, you know,

1333
00:46:03,530 --> 00:46:05,129
like self custody as it stands right

1334
00:46:05,130 --> 00:46:06,989
now, having seen this play out?

1335
00:46:06,990 --> 00:46:08,209
And do you think that we're getting

1336
00:46:08,210 --> 00:46:10,069
to a point where self custody really

1337
00:46:10,070 --> 00:46:11,349
isn't for anyone?

1338
00:46:11,350 --> 00:46:12,529
I don't wanna put any words in your

1339
00:46:12,530 --> 00:46:14,169
mouth. I just wanna get a sense of

1340
00:46:14,170 --> 00:46:15,649
where your head is at, just

1341
00:46:15,650 --> 00:46:18,189
considering the severity of this

1342
00:46:18,190 --> 00:46:19,649
exploits and what it might mean for

1343
00:46:19,650 --> 00:46:20,730
self custody going forward.

1344
00:46:23,120 --> 00:46:25,039
So although I don't think we've

1345
00:46:25,040 --> 00:46:26,619
had random number generator

1346
00:46:26,620 --> 00:46:29,439
vulnerabilities in hardware devices,

1347
00:46:29,440 --> 00:46:31,879
there have been many vulnerabilities

1348
00:46:31,880 --> 00:46:34,059
found in hardware and devices over

1349
00:46:34,060 --> 00:46:35,060
the years.

1350
00:46:36,100 --> 00:46:38,899
And there have many software

1351
00:46:38,900 --> 00:46:41,439
wallets and other,

1352
00:46:41,440 --> 00:46:43,279
I would say even more convenient

1353
00:46:43,280 --> 00:46:45,099
self custody

1354
00:46:45,100 --> 00:46:47,079
products that have

1355
00:46:47,080 --> 00:46:48,279
had random number generators

1356
00:46:48,280 --> 00:46:49,280
failures.

1357
00:46:51,730 --> 00:46:54,249
Self-custody is

1358
00:46:54,250 --> 00:46:56,369
for anyone

1359
00:46:56,370 --> 00:46:58,249
who is willing to

1360
00:46:58,250 --> 00:47:00,629
take on the responsibility that

1361
00:47:00,630 --> 00:47:01,630
comes with it.

1362
00:47:02,570 --> 00:47:04,509
So, you know, with great

1363
00:47:04,510 --> 00:47:06,449
power comes great responsibility.

1364
00:47:06,450 --> 00:47:08,329
This is true for self custody.

1365
00:47:08,330 --> 00:47:10,309
It greatly empowers the

1366
00:47:10,310 --> 00:47:11,929
end user.

1367
00:47:11,930 --> 00:47:13,789
And I think that

1368
00:47:13,790 --> 00:47:14,909
it's...

1369
00:47:16,470 --> 00:47:18,809
It can be a bit tricky,

1370
00:47:18,810 --> 00:47:20,769
like it may be too

1371
00:47:20,770 --> 00:47:22,609
easy for people to get set up

1372
00:47:22,610 --> 00:47:23,489
in self custody.

1373
00:47:23,490 --> 00:47:24,509
I know a lot of people are saying,

1374
00:47:24,510 --> 00:47:25,729
oh, self custody is hard.

1375
00:47:25,730 --> 00:47:27,549
Well, in order to

1376
00:47:27,550 --> 00:47:29,469
have a robust self custody with

1377
00:47:29,470 --> 00:47:31,369
no single points of failure,

1378
00:47:31,370 --> 00:47:32,969
yes, that does take some work.

1379
00:47:32,970 --> 00:47:35,109
And that's what I've spent over

1380
00:47:35,110 --> 00:47:36,949
a decade doing myself

1381
00:47:36,950 --> 00:47:39,009
is trying to make that easier for

1382
00:47:39,010 --> 00:47:40,009
people.

1383
00:47:40,010 --> 00:47:41,359
But...

1384
00:47:41,360 --> 00:47:43,339
It's deceptively simple to just

1385
00:47:43,340 --> 00:47:45,539
download a Bitcoin wallet,

1386
00:47:45,540 --> 00:47:47,059
a software wallet, and run it and

1387
00:47:47,060 --> 00:47:49,219
send money to it, and really have no

1388
00:47:49,220 --> 00:47:51,599
idea of all of the implications

1389
00:47:51,600 --> 00:47:53,219
behind that.

1390
00:47:53,220 --> 00:47:54,220
Um,

1391
00:47:55,120 --> 00:47:57,359
I think when one thing

1392
00:47:57,360 --> 00:47:58,839
that might.

1393
00:48:01,130 --> 00:48:02,969
Need further conversation

1394
00:48:02,970 --> 00:48:05,049
is, and

1395
00:48:05,050 --> 00:48:06,349
you were saying it earlier and I

1396
00:48:06,350 --> 00:48:07,549
think it's gonna be coming up a lot

1397
00:48:07,550 --> 00:48:09,729
in these discussions,

1398
00:48:09,730 --> 00:48:11,949
is the concept of don't trust,

1399
00:48:11,950 --> 00:48:12,950
verify.

1400
00:48:15,090 --> 00:48:16,669
I mean, obviously I think that

1401
00:48:16,670 --> 00:48:19,089
that's a good mantra and

1402
00:48:19,090 --> 00:48:21,049
that one should do

1403
00:48:21,050 --> 00:48:22,989
that whenever one can,

1404
00:48:22,990 --> 00:48:25,289
but you have to understand that

1405
00:48:25,290 --> 00:48:27,939
verification of

1406
00:48:27,940 --> 00:48:29,819
complex software and hardware

1407
00:48:29,820 --> 00:48:31,679
is simply not feasible

1408
00:48:31,680 --> 00:48:34,119
for 99.9%

1409
00:48:34,120 --> 00:48:35,819
of the population.

1410
00:48:35,820 --> 00:48:37,819
And I've had this idea floating

1411
00:48:37,820 --> 00:48:39,280
around in my head for a few years, I

1412
00:48:39,281 --> 00:48:41,099
actually have like a

1413
00:48:41,100 --> 00:48:42,859
draft blog post I've been working on

1414
00:48:42,860 --> 00:48:43,939
regarding it where,

1415
00:48:45,740 --> 00:48:47,859
ultimately what happens is because

1416
00:48:47,860 --> 00:48:49,899
it's just not feasible to

1417
00:48:49,900 --> 00:48:52,579
verify a lot of these things,

1418
00:48:52,580 --> 00:48:54,439
what we end up doing is

1419
00:48:54,440 --> 00:48:56,099
we end trusting.

1420
00:48:56,100 --> 00:48:58,199
Someone that that we suspect

1421
00:48:58,200 --> 00:48:59,959
has verified it, right?

1422
00:48:59,960 --> 00:49:02,339
And so you do end up with

1423
00:49:02,340 --> 00:49:04,879
various layers of trust,

1424
00:49:04,880 --> 00:49:06,199
which is understandable.

1425
00:49:07,600 --> 00:49:10,259
This is also in

1426
00:49:10,260 --> 00:49:12,179
a way how human civilization

1427
00:49:12,180 --> 00:49:14,299
works, like if you think of like

1428
00:49:14,300 --> 00:49:16,539
specialization of tasks, right.

1429
00:49:16,540 --> 00:49:18,579
So, you know, I think there

1430
00:49:18,580 --> 00:49:20,239
are a lot of people out there who

1431
00:49:20,240 --> 00:49:22,419
trust the various security experts

1432
00:49:22,420 --> 00:49:23,469
in the space.

1433
00:49:23,470 --> 00:49:25,589
Or just influencers in the space,

1434
00:49:25,590 --> 00:49:27,189
many of whom are gonna be getting a

1435
00:49:27,190 --> 00:49:29,629
lot of backlash for recommending

1436
00:49:29,630 --> 00:49:32,249
products that

1437
00:49:32,250 --> 00:49:34,209
resulted in people losing money.

1438
00:49:37,050 --> 00:49:37,859
But.

1439
00:49:37,860 --> 00:49:39,919
Perhaps it will be

1440
00:49:39,920 --> 00:49:41,979
easier if we just stick to

1441
00:49:41,980 --> 00:49:43,919
the don't trust aspect

1442
00:49:43,920 --> 00:49:45,679
and understand that you probably

1443
00:49:45,680 --> 00:49:47,359
can't verify.

1444
00:49:47,360 --> 00:49:49,239
And once again, and you know, this

1445
00:49:49,240 --> 00:49:51,019
is obviously shilling my own

1446
00:49:51,020 --> 00:49:52,999
product, but like the entire point

1447
00:49:53,000 --> 00:49:55,179
of CASA is to not trust

1448
00:49:55,180 --> 00:49:57,459
any one thing.

1449
00:49:57,460 --> 00:50:00,239
Not trust any one hardware vendor,

1450
00:50:00,240 --> 00:50:02,019
not trust anyone piece of software,

1451
00:50:02,020 --> 00:50:03,859
not even trust CASA ourselves.

1452
00:50:03,860 --> 00:50:05,779
And the main point that

1453
00:50:05,780 --> 00:50:08,639
I have here is that...

1454
00:50:08,640 --> 00:50:10,559
I understand there's

1455
00:50:10,560 --> 00:50:12,619
going to be a lot of hate and

1456
00:50:12,620 --> 00:50:14,839
vitriol and sadness

1457
00:50:14,840 --> 00:50:17,759
and other emotions and stuff over

1458
00:50:17,760 --> 00:50:21,279
this mistake.

1459
00:50:21,280 --> 00:50:23,759
I don't believe that

1460
00:50:23,760 --> 00:50:25,679
there's any evidence to show that it

1461
00:50:25,680 --> 00:50:26,959
was intentional.

1462
00:50:26,960 --> 00:50:28,799
I do believe that this

1463
00:50:28,800 --> 00:50:30,579
was an honest mistake.

1464
00:50:31,680 --> 00:50:33,539
And there will be good points

1465
00:50:33,540 --> 00:50:35,099
of like, well, this was their one

1466
00:50:35,100 --> 00:50:36,899
job and they failed at their one job

1467
00:50:36,900 --> 00:50:38,589
and that's completely fair.

1468
00:50:38,590 --> 00:50:40,499
But from an

1469
00:50:40,500 --> 00:50:42,859
end user perspective,

1470
00:50:42,860 --> 00:50:44,879
I believe it's

1471
00:50:44,880 --> 00:50:47,119
only fair to understand

1472
00:50:47,120 --> 00:50:50,159
that everyone is fallible.

1473
00:50:50,160 --> 00:50:52,079
Every company out there

1474
00:50:52,080 --> 00:50:53,219
is fallable.

1475
00:50:54,760 --> 00:50:56,999
Yes, like CoinKite has

1476
00:50:57,000 --> 00:50:58,779
been a smaller company with less

1477
00:50:58,780 --> 00:51:00,839
resources and so maybe more prone

1478
00:51:00,840 --> 00:51:02,959
to fallibility than others,

1479
00:51:02,960 --> 00:51:05,279
but ultimately there is

1480
00:51:05,280 --> 00:51:07,139
no such thing as

1481
00:51:07,140 --> 00:51:09,219
perfect software.

1482
00:51:09,220 --> 00:51:11,119
Perfect hardware,

1483
00:51:11,120 --> 00:51:12,360
perfect project.

1484
00:51:13,880 --> 00:51:16,099
Even Bitcoin Core, which is arguably

1485
00:51:16,100 --> 00:51:17,919
like one of the

1486
00:51:17,920 --> 00:51:21,019
most well engineered,

1487
00:51:21,020 --> 00:51:23,059
strongest, robust

1488
00:51:23,060 --> 00:51:24,959
open source projects

1489
00:51:24,960 --> 00:51:26,779
out there has had

1490
00:51:26,780 --> 00:51:28,179
critical vulnerabilities.

1491
00:51:30,160 --> 00:51:31,679
There have been critical

1492
00:51:31,680 --> 00:51:33,499
vulnerabilities over

1493
00:51:33,500 --> 00:51:36,089
the past year found in

1494
00:51:36,090 --> 00:51:38,649
low-level libraries like SSL

1495
00:51:38,650 --> 00:51:40,669
and SSH, like

1496
00:51:40,670 --> 00:51:42,729
code that runs on

1497
00:51:42,730 --> 00:51:45,049
billions of devices around

1498
00:51:45,050 --> 00:51:47,069
the world, and you could argue

1499
00:51:47,070 --> 00:51:49,029
is responsible for

1500
00:51:49,030 --> 00:51:51,409
the security of

1501
00:51:51,410 --> 00:51:53,349
many, many, or orders of

1502
00:51:53,350 --> 00:51:55,229
magnitude more time than

1503
00:51:55,230 --> 00:51:57,030
the entire Bitcoin ecosystem.

1504
00:51:59,070 --> 00:52:00,929
Open source is not panacea.

1505
00:52:00,930 --> 00:52:02,569
Like there, and there is no such

1506
00:52:02,570 --> 00:52:03,689
thing as perfection.

1507
00:52:03,690 --> 00:52:05,059
And so.

1508
00:52:05,060 --> 00:52:06,899
This is why I believe

1509
00:52:06,900 --> 00:52:09,479
that instead of

1510
00:52:09,480 --> 00:52:11,379
saying, OK, we're going to

1511
00:52:11,380 --> 00:52:12,999
trust any one given piece of

1512
00:52:13,000 --> 00:52:14,539
software or hardware, because we're

1513
00:52:14,540 --> 00:52:16,439
going to just verify the hell out of

1514
00:52:16,440 --> 00:52:18,959
it, whenever possible,

1515
00:52:18,960 --> 00:52:21,619
you should instead be

1516
00:52:21,620 --> 00:52:23,439
distributing your trust

1517
00:52:23,440 --> 00:52:25,779
across multiple different

1518
00:52:25,780 --> 00:52:27,260
diverse things.

1519
00:52:28,700 --> 00:52:30,839
Diversity is strength is one

1520
00:52:30,840 --> 00:52:33,869
of our mantras at CASA.

1521
00:52:33,870 --> 00:52:36,009
Because we expect that everything

1522
00:52:36,010 --> 00:52:37,429
can fail.

1523
00:52:37,430 --> 00:52:39,269
And so you want to put yourself in a

1524
00:52:39,270 --> 00:52:41,369
position where you

1525
00:52:41,370 --> 00:52:43,449
expect failure and you

1526
00:52:43,450 --> 00:52:45,509
are able to

1527
00:52:45,510 --> 00:52:47,469
recover from failure.

1528
00:52:47,470 --> 00:52:49,589
And so there's a million different

1529
00:52:49,590 --> 00:52:51,829
paths we could go down here.

1530
00:52:51,830 --> 00:52:53,129
And I was hearing and you were

1531
00:52:53,130 --> 00:52:54,509
talking about dice rolling a bit

1532
00:52:54,510 --> 00:52:55,510
earlier.

1533
00:52:56,310 --> 00:52:58,109
And yes, it's interesting that in

1534
00:52:58,110 --> 00:52:59,929
this particular case the

1535
00:52:59,930 --> 00:53:01,959
people who followed the

1536
00:53:01,960 --> 00:53:04,179
super high paranoid roll

1537
00:53:04,180 --> 00:53:06,779
your dice stuff were actually safe.

1538
00:53:06,780 --> 00:53:08,639
And, and I think it's,

1539
00:53:08,640 --> 00:53:10,779
it's somewhat ironic and unfortunate

1540
00:53:10,780 --> 00:53:13,179
that, um,

1541
00:53:13,180 --> 00:53:15,019
you know, I myself have

1542
00:53:15,020 --> 00:53:16,619
recommended rolling dice.

1543
00:53:16,620 --> 00:53:18,379
I have a blog post from three or

1544
00:53:18,380 --> 00:53:20,219
four years ago that I

1545
00:53:20,220 --> 00:53:22,199
entitled like how to generate

1546
00:53:22,200 --> 00:53:25,119
a secure seed phrase and,

1547
00:53:25,120 --> 00:53:27,019
and this is meant to be normie

1548
00:53:27,020 --> 00:53:29,779
friendly. And I said, basically

1549
00:53:29,780 --> 00:53:30,599
you want...

1550
00:53:30,600 --> 00:53:32,899
You want real-world entropy,

1551
00:53:32,900 --> 00:53:35,239
such as via playing

1552
00:53:35,240 --> 00:53:37,819
cards or throwing dice or whatever.

1553
00:53:37,820 --> 00:53:40,139
And in my guide,

1554
00:53:40,140 --> 00:53:42,119
I specifically say that the

1555
00:53:42,120 --> 00:53:43,999
easiest way to get a

1556
00:53:44,000 --> 00:53:46,019
super high security,

1557
00:53:46,020 --> 00:53:48,699
high entropy seed

1558
00:53:48,700 --> 00:53:50,719
phrase is to buy

1559
00:53:50,720 --> 00:53:52,679
casino-grade dice and buy a cold

1560
00:53:52,680 --> 00:53:54,599
card and use the built-in dice

1561
00:53:54,600 --> 00:53:55,579
rolling functionality.

1562
00:53:55,580 --> 00:53:57,599
And, as

1563
00:53:57,600 --> 00:53:59,539
far as we know, that claim

1564
00:53:59,540 --> 00:54:00,499
still ...

1565
00:54:00,500 --> 00:54:02,079
Holds true.

1566
00:54:02,080 --> 00:54:04,599
But I would note that,

1567
00:54:04,600 --> 00:54:06,559
you know, even that is

1568
00:54:06,560 --> 00:54:08,239
not a guarantee.

1569
00:54:08,240 --> 00:54:09,399
It's not panacea.

1570
00:54:09,400 --> 00:54:12,199
The reason for that is that

1571
00:54:12,200 --> 00:54:14,439
it would

1572
00:54:14,440 --> 00:54:16,759
have been and still is

1573
00:54:16,760 --> 00:54:18,619
entirely possible that there

1574
00:54:18,620 --> 00:54:20,479
is a bug in the

1575
00:54:20,480 --> 00:54:22,319
dice rolling algorithm

1576
00:54:22,320 --> 00:54:23,759
in the cold card.

1577
00:54:23,760 --> 00:54:25,019
Now, we...

1578
00:54:26,310 --> 00:54:27,689
As far as I know, that's not the

1579
00:54:27,690 --> 00:54:29,009
case. And people have actually been

1580
00:54:29,010 --> 00:54:30,529
looking at that over the past few

1581
00:54:30,530 --> 00:54:31,709
days.

1582
00:54:31,710 --> 00:54:33,269
And that code seems to be fine.

1583
00:54:33,270 --> 00:54:35,209
My point is, once

1584
00:54:35,210 --> 00:54:36,749
again, there is no guarantee that

1585
00:54:36,750 --> 00:54:38,349
that code is 100%.

1586
00:54:38,350 --> 00:54:40,069
There's no guarantee that any code

1587
00:54:40,070 --> 00:54:42,009
is a 100% perfect.

1588
00:54:42,010 --> 00:54:43,649
It just so happens that in this

1589
00:54:43,650 --> 00:54:44,969
case, that's the way to go.

1590
00:54:44,970 --> 00:54:45,970
Now,

1591
00:54:47,010 --> 00:54:48,989
going forward,

1592
00:54:48,990 --> 00:54:51,049
like if you wanted, for example,

1593
00:54:51,050 --> 00:54:53,569
to take kind of my perspective

1594
00:54:53,570 --> 00:54:55,429
of how you diversify your

1595
00:54:55,430 --> 00:54:57,009
trust since you can't really verify

1596
00:54:57,010 --> 00:54:59,069
any of the stuff, I

1597
00:54:59,070 --> 00:55:01,029
think the best way

1598
00:55:01,030 --> 00:55:02,869
to go forward would be to have

1599
00:55:02,870 --> 00:55:05,589
multiple different

1600
00:55:05,590 --> 00:55:08,549
means of

1601
00:55:08,550 --> 00:55:10,629
rolling the dice and

1602
00:55:10,630 --> 00:55:12,569
inputting that input and

1603
00:55:12,570 --> 00:55:14,389
and getting output from completely

1604
00:55:14,390 --> 00:55:16,679
different software and hardware.

1605
00:55:16,680 --> 00:55:19,139
That's how you verify that

1606
00:55:19,140 --> 00:55:22,059
it's unlikely to have bugs.

1607
00:55:22,060 --> 00:55:23,759
If you're doing the same thing

1608
00:55:23,760 --> 00:55:24,979
against multiple different software

1609
00:55:24,980 --> 00:55:26,799
and hardware, the

1610
00:55:26,800 --> 00:55:27,859
likelihood that they both

1611
00:55:27,860 --> 00:55:29,719
implemented the exact same bug

1612
00:55:29,720 --> 00:55:31,599
is very, very low.

1613
00:55:31,600 --> 00:55:33,519
And that's the same reason why

1614
00:55:33,520 --> 00:55:36,419
we are

1615
00:55:36,420 --> 00:55:38,259
highly supportive of

1616
00:55:38,260 --> 00:55:39,579
this idea of multi-vendor,

1617
00:55:39,580 --> 00:55:40,359
multi-SIG.

1618
00:55:40,360 --> 00:55:42,439
It's just because you expect

1619
00:55:42,440 --> 00:55:44,279
that anything can and will have

1620
00:55:44,280 --> 00:55:45,379
bugs, But.

1621
00:55:45,380 --> 00:55:47,299
As you diversify

1622
00:55:47,300 --> 00:55:49,519
your trust across many

1623
00:55:49,520 --> 00:55:51,599
different organizations,

1624
00:55:51,600 --> 00:55:53,419
then it's just a

1625
00:55:53,420 --> 00:55:55,239
sort of hedging game.

1626
00:55:56,480 --> 00:55:57,159
Yeah.

1627
00:55:57,160 --> 00:55:58,839
Talking about the hedging game,

1628
00:55:58,840 --> 00:56:00,739
obviously you've been at CUSA

1629
00:56:00,740 --> 00:56:02,359
a long time, you've been in the

1630
00:56:02,360 --> 00:56:04,519
security game a long, do you

1631
00:56:04,520 --> 00:56:05,959
think there needs to be a message

1632
00:56:05,960 --> 00:56:07,619
out there now considering everything

1633
00:56:07,620 --> 00:56:09,099
that has happened about A, best

1634
00:56:09,100 --> 00:56:10,979
practice for self custody, does

1635
00:56:10,980 --> 00:56:13,219
it need to be multi-seg,

1636
00:56:13,220 --> 00:56:15,159
where's your head at with regards to

1637
00:56:15,160 --> 00:56:16,160
that?

1638
00:56:17,920 --> 00:56:19,899
I mean, it's gonna be the same

1639
00:56:19,900 --> 00:56:21,899
thing that I've been preaching for

1640
00:56:21,900 --> 00:56:22,900
a decade.

1641
00:56:24,320 --> 00:56:26,359
It's very difficult to talk about

1642
00:56:26,360 --> 00:56:28,979
self-custody generally

1643
00:56:28,980 --> 00:56:31,159
because there are so many variables.

1644
00:56:32,740 --> 00:56:34,439
You know, I'm not gonna tell people

1645
00:56:34,440 --> 00:56:36,459
everybody needs to put

1646
00:56:36,460 --> 00:56:38,319
their Bitcoin into a

1647
00:56:38,320 --> 00:56:39,519
geographically distributed

1648
00:56:39,520 --> 00:56:41,739
multi-vendor multisig.

1649
00:56:41,740 --> 00:56:43,559
I believe that only makes

1650
00:56:43,560 --> 00:56:45,559
sense when you get to the point that

1651
00:56:45,560 --> 00:56:47,379
you have life

1652
00:56:47,380 --> 00:56:48,380
saving. Things.

1653
00:56:49,820 --> 00:56:51,639
When you have enough

1654
00:56:51,640 --> 00:56:53,719
Bitcoin that losing it

1655
00:56:53,720 --> 00:56:56,199
would be a sufficiently

1656
00:56:56,200 --> 00:56:58,039
large financial catastrophe

1657
00:56:58,040 --> 00:57:00,039
that you would never recover from

1658
00:57:00,040 --> 00:57:02,059
it, that's when you

1659
00:57:02,060 --> 00:57:04,079
should be getting to

1660
00:57:04,080 --> 00:57:05,919
the level of

1661
00:57:05,920 --> 00:57:07,619
paranoia where you're putting in

1662
00:57:07,620 --> 00:57:09,999
enough resources to ensure

1663
00:57:10,000 --> 00:57:12,039
you don't have single points

1664
00:57:12,040 --> 00:57:13,859
of failure, where if one thing goes

1665
00:57:13,860 --> 00:57:15,940
wrong, you have catastrophic loss.

1666
00:57:17,810 --> 00:57:20,029
But, you know, if you're just

1667
00:57:20,030 --> 00:57:21,529
playing around with Bitcoin, if you

1668
00:57:21,530 --> 00:57:22,809
only have a few hundred dollars

1669
00:57:22,810 --> 00:57:24,809
worth of Bitcoin, then yeah, leave

1670
00:57:24,810 --> 00:57:26,629
it on an exchange, put it

1671
00:57:26,630 --> 00:57:28,110
in a hot wallet.

1672
00:57:29,210 --> 00:57:31,129
If that blows up, then

1673
00:57:31,130 --> 00:57:33,009
it's not going to, you know,

1674
00:57:33,010 --> 00:57:34,569
negatively impact the rest of your

1675
00:57:34,570 --> 00:57:35,749
life.

1676
00:57:35,750 --> 00:57:37,789
If you have,

1677
00:57:37,790 --> 00:57:39,689
I generally say, if you have

1678
00:57:39,690 --> 00:57:41,589
thousands to tens of thousands

1679
00:57:41,590 --> 00:57:43,659
of dollars, That's

1680
00:57:43,660 --> 00:57:45,139
when it starts making sense to

1681
00:57:45,140 --> 00:57:46,939
invest a hundred or a few hundred

1682
00:57:46,940 --> 00:57:49,059
dollars into a hardware device

1683
00:57:49,060 --> 00:57:51,299
to get

1684
00:57:51,300 --> 00:57:53,559
your keys off the internet.

1685
00:57:53,560 --> 00:57:55,479
That protects you from

1686
00:57:55,480 --> 00:57:57,879
99% of

1687
00:57:57,880 --> 00:58:00,479
common loss

1688
00:58:00,480 --> 00:58:01,480
vectors.

1689
00:58:02,640 --> 00:58:04,699
But it's when you get beyond

1690
00:58:04,700 --> 00:58:06,519
that is when you should start

1691
00:58:06,520 --> 00:58:07,559
investing more.

1692
00:58:07,560 --> 00:58:09,419
I mean, generally just think of

1693
00:58:09,420 --> 00:58:10,469
it as.

1694
00:58:10,470 --> 00:58:12,669
And this is true for all types

1695
00:58:12,670 --> 00:58:14,269
of security.

1696
00:58:14,270 --> 00:58:16,289
The amount of

1697
00:58:16,290 --> 00:58:18,809
time, effort, and resources that

1698
00:58:18,810 --> 00:58:20,849
you invest into securing

1699
00:58:20,850 --> 00:58:23,069
a given asset should

1700
00:58:23,070 --> 00:58:25,189
be correlated to

1701
00:58:25,190 --> 00:58:26,730
the value of that asset.

1702
00:58:28,170 --> 00:58:30,249
So whether that's,

1703
00:58:30,250 --> 00:58:32,309
you're like insuring a house

1704
00:58:32,310 --> 00:58:35,149
or a car or

1705
00:58:35,150 --> 00:58:37,749
protecting the

1706
00:58:37,750 --> 00:58:40,469
physical security of your...

1707
00:58:40,470 --> 00:58:42,849
Family and loved ones,

1708
00:58:42,850 --> 00:58:44,689
you know, you should be spending

1709
00:58:44,690 --> 00:58:46,629
an amount on that that

1710
00:58:46,630 --> 00:58:48,489
is relative to how much you

1711
00:58:48,490 --> 00:58:50,529
care about it, how valuable it is,

1712
00:58:50,530 --> 00:58:52,090
and how irreplaceable it is.

1713
00:58:54,790 --> 00:58:56,829
I'm gonna quickly finish

1714
00:58:56,830 --> 00:58:58,609
off with a final question to you.

1715
00:58:58,610 --> 00:58:59,929
I mean, I know I think a lot of

1716
00:58:59,930 --> 00:59:01,309
people out there will naturally

1717
00:59:01,310 --> 00:59:03,149
gravitate towards some sort of

1718
00:59:03,150 --> 00:59:04,150
a custodian

1719
00:59:05,450 --> 00:59:07,929
or a custodeal solution.

1720
00:59:07,930 --> 00:59:10,109
CASA is still self-custodial,

1721
00:59:10,110 --> 00:59:11,209
right? Do you ever get people

1722
00:59:11,210 --> 00:59:12,069
knocking on your door and they're

1723
00:59:12,070 --> 00:59:13,149
like, I don't wanna use CASA because

1724
00:59:13,150 --> 00:59:14,549
I wanna be in full control and maybe

1725
00:59:14,550 --> 00:59:16,669
they don't understand that is

1726
00:59:16,670 --> 00:59:17,670
what it is.

1727
00:59:19,650 --> 00:59:21,469
And yeah, I don't know

1728
00:59:21,470 --> 00:59:23,249
if you want to just comment on the

1729
00:59:23,250 --> 00:59:25,249
other hardware wallets out there.

1730
00:59:25,250 --> 00:59:27,109
And I guess the coming together

1731
00:59:27,110 --> 00:59:28,789
of the industry, because it's been,

1732
00:59:28,790 --> 00:59:30,089
the last few weeks in Bitcoin has

1733
00:59:30,090 --> 00:59:31,470
been hugely divisive around, you

1734
00:59:31,471 --> 00:59:33,209
know, Bip 10, Bering in the

1735
00:59:33,210 --> 00:59:34,669
background. And this kind of seems

1736
00:59:34,670 --> 00:59:36,709
like everyone suddenly forgot about

1737
00:59:36,710 --> 00:59:38,549
that and got together and tried

1738
00:59:38,550 --> 00:59:39,949
to support each other as best as

1739
00:59:39,950 --> 00:59:41,769
possible as Bitcoin is first

1740
00:59:41,770 --> 00:59:42,770
and foremost.

1741
00:59:43,990 --> 00:59:45,829
Yeah. So, you know, the

1742
00:59:45,830 --> 00:59:47,869
way that I've framed

1743
00:59:47,870 --> 00:59:49,729
it is that,

1744
00:59:49,730 --> 00:59:51,709
um,

1745
00:59:51,710 --> 00:59:52,710
you

1746
01:00:00,051 --> 01:00:02,490
Diversity of

1747
01:00:02,491 --> 01:00:04,870
security properties and that

1748
01:00:04,871 --> 01:00:06,690
collectively added all together

1749
01:00:06,691 --> 01:00:08,670
uses this

1750
01:00:08,671 --> 01:00:11,030
really magic of

1751
01:00:11,031 --> 01:00:13,290
additive security properties that

1752
01:00:13,291 --> 01:00:15,290
puts you in a much more secure

1753
01:00:15,291 --> 01:00:17,270
position than you could ever be if

1754
01:00:17,271 --> 01:00:19,330
you only have one key of

1755
01:00:19,331 --> 01:00:20,331
any type.

1756
01:00:22,681 --> 01:00:25,920
But we're not reinventing

1757
01:00:25,921 --> 01:00:26,920
the wheel.

1758
01:00:26,921 --> 01:00:28,800
We're not doing anything novel from

1759
01:00:28,801 --> 01:00:30,680
a protocol point of view.

1760
01:00:30,681 --> 01:00:32,500
So it has always and will

1761
01:00:32,501 --> 01:00:35,060
always be possible for someone to

1762
01:00:35,061 --> 01:00:36,740
create their own multi-vendor

1763
01:00:36,741 --> 01:00:38,780
multi-SIG using nothing but

1764
01:00:38,781 --> 01:00:41,720
free open source software

1765
01:00:41,721 --> 01:00:43,300
that there's no subscription or

1766
01:00:43,301 --> 01:00:44,400
anything like that.

1767
01:00:44,401 --> 01:00:46,280
I think the reason you would

1768
01:00:46,281 --> 01:00:48,220
want to pay for

1769
01:00:48,221 --> 01:00:50,220
CASA Is

1770
01:00:50,221 --> 01:00:51,980
your your You're paying for

1771
01:00:51,981 --> 01:00:54,300
consulting, you're paying for

1772
01:00:54,301 --> 01:00:57,040
being given a

1773
01:00:57,041 --> 01:00:58,940
really a guided

1774
01:00:58,941 --> 01:01:01,460
software experience that

1775
01:01:01,461 --> 01:01:03,340
we

1776
01:01:03,341 --> 01:01:05,160
expressly do not support a lot

1777
01:01:05,161 --> 01:01:06,460
of advanced things that Bitcoin

1778
01:01:06,461 --> 01:01:08,140
supports because we think those are

1779
01:01:08,141 --> 01:01:09,960
foot guns that are too

1780
01:01:09,961 --> 01:01:11,800
easy to backfire.

1781
01:01:11,801 --> 01:01:14,060
So really, we're guiding

1782
01:01:14,061 --> 01:01:15,540
you through what we consider to be

1783
01:01:15,541 --> 01:01:16,950
the best practices.

1784
01:01:16,951 --> 01:01:19,030
And saving people a lot of time

1785
01:01:19,031 --> 01:01:21,610
from having to spend days,

1786
01:01:21,611 --> 01:01:23,630
weeks, months learning about all

1787
01:01:23,631 --> 01:01:25,090
of these best practices and

1788
01:01:25,091 --> 01:01:26,530
employing them.

1789
01:01:26,531 --> 01:01:28,990
And you're also, you're paying for

1790
01:01:28,991 --> 01:01:31,090
the CASA recovery key

1791
01:01:31,091 --> 01:01:33,270
that we safeguard and

1792
01:01:33,271 --> 01:01:35,990
build additional authorization

1793
01:01:35,991 --> 01:01:37,810
parameters around.

1794
01:01:37,811 --> 01:01:39,110
And there's plenty of other

1795
01:01:39,111 --> 01:01:40,890
functionality as well that

1796
01:01:40,891 --> 01:01:42,070
theoretically you could build

1797
01:01:42,071 --> 01:01:43,550
yourself, but we've been.

1798
01:01:43,551 --> 01:01:44,670
Building for eight years.

1799
01:01:44,671 --> 01:01:46,470
And so we've probably been able to

1800
01:01:46,471 --> 01:01:47,970
put a lot more resources into it

1801
01:01:47,971 --> 01:01:49,211
than any one person.

1802
01:01:50,871 --> 01:01:52,670
So optionality is great.

1803
01:01:52,671 --> 01:01:55,430
Just be aware that when

1804
01:01:55,431 --> 01:01:57,750
you are, if you're architecting

1805
01:01:57,751 --> 01:02:00,051
your self custody on your own,

1806
01:02:02,231 --> 01:02:04,610
it's problematic for

1807
01:02:04,611 --> 01:02:06,790
the same reason that these

1808
01:02:06,791 --> 01:02:08,390
other vulnerabilities have been

1809
01:02:08,391 --> 01:02:10,370
found, which like, as I

1810
01:02:10,371 --> 01:02:12,400
said, open source is not Patasia.

1811
01:02:12,401 --> 01:02:14,240
And it really comes down, I

1812
01:02:14,241 --> 01:02:15,980
think, to the principle of various

1813
01:02:15,981 --> 01:02:17,880
and numerous, which is strength

1814
01:02:17,881 --> 01:02:19,220
and numbers.

1815
01:02:19,221 --> 01:02:21,140
And when you're using CASA,

1816
01:02:21,141 --> 01:02:23,160
you're a system that a lot

1817
01:02:23,161 --> 01:02:25,260
of people are using that

1818
01:02:25,261 --> 01:02:27,080
we've had a lot

1819
01:02:27,081 --> 01:02:28,900
of engineering talent look at that

1820
01:02:28,901 --> 01:02:31,240
we get regular

1821
01:02:31,241 --> 01:02:33,380
security audits on.

1822
01:02:33,381 --> 01:02:35,300
And actually at this point, we have

1823
01:02:35,301 --> 01:02:37,120
our own AI pin test that

1824
01:02:37,121 --> 01:02:39,020
we're running all of the time.

1825
01:02:39,021 --> 01:02:39,950
So...

1826
01:02:39,951 --> 01:02:42,230
You're basically paying for

1827
01:02:42,231 --> 01:02:44,050
the expertise of a

1828
01:02:44,051 --> 01:02:46,370
security focused organization.

1829
01:02:46,371 --> 01:02:48,650
Whereas if you're doing it yourself,

1830
01:02:48,651 --> 01:02:50,370
it's highly likely that simply

1831
01:02:50,371 --> 01:02:52,090
because you're the only one looking

1832
01:02:52,091 --> 01:02:53,730
at it, you're just going to overlook

1833
01:02:53,731 --> 01:02:55,890
some potential vulnerabilities

1834
01:02:55,891 --> 01:02:57,650
and something over a long enough

1835
01:02:57,651 --> 01:02:59,171
time frame could go wrong.

1836
01:03:01,741 --> 01:03:03,560
Yeah, my final question to

1837
01:03:03,561 --> 01:03:05,560
you, I find this funny that,

1838
01:03:05,561 --> 01:03:07,340
you know, there was some influences

1839
01:03:07,341 --> 01:03:08,661
and stuff online being like, you

1840
01:03:08,662 --> 01:03:10,600
know, Bitcoin self custody is dead.

1841
01:03:10,601 --> 01:03:12,140
You're better off just, you know,

1842
01:03:12,141 --> 01:03:14,420
buying the Bitcoin ETFs.

1843
01:03:14,421 --> 01:03:16,120
Like my general take on that is

1844
01:03:16,121 --> 01:03:17,760
like, I fundamentally believe

1845
01:03:17,761 --> 01:03:19,500
Bitcoin is money and it will be

1846
01:03:19,501 --> 01:03:21,540
money that we all use

1847
01:03:21,541 --> 01:03:22,320
in the future.

1848
01:03:22,321 --> 01:03:23,840
And I want to own it and I want to

1849
01:03:23,841 --> 01:03:25,600
have total control of it.

1850
01:03:25,601 --> 01:03:26,740
I fundamentally believed that.

1851
01:03:26,741 --> 01:03:28,260
And I pushed back, you know, on

1852
01:03:28,261 --> 01:03:30,110
those assertions online, but Do

1853
01:03:30,111 --> 01:03:31,310
you think it's funny that people are

1854
01:03:31,311 --> 01:03:33,850
saying that online and is that just

1855
01:03:33,851 --> 01:03:35,590
completely misguided and we should

1856
01:03:35,591 --> 01:03:37,790
really still take Bitcoin

1857
01:03:37,791 --> 01:03:40,050
self-custody as the first prize and

1858
01:03:40,051 --> 01:03:41,910
really just be in control of

1859
01:03:41,911 --> 01:03:42,911
our own keys?

1860
01:03:45,371 --> 01:03:47,250
I mean the freedom of the protocol

1861
01:03:47,251 --> 01:03:49,150
is that people

1862
01:03:49,151 --> 01:03:51,490
can put

1863
01:03:51,491 --> 01:03:53,070
themselves into what security model

1864
01:03:53,071 --> 01:03:55,210
they want. They can choose whether

1865
01:03:55,211 --> 01:03:57,030
they prefer convenience

1866
01:03:57,031 --> 01:03:58,790
or security. That's really what this

1867
01:03:58,791 --> 01:04:01,490
all comes down to is that

1868
01:04:01,491 --> 01:04:03,590
if you want to

1869
01:04:03,591 --> 01:04:05,530
be empowered, if want

1870
01:04:05,531 --> 01:04:07,430
to actually

1871
01:04:07,431 --> 01:04:10,220
take full advantage of.

1872
01:04:10,221 --> 01:04:12,280
What I would say Bitcoin's

1873
01:04:12,281 --> 01:04:14,540
fundamental premise of not trusting

1874
01:04:14,541 --> 01:04:16,560
third parties is,

1875
01:04:16,561 --> 01:04:18,720
you gotta put in some work to do so.

1876
01:04:18,721 --> 01:04:21,020
If all you want is financial

1877
01:04:21,021 --> 01:04:22,880
exposure to Bitcoin as a

1878
01:04:22,881 --> 01:04:24,940
volatile asset, and

1879
01:04:24,941 --> 01:04:26,800
you don't

1880
01:04:26,801 --> 01:04:28,660
care too much about

1881
01:04:28,661 --> 01:04:30,720
the risks of

1882
01:04:30,721 --> 01:04:32,800
custodial failure, despite the

1883
01:04:32,801 --> 01:04:34,700
history of that happening in

1884
01:04:34,701 --> 01:04:37,040
the space, then you're

1885
01:04:37,041 --> 01:04:38,880
free to do that as well.

1886
01:04:38,881 --> 01:04:40,940
The the only thing that worries me

1887
01:04:40,941 --> 01:04:42,880
and the reason why I've spent the

1888
01:04:42,881 --> 01:04:45,280
past decade focused on self-custody

1889
01:04:45,281 --> 01:04:47,420
is That I do believe that

1890
01:04:47,421 --> 01:04:49,540
if we don't continue to iterate

1891
01:04:49,541 --> 01:04:51,360
on it and make it

1892
01:04:51,361 --> 01:04:53,260
easier for people make

1893
01:04:53,261 --> 01:04:55,720
people Believe

1894
01:04:55,721 --> 01:04:57,920
that they are capable

1895
01:04:57,921 --> 01:04:59,940
of taking on this Responsibility

1896
01:04:59,941 --> 01:05:01,820
then it will end

1897
01:05:01,821 --> 01:05:03,760
up being a niche thing that

1898
01:05:03,761 --> 01:05:05,820
only like super nerds

1899
01:05:05,821 --> 01:05:07,820
and paranoid folks are

1900
01:05:07,821 --> 01:05:09,880
going to end up bothering to do

1901
01:05:09,881 --> 01:05:12,300
and ultimately that is

1902
01:05:12,301 --> 01:05:15,340
arguably not great for

1903
01:05:15,341 --> 01:05:18,180
the fragility

1904
01:05:18,181 --> 01:05:20,300
of the overall ecosystem because

1905
01:05:20,301 --> 01:05:22,400
it would result in massive

1906
01:05:22,401 --> 01:05:24,180
amounts of Bitcoin being centralized

1907
01:05:24,181 --> 01:05:25,940
into a few trusted third parties and

1908
01:05:25,941 --> 01:05:27,360
then we're just recreating the old

1909
01:05:27,361 --> 01:05:28,780
system all over again.

1910
01:05:28,781 --> 01:05:29,990
Yep.

1911
01:05:29,991 --> 01:05:31,010
Completely agree with you.

1912
01:05:31,011 --> 01:05:32,150
I don't think I could have said it

1913
01:05:32,151 --> 01:05:33,330
better than that.

1914
01:05:33,331 --> 01:05:34,330
Jamison, thank you so much for

1915
01:05:34,331 --> 01:05:35,310
taking the time to chat to us.

1916
01:05:35,311 --> 01:05:36,770
I know you're probably busy on this

1917
01:05:36,771 --> 01:05:37,970
Monday with the amount of people

1918
01:05:37,971 --> 01:05:39,690
that are just looking to migrate

1919
01:05:39,691 --> 01:05:42,030
funds or cold card,

1920
01:05:42,031 --> 01:05:43,710
but salute you and every other

1921
01:05:43,711 --> 01:05:45,070
Bitcoin out there that's doing what

1922
01:05:45,071 --> 01:05:46,470
they can to help people in these

1923
01:05:46,471 --> 01:05:47,610
pretty tough times.

1924
01:05:48,731 --> 01:05:49,930
So Jamison we'll say goodbye to you.

1925
01:05:49,931 --> 01:05:52,370
Feel free to drop off the stream.

1926
01:05:52,371 --> 01:05:54,770
Thank you so for watching the

1927
01:05:54,771 --> 01:05:55,930
starting block today.

1928
01:05:55,931 --> 01:05:57,830
We broadcast live every day,

1929
01:05:57,831 --> 01:05:59,870
8.30 a.m. Eastern Time, 2.30

1930
01:05:59,871 --> 01:06:01,410
p.m., Central European Time, if

1931
01:06:01,411 --> 01:06:03,270
you're in Europe like I

1932
01:06:03,271 --> 01:06:04,490
am.

1933
01:06:04,491 --> 01:06:05,530
Please come back tomorrow.

1934
01:06:05,531 --> 01:06:07,390
We have Ophelia Snyder, who is

1935
01:06:07,391 --> 01:06:09,670
one of the co-founders of 21 shares.

1936
01:06:09,671 --> 01:06:11,150
She exited the company last year.

1937
01:06:11,151 --> 01:06:12,850
We're going to be chatting about the

1938
01:06:12,851 --> 01:06:14,690
institutionalization of Bitcoin and

1939
01:06:14,691 --> 01:06:15,850
many other cryptocurrencies.

1940
01:06:15,851 --> 01:06:17,990
So we'd love to have you join us

1941
01:06:17,991 --> 01:06:20,070
on the show tomorrow,

1942
01:06:20,071 --> 01:06:21,990
wherever you are around the world.

1943
01:06:21,991 --> 01:06:22,810
Look after yourself.

1944
01:06:22,811 --> 01:06:24,250
Look after your Bitcoin.

1945
01:06:24,251 --> 01:06:25,880
We'll be back tomorrow Until then,

1946
01:06:25,881 --> 01:06:26,881
bye-bye.