1
00:00:00,020 --> 00:00:01,220
Nobody wants to talk about it.

2
00:00:01,500 --> 00:00:02,380
Everyone's afraid of it.

3
00:00:02,600 --> 00:00:05,580
Everyone thinks that encryption is sacred and it shouldn't be compromised.

4
00:00:05,760 --> 00:00:06,900
And I 100% agree.

5
00:00:09,000 --> 00:00:11,040
Hello, everybody, and welcome to Techlore Talks.

6
00:00:11,140 --> 00:00:14,040
Today, I am really excited to invite back on my good friend,

7
00:00:14,220 --> 00:00:18,180
John Ozbay from Cryptee, to discuss a very sensitive topic.

8
00:00:18,300 --> 00:00:22,280
So I ask us all to make sure that we go into this in a proper mindset.

9
00:00:22,640 --> 00:00:25,020
We're going to discuss some exciting things along the way,

10
00:00:25,080 --> 00:00:27,880
like Cryptee announcing end-to-end encrypted photo sharing,

11
00:00:28,000 --> 00:00:30,940
which is something they've held off on shipping for a long time.

12
00:00:31,060 --> 00:00:35,340
They are an open source, end-to-end encrypted photo management and document management platform,

13
00:00:35,600 --> 00:00:37,480
and they have now finally shipped sharing.

14
00:00:37,680 --> 00:00:42,900
You might be wondering why that took so long for them to do, and this is overall what we talk about.

15
00:00:42,920 --> 00:00:46,140
We talk about their implementation, but we also talk about their hesitation,

16
00:00:46,540 --> 00:00:49,200
involving things like CSAM and other concerns they have.

17
00:00:49,360 --> 00:00:53,539
We're going to talk about the legal and ethical considerations while still balancing user privacy

18
00:00:53,540 --> 00:00:58,380
and some innovative solutions that John has created at Cryptee to try to balance these two things.

19
00:00:58,640 --> 00:01:01,900
In this interview, we're going to talk about something quite sensitive in the privacy community,

20
00:01:02,030 --> 00:01:07,340
which is the relationship between end-to-end encryption, as well as CSAM, which is child sexual abuse material.

21
00:01:07,560 --> 00:01:10,600
This is something that in the past has not gone too well in our communities,

22
00:01:10,900 --> 00:01:15,820
and I hope that this conversation can be a good starting point for us to really begin anew

23
00:01:16,200 --> 00:01:18,660
and actually have good conversations around this.

24
00:01:19,040 --> 00:01:23,480
I ask us all to be sensitive in the comments, as some people may be impacted by things like CSAM,

25
00:01:23,840 --> 00:01:30,060
And this is something that I wish all of us keep in mind when we think about privacy, encryption, and the other complexities in the world.

26
00:01:30,620 --> 00:01:32,540
Without further ado, let's get into the interview.

27
00:01:33,140 --> 00:01:37,500
All right, I am really happy to bring back on John Ozbay from Cryptee.

28
00:01:37,680 --> 00:01:40,520
John's been on here before. I'll let you properly introduce yourself in a second.

29
00:01:40,900 --> 00:01:47,040
But John, we had a phone call or you texted me a few weeks ago saying that you guys had a new release at Cryptee.

30
00:01:47,600 --> 00:01:58,520
And you had to do a lot of explaining to talk about the nuances of end-to-end encrypted photo sharing in context of chat control, in context of CSAM.

31
00:01:59,520 --> 00:02:04,360
And I think you said, oh, let me just explain this to you in a couple sentences.

32
00:02:05,060 --> 00:02:08,740
And I think it was like 10 minutes later, I was still just listening to you explain this.

33
00:02:10,060 --> 00:02:27,140
But it got me thinking, this is actually a really interesting topic because this was a common discussion on our forum back in the day before it was archived of people talking about how to balance CSAM, for those who don't know, child sexual abuse material, with end-to-end encryption, if that's a real concern at all.

34
00:02:27,170 --> 00:02:30,520
I know there's some takes online that that's just not really the provider's responsibility.

35
00:02:31,110 --> 00:02:35,840
But you actually run an end-to-end encrypted service, and you recently just released photo sharing.

36
00:02:36,220 --> 00:02:37,840
And so this was something you guys had to discuss.

37
00:02:38,240 --> 00:02:44,180
And I kind of just want to start by really laying out what we discussed and why we're here.

38
00:02:44,320 --> 00:02:48,300
And so I really wanted to open up this conversation as hopefully something for all of you to think about.

39
00:02:48,500 --> 00:02:50,280
I don't really know what to think going into this.

40
00:02:50,320 --> 00:02:51,500
I haven't made any conclusions.

41
00:02:51,660 --> 00:02:54,600
So I just want to hear what John did and maybe ask questions to see how things work.

42
00:02:55,140 --> 00:02:56,160
Is that a good starting point?

43
00:02:56,800 --> 00:02:57,680
Yeah, that's great.

44
00:02:57,940 --> 00:02:58,680
That's a great starting point.

45
00:02:58,960 --> 00:03:04,660
So do you want to just quickly reintroduce yourself for those of you people who might listen to you for the first time here?

46
00:03:04,960 --> 00:03:05,100
Yeah.

47
00:03:05,620 --> 00:03:06,420
Sounds good, of course.

48
00:03:06,900 --> 00:03:07,280
Hi, everyone.

49
00:03:07,680 --> 00:03:08,500
I'm John Ozbay.

50
00:03:08,580 --> 00:03:13,780
I'm the founder and CEO of a small end-to-end encrypted documents and photo storage service called Cryptee.

51
00:03:14,020 --> 00:03:17,060
We've been around for a very long time now, almost eight years.

52
00:03:17,590 --> 00:03:24,020
And for the longest time, we did not want to ship sharing for various legal reasons.

53
00:03:24,880 --> 00:03:28,220
And we decided that now might be the perfect time.

54
00:03:28,420 --> 00:03:32,240
And we've been sitting on this for almost five or six years now at this point.

55
00:03:32,680 --> 00:03:41,120
So it gives me great relief to say that we're finally ready to ship sharing for your photos and your videos.

56
00:03:41,540 --> 00:03:45,800
And a lot has happened over the course of the last four to five years.

57
00:03:45,950 --> 00:03:50,680
And we have to keep our code base up to date so that we can one day flip that switch and enable sharing.

58
00:03:50,960 --> 00:03:54,960
But finally, this week felt like the right week.

59
00:03:55,280 --> 00:03:59,300
And I texted Henry saying, yo, you're never going to believe what we're about to do.

60
00:03:59,940 --> 00:04:02,520
And Henry was kind enough to help me on this podcast.

61
00:04:02,820 --> 00:04:04,720
And so I'm very happy and grateful.

62
00:04:05,380 --> 00:04:05,580
Yeah.

63
00:04:06,440 --> 00:04:07,280
So do you mind expanding?

64
00:04:07,480 --> 00:04:09,880
You know, you started building this for five years ago,

65
00:04:10,140 --> 00:04:13,000
and you guys have been secretly, I guess, behind the scenes maintaining it.

66
00:04:13,200 --> 00:04:17,359
But what were the reasons back then, and even, I guess, as of a month ago, maybe,

67
00:04:18,260 --> 00:04:19,320
why were you not shipping it?

68
00:04:19,540 --> 00:04:21,500
Like, do you mind expanding on some of those reasons?

69
00:04:22,640 --> 00:04:26,940
Well, so we first had a long call and chat with our legal team.

70
00:04:27,260 --> 00:04:30,020
And this is, we're talking five, six years ago.

71
00:04:30,370 --> 00:04:32,520
And long story short, they basically said,

72
00:04:32,780 --> 00:04:34,740
there are a bunch of legal edge case scenarios

73
00:04:35,200 --> 00:04:37,360
that we should be extremely careful of

74
00:04:37,470 --> 00:04:40,340
if we are to do any end-to-end encrypt sharing

75
00:04:40,640 --> 00:04:42,760
and distribution of content in general.

76
00:04:43,130 --> 00:04:44,920
It doesn't have to be photographic content.

77
00:04:45,010 --> 00:04:47,320
It could also be like documents and whatnot.

78
00:04:47,880 --> 00:04:50,480
Essentially, we were just told, be extremely careful.

79
00:04:50,920 --> 00:04:54,160
And when we asked for details, like, okay,

80
00:04:54,220 --> 00:04:57,660
Just explain to us, run me through the scenario.

81
00:04:58,800 --> 00:05:02,400
One big scenario that they flagged, especially with photos and video sharing,

82
00:05:03,090 --> 00:05:07,220
and this is to put some perspective around pandemic era,

83
00:05:07,740 --> 00:05:12,220
2020 and politically everything was kind of difficult in the United States

84
00:05:12,230 --> 00:05:13,660
and around the world for that matter.

85
00:05:14,280 --> 00:05:17,880
Our team basically said, look, there is a really strange edge case scenario.

86
00:05:18,240 --> 00:05:21,599
Say, for example, someone's spreading CSAM on the internet

87
00:05:21,620 --> 00:05:25,240
and they have an account with Cryptee and they shared this with five other people.

88
00:05:25,980 --> 00:05:30,240
And let's say a law enforcement team from the United States catches one of the recipients

89
00:05:30,840 --> 00:05:34,520
and this person gives them their key and says, here it is.

90
00:05:35,100 --> 00:05:38,860
Now, the law enforcement has access to the content and they can see this content.

91
00:05:39,540 --> 00:05:43,080
But in the United States, by law, including the FBI,

92
00:05:43,250 --> 00:05:45,960
no one is allowed to send that key to someone else.

93
00:05:46,100 --> 00:05:52,680
The only party they're allowed to send this is the National Database of Missing and Exploited Children, NCMEC.

94
00:05:53,060 --> 00:05:55,480
That's the only legal party that they're allowed to send this to.

95
00:05:55,540 --> 00:06:00,000
So if FBI discovers one of these recipients and this key is in their hands,

96
00:06:00,580 --> 00:06:03,140
FBI legally cannot send us the key to say,

97
00:06:03,140 --> 00:06:06,200
hey, we've discovered CSAM on your platform, and here's the key.

98
00:06:06,800 --> 00:06:08,620
You can go verify it or something.

99
00:06:09,380 --> 00:06:10,860
Can you please take this content down?

100
00:06:11,260 --> 00:06:16,540
They cannot send it to us because that would constitute legally as sharing CSAM for them.

101
00:06:16,850 --> 00:06:17,960
So they are not allowed to do that.

102
00:06:17,990 --> 00:06:19,140
So they can't share the keys with us.

103
00:06:19,280 --> 00:06:23,680
And to be honest, I don't want to have to worry about employing a content management team

104
00:06:23,820 --> 00:06:29,140
that has to go through with the incredibly brittle task of having to look at the CSAM material,

105
00:06:30,560 --> 00:06:35,080
CSAM content in general, to be able to say, hey, this is indeed CSAM or this is not CSAM.

106
00:06:35,080 --> 00:06:36,340
And we cannot decide.

107
00:06:36,490 --> 00:06:39,080
We can't be the arbiters of that decision either.

108
00:06:40,020 --> 00:06:45,320
And NCMEC is basically, at least in the United States and in Northern America, the correct party to do that.

109
00:06:46,060 --> 00:06:53,340
So we thought this is a problem because if they cannot share this content with us and if we can't verify because everything is end-to-end encrypted,

110
00:06:54,020 --> 00:06:59,660
then how do we know that what they are asking to take down is in fact CSAM?

111
00:06:59,880 --> 00:07:09,240
What if it's police brutality that someone in the United States documented and the United States legal law enforcement is saying, hey, we want you to take this down?

112
00:07:09,440 --> 00:07:12,880
They have access to the recipients key, but not the sharing party's key.

113
00:07:13,280 --> 00:07:16,780
And if they're asking us, hey, we need you to take this down,

114
00:07:16,900 --> 00:07:18,920
we would need to somehow be able to verify this.

115
00:07:19,080 --> 00:07:20,420
Otherwise, we're taking their word.

116
00:07:21,040 --> 00:07:23,960
And it was a lose-lose scenario, because if we take it down

117
00:07:23,980 --> 00:07:27,160
and if it's police brutality, we're effectively backing down

118
00:07:27,200 --> 00:07:29,940
on our promise to do our best for reporters around the world,

119
00:07:30,160 --> 00:07:32,980
to do our best for war photographers around the world.

120
00:07:33,440 --> 00:07:36,560
But if we don't take it down, and if it is in fact CSAM,

121
00:07:36,600 --> 00:07:39,000
then we are hosting CSAM and we're all going to go to jail.

122
00:07:39,200 --> 00:07:40,000
And that's not great either.

123
00:07:40,690 --> 00:07:44,680
So we thought, okay, this is a bad lose-lose scenario

124
00:07:44,960 --> 00:07:47,440
that we need to think very thoroughly about

125
00:07:47,700 --> 00:07:49,620
and figure out how this is going to pan out.

126
00:07:50,360 --> 00:07:52,020
We don't want this to become a thing

127
00:07:52,280 --> 00:07:54,800
where we are doing the wrong thing

128
00:07:55,520 --> 00:07:57,380
and we're acting kind of like Silicon Valley

129
00:07:57,600 --> 00:08:00,120
and sort of like releasing our code fast

130
00:08:00,320 --> 00:08:02,640
and breaking things and breaking the society.

131
00:08:02,820 --> 00:08:05,240
We want to do the right responsible thing and take our time.

132
00:08:05,980 --> 00:08:07,840
So I thought, let's just take our time.

133
00:08:08,090 --> 00:08:08,420
Fuck it.

134
00:08:08,740 --> 00:08:09,640
We have nothing to lose.

135
00:08:10,200 --> 00:08:12,060
We just need to do the right thing here.

136
00:08:12,200 --> 00:08:15,720
And I didn't want us to unleash something onto the internet that we're not comfortable with.

137
00:08:16,580 --> 00:08:21,800
And over the course of the years, what happened is, I mean, fast forward now, multiple years in,

138
00:08:22,220 --> 00:08:25,220
I started going to Brussels a lot more often for various reasons,

139
00:08:25,480 --> 00:08:29,720
which you've mainly discussed about in the context of Apple.

140
00:08:29,900 --> 00:08:35,900
But there's also other topics like browser choice and also European digital rights and digital sovereignty.

141
00:08:36,180 --> 00:08:48,740
So I ended up doing a lot of trips to Brussels and ended up talking to a lot of these regulators from the European Commission, from the European Parliament, also to civil rights organizations, to human rights organizations, to women's rights organizations.

142
00:08:49,160 --> 00:08:58,060
And the more I ended up making these trips and talking to these people, the more I understood and had a better picture of what everyone is actually afraid of.

143
00:08:58,300 --> 00:09:03,100
And I feel like CSAM is one of those, I don't want to say difficult topics, but one of those

144
00:09:04,500 --> 00:09:08,360
thorny topics, pun intended, that nobody really wants to talk about.

145
00:09:09,200 --> 00:09:13,620
And especially coming from a company like Cryptee, where we're talking about end-to-end encryption

146
00:09:13,770 --> 00:09:16,280
and making sure that everyone's data stays safe.

147
00:09:17,220 --> 00:09:21,060
The automatic question that a lot of people ask is, how do you make sure people don't spread

148
00:09:21,180 --> 00:09:22,140
bad shit on your platform?

149
00:09:22,310 --> 00:09:23,140
And they're right.

150
00:09:23,839 --> 00:09:25,400
That's perfectly fair to ask.

151
00:09:26,120 --> 00:09:31,840
And I thought that it's very important that someone just sits down and talks to the regulators

152
00:09:32,380 --> 00:09:36,640
and talks to these human rights organizations and talks to these civil rights organizations

153
00:09:36,900 --> 00:09:38,420
and psychologists and sociologists.

154
00:09:38,440 --> 00:09:41,200
Just chat control especially is very, very big.

155
00:09:41,360 --> 00:09:42,100
Yeah, especially chat control.

156
00:09:42,620 --> 00:09:42,780
Exactly.

157
00:09:43,100 --> 00:09:48,200
And chat control has been sort of like this dark cloud looming all over Europe for a very

158
00:09:48,220 --> 00:09:48,920
long time now.

159
00:09:48,940 --> 00:09:52,840
And thankfully, last week, we finally killed it for good, it seems.

160
00:09:53,640 --> 00:10:02,460
And I thought, okay, I want to just talk to everyone, understand why they're so afraid of this particular thing, and what can we as tech companies do about it?

161
00:10:02,660 --> 00:10:11,700
It was very important for me to be able to say, hey, I want to do the right thing, and I'm doing the right thing, so tell me what you're worried about, and I'm going to find an engineering solution to this problem.

162
00:10:11,940 --> 00:10:18,140
Because I believe that it's not the fact that it's end-to-end encryption that scares a lot of these people, it is something else.

163
00:10:19,200 --> 00:10:25,580
And the more I talked, both to the regulators and to the civil rights org, one thing became

164
00:10:25,740 --> 00:10:30,880
clear is that what they're afraid of is not one person sharing CSAM with another singular

165
00:10:31,220 --> 00:10:31,340
person.

166
00:10:31,740 --> 00:10:34,500
What they're afraid of is mass spread of CSAM.

167
00:10:35,120 --> 00:10:40,240
So if you think about sharing CSAM with one person to another, one could put it on a USB

168
00:10:40,440 --> 00:10:42,280
flash drive and mail it to the other person.

169
00:10:42,500 --> 00:10:44,220
There is no way to detect it, stop it.

170
00:10:44,980 --> 00:10:50,040
It's effectively invisible to the law enforcement, and then someone could just do peer-to-peer sharing.

171
00:10:50,600 --> 00:10:55,260
That's not the scenario that they're most afraid of, because that's already been around, and that's very difficult to prevent.

172
00:10:55,820 --> 00:11:03,620
What they're worried about is people just uploading CSAM to a website, and then posting a link to the dark web with thousands of people downloading it.

173
00:11:04,060 --> 00:11:05,540
That's the scenario that they want to get ahead of.

174
00:11:06,320 --> 00:11:08,420
And they don't know how to address this.

175
00:11:08,820 --> 00:11:17,140
The one scenario that they can keep thinking of is like, well, there's all these chat apps, there's all these social media platforms like Facebook and Google and iCloud and whatnot.

176
00:11:17,230 --> 00:11:22,040
And realistically, these are the most popular vectors for CSAM to spread.

177
00:11:22,220 --> 00:11:24,580
People just upload bad shit on these platforms.

178
00:11:24,930 --> 00:11:26,960
And it's very easy to spread it to thousands of people.

179
00:11:27,030 --> 00:11:27,780
And that's how it happens.

180
00:11:28,380 --> 00:11:44,400
And so they basically said, hey, if you can build this in a responsible way where it is almost more impractical for someone to try to use Cryptee than it is for them to use any other fucking means of distributing CSAM, chances are they're not going to use Cryptee.

181
00:11:44,620 --> 00:11:52,680
But you need to make sure that barrier is so fucking high that no one in their same fucking mind who wants to share CSAM would use your platform.

182
00:11:53,040 --> 00:11:58,100
But anyone who wants to just share their family pictures with four or five people would do so.

183
00:11:58,520 --> 00:12:06,040
And that was sort of like a awakening, oh my God, I finally understand what these people are afraid of moment.

184
00:12:06,230 --> 00:12:10,520
Because up until this point, there was always this rage around the topic of encryption.

185
00:12:10,680 --> 00:12:12,300
Everyone's like, no, zero compromises.

186
00:12:12,500 --> 00:12:14,040
It's like, yeah, we're not going to make compromises.

187
00:12:14,220 --> 00:12:15,360
We just want to do this right.

188
00:12:17,839 --> 00:12:20,960
Yeah, so I feel, not to cut you off,

189
00:12:20,960 --> 00:12:24,960
but I just want to make sure we dive into these things

190
00:12:24,980 --> 00:12:25,700
a little bit more here.

191
00:12:25,960 --> 00:12:29,020
So I think in our initial very quick call

192
00:12:29,060 --> 00:12:30,480
when you were explaining things to me,

193
00:12:30,500 --> 00:12:33,080
you mentioned how end-to-end encrypted services

194
00:12:33,500 --> 00:12:38,500
and someone having bad content on their own platform,

195
00:12:38,760 --> 00:12:39,860
let's say, that's not being shared around,

196
00:12:40,120 --> 00:12:43,180
is no different than mailing a flash drive to a friend.

197
00:12:43,680 --> 00:12:45,760
No one's going to be opening up the envelope.

198
00:12:46,110 --> 00:12:48,300
No one's going to be analyzing what's on there,

199
00:12:48,460 --> 00:12:49,640
especially if it's an encrypted flash drive.

200
00:12:49,910 --> 00:12:51,020
But that's not the real concern.

201
00:12:51,140 --> 00:12:52,840
The concern is, in that analogy,

202
00:12:52,910 --> 00:12:55,660
the equivalent would be hosting a theater event

203
00:12:56,440 --> 00:13:00,500
where hundreds of people are coming to go watch a movie.

204
00:13:00,780 --> 00:13:02,340
And that's what is the worst problem

205
00:13:02,390 --> 00:13:03,720
and most preventable as well.

206
00:13:04,210 --> 00:13:05,600
And that is what you're saying is the issue,

207
00:13:05,780 --> 00:13:08,660
is just the accessibility of the content

208
00:13:09,080 --> 00:13:11,240
and whether or not you're actually building a platform

209
00:13:11,590 --> 00:13:13,460
rather than, obviously both are problems.

210
00:13:13,560 --> 00:13:15,500
So it would be awesome to deal with both of those issues.

211
00:13:15,660 --> 00:13:18,520
But it sounds like the biggest issue is that latter one, which is.

212
00:13:19,100 --> 00:13:19,460
That is correct.

213
00:13:20,259 --> 00:13:21,360
Widespread, mass spread.

214
00:13:21,820 --> 00:13:33,220
And essentially, to go with your analogy, like if we're talking about a theater, right, you invite 100 people, then those 100 people leave their own video copies and they get to have a 100% theater event, so to speak.

215
00:13:33,280 --> 00:13:34,800
And that's what they're trying to get ahead of.

216
00:13:34,960 --> 00:13:42,360
Because there is an R vector to this and you want to keep that as low as possible to prevent the spread of bad stuff on the internet.

217
00:13:43,020 --> 00:13:45,280
That's the ideal scenario, at least.

218
00:13:45,740 --> 00:13:50,340
So what you're saying is, yes, encryption makes it harder to moderate,

219
00:13:50,840 --> 00:13:54,880
which is probably not, you know, that's not a debatable thing.

220
00:13:55,080 --> 00:13:59,760
It is debatable if the moderation is, and breaking end-dead encryption is worth it,

221
00:13:59,960 --> 00:14:01,400
which I think we can talk about soon here.

222
00:14:01,440 --> 00:14:04,080
But the main issue you're saying is the shareability of it,

223
00:14:04,720 --> 00:14:06,720
which leads me to my next question for you,

224
00:14:06,880 --> 00:14:08,760
which is, you mentioned the Silicon Valley companies,

225
00:14:09,080 --> 00:14:11,220
you know, Apple, Google, they all have photo providers.

226
00:14:11,340 --> 00:14:12,140
They all allow sharing.

227
00:14:13,060 --> 00:14:18,040
In fact, Apple does have ADP, which is end-to-end encrypted, and it still enables sharing.

228
00:14:18,730 --> 00:14:21,480
You can do a public shared album on Apple.

229
00:14:22,010 --> 00:14:25,240
So how are regular big tech companies handling this?

230
00:14:25,970 --> 00:14:29,280
Is it still an issue on their platforms because they still have the shareability concerns,

231
00:14:30,340 --> 00:14:32,080
even though they don't have the encryption for most users?

232
00:14:32,880 --> 00:14:35,980
To the best of my understanding, now I can't speak for the tech companies

233
00:14:36,450 --> 00:14:39,580
because I don't know how their process is behind the scenes.

234
00:14:40,260 --> 00:14:45,580
But to the best of my understanding, what companies like Apple can afford to do is throw lawyers at problems.

235
00:14:46,660 --> 00:14:54,140
And essentially, Apple trusts on the fact that something like 0.5% of their users probably enable advanced data protection.

236
00:14:54,820 --> 00:15:02,100
And of those, we can, let's say, speculate, I don't know, 10%, 5% of it maybe actually uses it to spread CSAM, whatever.

237
00:15:02,300 --> 00:15:04,000
There's probably a percentage that we don't know.

238
00:15:04,010 --> 00:15:05,140
I'm making this up right now.

239
00:15:05,880 --> 00:15:12,680
And they probably arrived at the conclusion that this may be a tolerable number given their multi-billion user base.

240
00:15:13,620 --> 00:15:18,920
But companies like Google Photos, for example, they are not encrypted to the best of my knowledge.

241
00:15:19,420 --> 00:15:24,680
Therefore, they could just scan the images as they please and essentially shut down accounts beforehand.

242
00:15:24,800 --> 00:15:32,040
So they can probably use things like AI-based fingerprinting and they can use solutions like photo DNA developed by Microsoft and the NSA.

243
00:15:32,420 --> 00:15:35,500
And they can use a bunch of solutions to sort of fingerprint these existing images.

244
00:15:36,100 --> 00:15:41,920
but this sort of wouldn't work in the context of end-to-end encrypted sharing needless to say so i

245
00:15:42,060 --> 00:15:46,880
can't comment on what the big tech companies do but they have enough legal firepower to throw at

246
00:15:46,880 --> 00:15:54,200
these problems that smaller providers like us cannot do so to speak and a good example is like

247
00:15:54,380 --> 00:16:00,159
apple's legal budget for the dma hearings is one billion dollars so that's just for the dma alone

248
00:16:00,160 --> 00:16:04,020
And they can throw an obscene amount of money at problems.

249
00:16:04,440 --> 00:16:07,740
And I think that's the solution that someone else can't do.

250
00:16:08,060 --> 00:16:14,480
So that's how I see it, is that if we did this, there would be no way for us to get out of this.

251
00:16:14,640 --> 00:16:19,540
If Apple does it, they will probably be able to find a very expensive solution to get out of it.

252
00:16:20,279 --> 00:16:20,720
Yeah.

253
00:16:22,260 --> 00:16:25,800
Can you explain maybe a little bit before we get into my next questions,

254
00:16:26,010 --> 00:16:28,620
why the end-to-end encryption prevents your ability to do this?

255
00:16:28,620 --> 00:16:29,900
I know you guys can't.

256
00:16:30,100 --> 00:16:34,540
The whole point of end-to-end encryption is that only the account or whoever they share that account with,

257
00:16:34,710 --> 00:16:40,000
whoever has the deCrypteeon keys can view the data, which you guys in theory shouldn't have because you're end-to-end encrypted.

258
00:16:40,250 --> 00:16:44,880
But what about, I know Apple had the whole thing about local scanning on devices.

259
00:16:45,980 --> 00:16:51,660
Theoretically, could an end-to-end encrypted service still do local scanning without compromising the end-to-end encryption?

260
00:16:52,780 --> 00:16:57,380
Now, they could theoretically do that, but they would be compromising the scanning algorithm.

261
00:16:57,780 --> 00:17:17,439
Now, the thing that a lot of people are afraid of, and we looked into this, just to be very clear, to understand what is it that everyone's trying to market around and what's this thing that everyone's saying, hey, we have this incredible pre-encryption scanning, whatever, magic solution that, of course, doesn't work, needless to say.

262
00:17:17,520 --> 00:17:22,500
What these scanning solutions do is they've scrapped together a bunch of AI algorithms.

263
00:17:23,300 --> 00:17:29,500
And essentially, they've tried combining things like age detection and nudity in the same algorithm.

264
00:17:29,780 --> 00:17:33,860
So they basically try to score an image saying, hey, is there an underage person?

265
00:17:34,320 --> 00:17:35,500
Is there some nudity in this?

266
00:17:36,080 --> 00:17:40,000
And maybe some context like, is this a photo of a kid on a beach type of thing?

267
00:17:40,020 --> 00:17:42,880
And they try to eliminate some false positives in a terrible way.

268
00:17:43,420 --> 00:17:47,040
And needless to say, that's not what all CSAM is about.

269
00:17:47,260 --> 00:17:56,980
And every single activism and every single human rights group and every single women's rights group and children's rights group that we talked to said, that's not what all CSAM is about.

270
00:17:57,280 --> 00:17:59,340
There's a lot worse stuff than that.

271
00:17:59,920 --> 00:18:02,400
And so you can't just casually detect with that.

272
00:18:02,520 --> 00:18:03,060
There's gore.

273
00:18:03,300 --> 00:18:04,400
There's all sorts of other things.

274
00:18:04,560 --> 00:18:10,160
There's, not to get into super gory and boring details here, essentially, it just doesn't work.

275
00:18:10,280 --> 00:18:11,360
You can't just do that.

276
00:18:11,560 --> 00:18:18,880
And the other alternative to this is to use something like PhotoDNA developed by Microsoft and NSA, which essentially fingerprints against an existing database.

277
00:18:19,400 --> 00:18:25,560
The problem with that is that their hashing algorithm for these photos need to remain private.

278
00:18:26,580 --> 00:18:36,160
If you ship it in your front end, then suddenly people realize, oh, if I upload like 10 photos and do this edit and do that edit, it doesn't detect.

279
00:18:36,200 --> 00:18:40,140
And you can essentially brute force your way through the hashing algorithm and it loses all effectiveness.

280
00:18:40,800 --> 00:18:44,680
So in order for photo DNA to be effective, it needs to remain private.

281
00:18:44,890 --> 00:18:46,460
You can't just ship it with your client.

282
00:18:46,890 --> 00:18:51,000
So needless to say, we're never going to be able to ship any solution like that

283
00:18:51,140 --> 00:18:56,060
or any end-to-end encrypted provider who prefers to have their code open source like we do

284
00:18:56,270 --> 00:19:00,620
will never be able to do that because it effectively kills the whole point of the hashing algorithm

285
00:19:01,600 --> 00:19:04,400
as someone's going to reverse engineer that and defeats the whole point.

286
00:19:04,640 --> 00:19:08,880
And you can maybe ship it as a compiled version with the app somehow.

287
00:19:09,500 --> 00:19:16,020
And it's still not going to work because then someone's going to capture the network traffic tomorrow and figure out what it's actually uploading.

288
00:19:16,480 --> 00:19:17,760
And that's game over also.

289
00:19:17,900 --> 00:19:23,260
So none of these pre-encryption scanning things that they're trying to sell, it doesn't make any sense at all.

290
00:19:23,400 --> 00:19:24,760
It's just really broken in general.

291
00:19:25,200 --> 00:19:27,120
How do chat apps right now kind of handle this?

292
00:19:27,120 --> 00:19:32,000
I mean, so here's, you know, we always dive into this in other contexts when people ask, oh, do you like Telegram?

293
00:19:32,060 --> 00:19:37,760
I say, no, I don't like Telegram because Telegram by default doesn't have any end encryption in groups or channels.

294
00:19:38,240 --> 00:19:41,060
you have to opt into it for even regular DMs.

295
00:19:41,520 --> 00:19:43,340
And it puts them in this really interesting position

296
00:19:43,540 --> 00:19:46,420
where now you actually have to trust Telegram a lot more

297
00:19:46,460 --> 00:19:47,740
than you have to trust something like Signal

298
00:19:48,020 --> 00:19:50,160
because now Telegram is holding your data

299
00:19:50,380 --> 00:19:53,480
and you have to trust them to respond to legal requests

300
00:19:53,940 --> 00:19:56,460
and you have to trust that they're sticking to their privacy policy.

301
00:19:57,520 --> 00:19:59,420
Signal, you don't really have to trust them as much

302
00:19:59,640 --> 00:20:00,940
because there's only so much they can do

303
00:20:01,000 --> 00:20:02,620
even if they wanted to get access to your messages.

304
00:20:03,920 --> 00:20:04,440
Signal's amazing.

305
00:20:04,880 --> 00:20:07,380
To the best of my understanding, they don't.

306
00:20:07,840 --> 00:20:13,700
First of all, but they do things that are very similar to what we've implemented in our platform.

307
00:20:14,010 --> 00:20:17,260
Well, Signal handles it by means of not storing anything.

308
00:20:18,120 --> 00:20:19,100
That's a solution, right?

309
00:20:19,740 --> 00:20:20,400
It's ephemeral.

310
00:20:20,670 --> 00:20:25,780
You send something, Signal just forwards it to the next person, and then it's gone.

311
00:20:25,830 --> 00:20:27,340
It doesn't stay on Signal's servers.

312
00:20:27,680 --> 00:20:33,000
So that, because your sharing is more permanent, is actually where the problem arises, I see.

313
00:20:33,400 --> 00:20:35,620
Correct, because we are the storage facility.

314
00:20:36,160 --> 00:20:38,720
By law, when you're transiting something,

315
00:20:38,750 --> 00:20:41,080
it's kind of like you can't keep the mail service

316
00:20:41,460 --> 00:20:43,580
responsible for sending the flash drive.

317
00:20:44,050 --> 00:20:45,200
They didn't know what's inside.

318
00:20:45,460 --> 00:20:47,700
Because you're a photo storage provider.

319
00:20:48,380 --> 00:20:50,060
Exactly, then we become liable.

320
00:20:50,490 --> 00:20:52,460
So WhatsApp is a bit more liable

321
00:20:52,780 --> 00:20:55,580
because they actually do store encrypted content

322
00:20:55,690 --> 00:20:56,280
on their servers.

323
00:20:57,100 --> 00:20:59,360
But to combat some of this,

324
00:20:59,800 --> 00:21:01,900
they have recently implemented solutions

325
00:21:02,220 --> 00:21:04,140
like limit the number of users

326
00:21:04,170 --> 00:21:05,460
who can be in a single group.

327
00:21:05,680 --> 00:21:07,980
omit the number of times an image can be forwarded, etc.

328
00:21:08,880 --> 00:21:11,620
And I think that is effectively the solution,

329
00:21:11,840 --> 00:21:13,880
and that's what we've arrived at also.

330
00:21:13,980 --> 00:21:17,240
And I suppose this is a very long way of getting to the solutions

331
00:21:17,440 --> 00:21:20,460
that we come up with to address this problem.

332
00:21:21,840 --> 00:21:24,360
Before we get there, one final question here.

333
00:21:24,580 --> 00:21:26,640
Because Signal recently released NDEA encrypted backups,

334
00:21:26,880 --> 00:21:29,320
and WhatsApp also has backups to their own servers.

335
00:21:29,700 --> 00:21:30,460
Is that not storage?

336
00:21:31,680 --> 00:21:35,840
Well, it's different because your backups are not meant to be shared with someone else.

337
00:21:36,130 --> 00:21:37,420
Your backups are for you.

338
00:21:39,120 --> 00:21:41,980
It's the sharing aspect, basically, that becomes a problem.

339
00:21:42,050 --> 00:21:46,660
So you're more liable to this problem if you're a communications provider of any sort.

340
00:21:47,360 --> 00:21:49,940
It could be messaging, it could be photo sharing, whatnot.

341
00:21:50,050 --> 00:21:54,840
But if you're just doing personal backups, it's no different than putting your hard drive in a storage facility in the city somewhere.

342
00:21:57,160 --> 00:22:01,200
I want to touch on the solutions, and then after I want to zoom out and just ask more broadly,

343
00:22:02,240 --> 00:22:04,100
The philosophical angle of this, right?

344
00:22:04,250 --> 00:22:06,000
Because I feel like it's a very touchy subject.

345
00:22:06,650 --> 00:22:09,240
I don't like talking about it, but I feel like it's something that's not discussed.

346
00:22:09,240 --> 00:22:10,460
I don't think anyone does.

347
00:22:10,610 --> 00:22:13,660
And I think that's the difficult part, is that nobody wants to talk about it.

348
00:22:14,060 --> 00:22:15,100
Everyone's afraid of it.

349
00:22:15,320 --> 00:22:18,800
Everyone thinks that encryption's sacred and it shouldn't be compromised.

350
00:22:19,080 --> 00:22:20,100
And I 100% agree.

351
00:22:20,400 --> 00:22:25,800
And that's the thing where it's like we have this inherent issue around nobody wants to talk about.

352
00:22:27,140 --> 00:22:29,800
So I want to get to that and the solutions.

353
00:22:29,980 --> 00:22:32,740
But last question before I dive into the solutions,

354
00:22:32,940 --> 00:22:34,480
and also we talk about chat control later too.

355
00:22:34,740 --> 00:22:37,760
I know you probably do have discussions with other providers

356
00:22:37,840 --> 00:22:39,380
that care about privacy and security out there.

357
00:22:40,580 --> 00:22:41,620
How much is this considered?

358
00:22:42,120 --> 00:22:44,220
I feel like when people...

359
00:22:45,640 --> 00:22:46,040
Heavily.

360
00:22:47,660 --> 00:22:49,360
Yeah, when people think of Signal,

361
00:22:49,480 --> 00:22:51,840
when they think of Tor,

362
00:22:52,160 --> 00:22:53,480
when they think of all these services,

363
00:22:53,560 --> 00:22:55,540
they just think, wow, they just built a private and secure service

364
00:22:55,640 --> 00:22:58,160
and they're just rolling out new features as they're ready to build them.

365
00:22:59,020 --> 00:23:01,280
But how much, you know, this isn't talked about.

366
00:23:01,690 --> 00:23:02,720
No one talks about this.

367
00:23:03,060 --> 00:23:08,140
So is this secretly something that every team behind the scenes is taking into consideration

368
00:23:08,560 --> 00:23:11,400
when they're rolling out their services that they just don't want to talk about publicly?

369
00:23:12,380 --> 00:23:13,740
I would certainly hope so.

370
00:23:15,280 --> 00:23:16,880
I mean, I can at least say this.

371
00:23:17,030 --> 00:23:21,540
I haven't spoken to any other founders about this particular topic in depth.

372
00:23:21,930 --> 00:23:26,760
But I have spoken to some founders who had companies that were dedicated,

373
00:23:27,020 --> 00:23:32,560
Like one of my really good friends who used to be my neighbor, she started an OnlyFans competitor.

374
00:23:32,860 --> 00:23:37,560
And we had a very extensive two and a half hour podcast session about this and how she gets ahead of all this.

375
00:23:37,760 --> 00:23:39,900
And she's an incredibly brilliant person.

376
00:23:40,380 --> 00:23:43,900
And I'll send you the link if you wish to share the details of that podcast.

377
00:23:44,620 --> 00:23:47,680
There's a very long, extensive conversation we had about this.

378
00:23:48,540 --> 00:23:56,380
And long story short is that I think for companies that matter, they do hopefully think about it.

379
00:23:56,540 --> 00:23:58,360
Within the framework of chat control,

380
00:23:59,160 --> 00:24:02,380
I do know that pretty much every messenger service out there

381
00:24:02,510 --> 00:24:05,240
thinks about it extensively, and it's a big concern for them.

382
00:24:05,980 --> 00:24:08,880
To give you one very short personal experience,

383
00:24:09,210 --> 00:24:12,380
as an example, where a couple months ago I was in Amsterdam,

384
00:24:13,140 --> 00:24:14,940
and as I'm walking on the street,

385
00:24:15,070 --> 00:24:18,100
and this was right on the week of the chat control vote,

386
00:24:18,100 --> 00:24:19,800
I think I sent you a message when it happened to you,

387
00:24:20,120 --> 00:24:22,420
but this was the week of the chat control vote,

388
00:24:22,430 --> 00:24:25,100
so I was on my way to Brussels, and I was in Amsterdam.

389
00:24:25,680 --> 00:24:31,840
And literally on the street, outside of a cafe, I bumped into Mary Beth Withaker, president of Signal.

390
00:24:32,160 --> 00:24:34,040
And I said, like, hey, how's it going?

391
00:24:34,110 --> 00:24:37,900
And we very briefly for like two minutes had a conversation.

392
00:24:38,090 --> 00:24:40,520
And the entirety of those two minutes were about chat control.

393
00:24:40,980 --> 00:24:41,700
That says something.

394
00:24:42,400 --> 00:24:44,700
And Mary Beth was like, we got to do something about this.

395
00:24:44,710 --> 00:24:46,520
And I said, yeah, we got to do something about this.

396
00:24:46,520 --> 00:24:55,140
And that was the week where I was flying around like headless chicken crazy, running around trying to talk to every regulator that I can talk to.

397
00:24:55,300 --> 00:24:57,620
And I told Mirabeth, and I think I told you,

398
00:24:57,800 --> 00:24:59,480
I've spoken to everyone in Estonia and Finland

399
00:24:59,560 --> 00:25:01,900
that I can speak to to try to get the countries

400
00:25:02,320 --> 00:25:03,440
to line up as much as we could.

401
00:25:03,700 --> 00:25:05,880
And thankfully Meredith was talking about that too

402
00:25:05,900 --> 00:25:07,940
and she was very much concerned about that too.

403
00:25:08,120 --> 00:25:11,500
And yeah, everyone was thinking about that at that time

404
00:25:11,620 --> 00:25:14,640
and still is thinking about that, I think, after this week.

405
00:25:15,519 --> 00:25:18,360
I want to talk, let's put a pin in that.

406
00:25:18,440 --> 00:25:20,880
I want to ask you more about the regulation aspect of things

407
00:25:21,060 --> 00:25:24,720
because I know a lot of people are wondering why that needs to happen.

408
00:25:24,960 --> 00:25:26,080
You just do your job.

409
00:25:26,300 --> 00:25:27,800
Just roll out good tech,

410
00:25:28,160 --> 00:25:28,980
roll out good encryption,

411
00:25:29,680 --> 00:25:30,900
and bury your head in that.

412
00:25:31,620 --> 00:25:33,600
So why take all that time to go to process?

413
00:25:33,820 --> 00:25:35,060
But I want to talk about that

414
00:25:35,340 --> 00:25:37,180
because I think it's probably going to lead

415
00:25:37,280 --> 00:25:38,040
to some interesting discussions.

416
00:25:38,320 --> 00:25:39,480
Now, I want to get into your solution

417
00:25:39,820 --> 00:25:40,940
and how you guys approach this

418
00:25:41,100 --> 00:25:42,160
because you're not compromising

419
00:25:42,220 --> 00:25:43,420
on the encryption aspect of things.

420
00:25:43,500 --> 00:25:44,440
So I'm just going to assume

421
00:25:44,460 --> 00:25:45,220
that it's a sharing thing.

422
00:25:45,460 --> 00:25:47,260
Actually, you sent me a couple of test albums

423
00:25:47,680 --> 00:25:48,660
and they expired before

424
00:25:48,940 --> 00:25:50,280
because I've been ignoring you on Signal.

425
00:25:51,520 --> 00:25:52,880
Because I've just been so busy this week.

426
00:25:53,240 --> 00:25:54,040
I'll send them to me again.

427
00:25:54,100 --> 00:25:54,640
You'll send them to me again.

428
00:25:55,400 --> 00:25:57,940
You have some photos from when I was there visiting you.

429
00:25:58,160 --> 00:26:00,180
So you're sharing these photos with me.

430
00:26:00,320 --> 00:26:02,960
What does this look like now with the way that you're implementing this in Cryptee?

431
00:26:04,159 --> 00:26:09,220
We decided that there's got to be a healthy framework that we set

432
00:26:09,500 --> 00:26:13,540
to figure out what it is that we want to address first.

433
00:26:13,760 --> 00:26:15,720
Let's figure out what is everyone afraid of

434
00:26:15,820 --> 00:26:17,480
and what can we do to actually address these things.

435
00:26:17,640 --> 00:26:19,300
And the first thing that we decided is,

436
00:26:19,460 --> 00:26:22,580
we don't want this to be spread en masse.

437
00:26:23,040 --> 00:26:26,060
This needs to be a solution where you can easily share it with your family,

438
00:26:26,180 --> 00:26:27,160
you can share it with your client.

439
00:26:27,250 --> 00:26:30,240
If you're a war photographer, you can send the photos to your editors,

440
00:26:30,430 --> 00:26:31,500
to your news organization.

441
00:26:32,120 --> 00:26:35,960
The idea here is to just make it so that you can very easily share

442
00:26:36,010 --> 00:26:37,680
with the number of people you would want to share,

443
00:26:38,480 --> 00:26:42,900
but not spread it en masse to very large numbers of people.

444
00:26:43,940 --> 00:26:45,860
Now, if you want to do link sharing,

445
00:26:46,210 --> 00:26:48,480
that just automatically complicates things,

446
00:26:48,650 --> 00:26:51,660
because how do you verify the number of people visiting that link?

447
00:26:52,100 --> 00:26:54,020
So you set it just to IP addresses?

448
00:26:54,640 --> 00:26:57,100
Well, then how do you deal with people using VPN services?

449
00:26:57,350 --> 00:26:58,760
What if they all use the same VPN?

450
00:26:59,420 --> 00:27:03,380
Well, how do you do, like if it's like a one-time visit in that browser,

451
00:27:03,580 --> 00:27:05,400
what if your client couldn't open it on their desktop,

452
00:27:05,620 --> 00:27:09,260
but wants to open it on their phone or their whatever other browser and whatnot?

453
00:27:09,460 --> 00:27:10,080
That's a problem.

454
00:27:10,340 --> 00:27:10,900
That's not going to work.

455
00:27:11,460 --> 00:27:13,780
So after going through a bunch of iterations, we thought,

456
00:27:13,850 --> 00:27:15,860
okay, there's got to be a better way to do this

457
00:27:15,970 --> 00:27:18,560
because we also don't want people to just have to worry about

458
00:27:19,020 --> 00:27:21,320
hoarding links somewhere that they received.

459
00:27:21,720 --> 00:27:25,500
And there's got to be a better way to just sort of make this more permanent and easy for people to access.

460
00:27:26,440 --> 00:27:31,520
So we thought, OK, why don't we just make it so that the recipients of the links have to enter their email address.

461
00:27:31,880 --> 00:27:39,800
And once they verify their email address with a one-time password, then we know for sure that only that email can access that album.

462
00:27:40,680 --> 00:27:45,400
And that's it. That email is added to the recipient invitees or the recipients.

463
00:27:45,880 --> 00:27:50,700
And that way we know there can be maximum five email addresses that can receive an album.

464
00:27:51,120 --> 00:27:52,800
So if you have a family, that's enough.

465
00:27:52,970 --> 00:27:54,380
If you have an editor, that's enough.

466
00:27:54,560 --> 00:27:58,760
If you shot wedding photography and you just want to send it to the bride and groom's family,

467
00:27:59,080 --> 00:27:59,640
five is enough.

468
00:27:59,730 --> 00:28:03,080
Like we realized that's a very sufficient number generally.

469
00:28:03,170 --> 00:28:06,440
And if you went to a party with your friends, like five is probably enough.

470
00:28:06,700 --> 00:28:11,000
And please contact us if you feel that five is too small and we'll make it 10.

471
00:28:11,540 --> 00:28:16,000
But basically the point is that five generally felt like a very healthy number.

472
00:28:16,340 --> 00:28:19,580
And we thought email verification should be fine, right?

473
00:28:19,940 --> 00:28:24,280
If you're a wedding photographer and you're sharing photos that you shot at someone's wedding,

474
00:28:25,020 --> 00:28:28,820
you wouldn't mind your client adding their email there.

475
00:28:29,100 --> 00:28:32,360
Or if you're sharing photos as a war photographer and you send it to New York Times,

476
00:28:32,640 --> 00:28:35,700
you wouldn't mind if someone put it at newyorktimes.com email.

477
00:28:36,280 --> 00:28:37,260
That's not going to change anything.

478
00:28:37,500 --> 00:28:39,220
And if you share it with your friends, it's totally fine.

479
00:28:39,260 --> 00:28:40,320
You already have their email address.

480
00:28:41,040 --> 00:28:45,278
But if you were trying to share CSAM, as a recipient, you would probably question

481
00:28:45,300 --> 00:28:49,960
what the fuck if you have to enter your email address and verify and send it to the person

482
00:28:50,280 --> 00:28:55,640
sharing it how do you know if someone's not trying to bait you into something and catch you something

483
00:28:55,980 --> 00:29:00,660
and this is we thought this could be a good way to actually prevent a little bit of the bad stuff

484
00:29:01,440 --> 00:29:05,980
because you will have no way of knowing who the person sharing this is and maybe you're being

485
00:29:06,180 --> 00:29:11,378
baited you'll have no idea so it inherently raises the bar this is the type of bar raising that i was

486
00:29:11,400 --> 00:29:15,640
talking about earlier, that if we make it difficult enough for people who wish to share bad stuff on

487
00:29:15,640 --> 00:29:21,140
our platform, they will simply just not do so. And they will use significantly simpler alternatives

488
00:29:21,270 --> 00:29:26,520
to do so. And when they do, we reach our goal. So that's how we decided, okay, five recipients,

489
00:29:27,040 --> 00:29:30,640
everyone has to verify their email address. And when you do enter your email address,

490
00:29:30,790 --> 00:29:34,380
we assign a token that's valid for six hours, we assign it to your IP address,

491
00:29:35,040 --> 00:29:39,298
and we assign it to that email address. And all these things happen behind the scene. And we

492
00:29:39,360 --> 00:29:40,960
We try to simplify it as much as we can.

493
00:29:41,370 --> 00:29:47,400
And to add to that, the sharing parties, the original source of the content, has to be a paid user.

494
00:29:48,340 --> 00:29:56,440
This allows us to make sure that free users cannot just sign up and create tons of accounts and try to mass spread the content that way.

495
00:29:57,100 --> 00:30:01,540
But instead allows us to limit the person who would be sharing this to one.

496
00:30:01,940 --> 00:30:05,760
Or let's say two, if you happen to have two cards and you created two accounts that way.

497
00:30:06,100 --> 00:30:20,540
But the idea here is that we can actually use payment methods as a means of also KYC and our customers to say, hey, we cannot see your photos, but we need to know that you are a real human and not a bot creating 1000 accounts.

498
00:30:21,040 --> 00:30:23,820
If you have a payment method on file, that sort of establishes that.

499
00:30:24,340 --> 00:30:27,700
We don't need to see your photos to be able to prevent bad stuff from happening.

500
00:30:28,070 --> 00:30:34,940
If we can just prevent bots from signing up and sending thousands, or if we can stop you from creating thousand accounts and spreading that way.

501
00:30:35,380 --> 00:30:37,180
So that's sort of the solution that we arrived to.

502
00:30:37,220 --> 00:30:39,020
And also, if you're trying to spread CSAM,

503
00:30:39,020 --> 00:30:40,020
the last thing you would want to do

504
00:30:40,020 --> 00:30:41,900
is probably actually give your payment information

505
00:30:42,060 --> 00:30:43,360
and billing information to a service.

506
00:30:43,960 --> 00:30:44,900
If something goes wrong,

507
00:30:45,300 --> 00:30:47,200
we can then report it to the law enforcement

508
00:30:47,460 --> 00:30:48,980
and that raises the bar yet again.

509
00:30:49,480 --> 00:30:51,660
So that's sort of the solution that we arrived.

510
00:30:51,720 --> 00:30:52,960
It's like only five people.

511
00:30:53,140 --> 00:30:53,920
There needs to be emails.

512
00:30:54,300 --> 00:30:55,860
The sharing party needs to have a payment method.

513
00:30:56,420 --> 00:30:57,740
And to make it even more difficult

514
00:30:57,960 --> 00:30:59,940
so that you can't just post a link in the dark web

515
00:31:00,400 --> 00:31:02,140
and let anyone join this thing,

516
00:31:02,420 --> 00:31:03,340
links need to expire,

517
00:31:03,680 --> 00:31:07,660
or at least the invites to the links need to expire after 48 hours.

518
00:31:08,720 --> 00:31:12,200
So if you didn't accept the invite, I can create another link and just give it to you.

519
00:31:12,420 --> 00:31:12,680
That's okay.

520
00:31:13,340 --> 00:31:16,840
But that link expires again in 48 hours unless you accept it.

521
00:31:16,940 --> 00:31:19,300
If you do accept it, you can continue to access it forever.

522
00:31:19,780 --> 00:31:21,680
So if you think about it, you just send the link to your mom.

523
00:31:22,160 --> 00:31:23,160
Your mom accepts it.

524
00:31:23,390 --> 00:31:24,340
She can access the photos.

525
00:31:24,860 --> 00:31:25,220
Problem solved.

526
00:31:25,500 --> 00:31:25,640
Forever.

527
00:31:26,140 --> 00:31:27,340
She's going to have access to that.

528
00:31:27,570 --> 00:31:27,960
No problem.

529
00:31:28,420 --> 00:31:31,200
But if you decide to post that link in a dark web form,

530
00:31:31,600 --> 00:31:33,440
It's not going to live longer than 48 hours,

531
00:31:33,540 --> 00:31:35,600
and that's sort of the point that we're trying to address here.

532
00:31:37,000 --> 00:31:39,240
But the link is almost like a signal username.

533
00:31:40,140 --> 00:31:42,920
It's just meant to add someone, and you can change the username afterwards,

534
00:31:43,300 --> 00:31:44,860
and you're already linked at that point.

535
00:31:45,400 --> 00:31:46,560
Correct, that's the idea.

536
00:31:46,780 --> 00:31:48,940
And you can revoke the link at any point you want.

537
00:31:49,060 --> 00:31:52,480
So say, for example, you wish to share private photos with your partner.

538
00:31:52,800 --> 00:31:54,320
That's totally fine by us, right?

539
00:31:54,720 --> 00:31:56,580
But let's say you broke up with your partner,

540
00:31:56,740 --> 00:31:59,080
and you don't want them to have those private photos anymore.

541
00:31:59,540 --> 00:32:03,500
then you should be able to very easily revoke that link and get your access back.

542
00:32:03,780 --> 00:32:06,240
And we thought that the most important thing for us right now,

543
00:32:06,400 --> 00:32:10,460
after having spoken to both women's rights organizations and children's rights organizations,

544
00:32:11,280 --> 00:32:12,340
consent is the keyword.

545
00:32:13,260 --> 00:32:19,520
As long as you're sharing your photos and videos and your consent is still there, that's good.

546
00:32:20,180 --> 00:32:22,460
If you wish to revoke that link, consent is the keyword.

547
00:32:22,640 --> 00:32:25,000
You should be able to revoke that link and it should be gone instantly.

548
00:32:25,420 --> 00:32:26,960
So that's basically how we built this.

549
00:32:27,100 --> 00:32:29,080
So that if you revoke the link, it's just gone instantly

550
00:32:29,170 --> 00:32:30,760
and you can share the link again if you want to,

551
00:32:30,970 --> 00:32:31,500
if this makes sense.

552
00:32:33,080 --> 00:32:34,540
So this isn't for quick sharing.

553
00:32:34,960 --> 00:32:36,600
I feel like people think of photo sharing

554
00:32:36,670 --> 00:32:39,780
and they think, oh, I met this person at a conference

555
00:32:40,400 --> 00:32:41,680
and now I'm just going to get them a link

556
00:32:42,180 --> 00:32:44,340
and I'm going to share with them these 10 photos I took.

557
00:32:44,700 --> 00:32:46,760
Theoretically, that could still work, but I feel like...

558
00:32:46,760 --> 00:32:49,020
I mean, it's just as simple as you give them a link,

559
00:32:49,170 --> 00:32:50,840
they enter their email address and that's it.

560
00:32:50,970 --> 00:32:52,140
They have access to those links now.

561
00:32:54,180 --> 00:32:54,680
Yeah, that's true.

562
00:32:55,580 --> 00:32:59,120
Everything that happens behind the scenes that I've just described is actually...

563
00:32:59,120 --> 00:33:00,460
It is pretty seamless, it seems.

564
00:33:00,600 --> 00:33:01,440
It's pretty seamless.

565
00:33:01,690 --> 00:33:04,920
Basically, you get a link, you just type your email address, and you have access to it.

566
00:33:05,060 --> 00:33:05,300
That's it.

567
00:33:05,420 --> 00:33:09,180
And the process is actually simpler than the WeTransfer's process, for context.

568
00:33:09,880 --> 00:33:12,380
With WeTransfer, you have to enter your email each time you send,

569
00:33:12,520 --> 00:33:14,620
and the recipient has to enter their email each time they receive.

570
00:33:15,480 --> 00:33:18,540
In our case, you just have to create a link to whatever album you want,

571
00:33:18,540 --> 00:33:19,100
and that's it.

572
00:33:19,100 --> 00:33:21,580
The recipient just enters their email and accepts it.

573
00:33:23,240 --> 00:33:25,920
Let's say someone is new, they've never used Cryptee before,

574
00:33:26,100 --> 00:33:27,800
I'm at a conference, I have these 10 photos,

575
00:33:27,980 --> 00:33:30,120
I give it to the person, they have to put in an email.

576
00:33:30,280 --> 00:33:32,920
Is that creating a free Cryptee account, or is that just to access?

577
00:33:34,280 --> 00:33:36,040
But then where do the photos go?

578
00:33:36,840 --> 00:33:38,500
Because the link will expire in two days.

579
00:33:39,260 --> 00:33:43,740
The link is essentially the same, but the invite expires after 48 days.

580
00:33:44,280 --> 00:33:46,160
At that point, it's linked to the email,

581
00:33:47,680 --> 00:33:49,820
and how do they access the photos after two days?

582
00:33:50,700 --> 00:33:51,540
With the email address.

583
00:33:51,940 --> 00:33:55,680
Whenever they want to access, they just do the one-time password with that email.

584
00:33:55,900 --> 00:33:58,100
If you did it on your browser and then you want to do it on your phone,

585
00:33:58,580 --> 00:34:01,040
you just enter your email on your phone, you get a one-time password.

586
00:34:02,660 --> 00:34:05,480
Are they logging into Cryptee?

587
00:34:06,480 --> 00:34:09,840
No, they're just accessing the album with a one-time password using that email.

588
00:34:10,320 --> 00:34:11,899
The link has the encryption key.

589
00:34:13,240 --> 00:34:15,100
So they are saving a link somewhere.

590
00:34:15,260 --> 00:34:16,540
There is a link to the album.

591
00:34:17,260 --> 00:34:20,860
But the invite link is separate from the actual permanent link to the album.

592
00:34:21,399 --> 00:34:22,139
They're the same link.

593
00:34:23,100 --> 00:34:24,220
But you said the link expires.

594
00:34:25,159 --> 00:34:27,159
So the invite portion of the link expires.

595
00:34:27,419 --> 00:34:30,600
Essentially, if you want, I can do a simpler explanation.

596
00:34:30,940 --> 00:34:32,940
But basically, when you send a link,

597
00:34:33,460 --> 00:34:34,800
let's say you receive that link,

598
00:34:35,300 --> 00:34:38,280
if you don't accept invite to that link within the 48 hours,

599
00:34:38,740 --> 00:34:41,000
you cannot accept invite any longer.

600
00:34:41,379 --> 00:34:43,780
But after 48 hours, if you've accepted it,

601
00:34:43,780 --> 00:34:45,879
you can continue to access the images from that link.

602
00:34:46,040 --> 00:34:47,639
There is no separate link involved, basically.

603
00:34:48,100 --> 00:34:49,000
It's the same link.

604
00:34:49,679 --> 00:34:54,560
And if you already have a Cryptee account and you put in your Cryptee email, does that add it to your Cryptee account?

605
00:34:55,600 --> 00:34:56,760
Not yet, but coming soon.

606
00:34:56,960 --> 00:35:02,700
So basically what we're working on now is the ability to be able to simply click one button and say,

607
00:35:02,750 --> 00:35:05,120
hey, I'm already logged into Cryptee, just add this to my account.

608
00:35:05,150 --> 00:35:08,060
And you don't have to enter your email, we already know who you are if you're logged in.

609
00:35:08,600 --> 00:35:13,600
And it's going to be as simple as literally just open in my Cryptee account and it will be permanently there.

610
00:35:13,670 --> 00:35:14,720
You won't need the story link.

611
00:35:15,060 --> 00:35:18,260
Whose storage is that quota being used against?

612
00:35:19,080 --> 00:35:20,700
The person who shared the album.

613
00:35:21,460 --> 00:35:24,260
So essentially the photos are stored in, let's say,

614
00:35:24,360 --> 00:35:26,780
the photos I shared with you, they're stored in my account

615
00:35:26,960 --> 00:35:29,300
and I am giving you access to those files.

616
00:35:29,680 --> 00:35:30,940
That's how we're designing this.

617
00:35:31,440 --> 00:35:34,200
So if I revoke it or I delete those files, they're gone.

618
00:35:34,240 --> 00:35:36,740
You don't get to have a copy of it because that assumes

619
00:35:37,360 --> 00:35:43,240
that I am removing the content of someone else accessing it, basically.

620
00:35:43,900 --> 00:35:44,480
I see, that's cool.

621
00:35:45,020 --> 00:35:46,000
That's the way we designed it.

622
00:35:46,740 --> 00:35:48,620
The big thing that comes up, actually,

623
00:35:48,720 --> 00:35:51,040
and this is maybe a little mini rabbit hole

624
00:35:51,080 --> 00:35:52,120
that we should have to go down

625
00:35:52,840 --> 00:35:54,400
before we zoom out again.

626
00:35:54,660 --> 00:35:55,680
A lot of the things you brought up

627
00:35:55,820 --> 00:35:57,200
don't sound very privacy-respecting,

628
00:35:57,400 --> 00:35:57,940
just at first glance.

629
00:35:59,160 --> 00:36:00,980
You mentioned, first off, requiring an email.

630
00:36:01,500 --> 00:36:04,040
It's not just an anonymous link that anyone can access.

631
00:36:04,260 --> 00:36:05,800
I understand the context.

632
00:36:06,820 --> 00:36:08,260
We've been talking about CSAM and everything.

633
00:36:09,140 --> 00:36:10,060
I understand that.

634
00:36:10,200 --> 00:36:12,460
But payment method as well, it needs a paid account.

635
00:36:13,240 --> 00:36:15,600
I guess I would ask, is it a problem?

636
00:36:15,780 --> 00:36:18,560
Because you mentioned a New York Times reporter needing to put in an email.

637
00:36:18,890 --> 00:36:20,360
Do you see a concern with that?

638
00:36:20,510 --> 00:36:30,180
And would it make it any easier for Cryptee to potentially notice that certain accounts are tied to journalists and put you guys in greater liability?

639
00:36:30,720 --> 00:36:34,700
And that could also implicate any kind of journalist on your platform if there is that kind of data.

640
00:36:35,240 --> 00:36:38,260
So I'm kind of curious what the privacy implications are of everything you shared there.

641
00:36:39,240 --> 00:36:42,180
So good news is that we wouldn't know who the sharing person is.

642
00:36:42,320 --> 00:36:45,400
So if you have a Cryptee account, and let's say you're a war photographer, right?

643
00:36:46,200 --> 00:36:51,480
we still don't know who you work for we have no clue no idea and otherwise we would have your

644
00:36:51,620 --> 00:36:55,180
payment method just to be able to store your photos anyway that doesn't change anything and

645
00:36:55,280 --> 00:37:01,040
as for the recipients let's say the recipient is new york times that doesn't mean i the sharing

646
00:37:01,280 --> 00:37:05,500
party actually work for new york times it could be that i am just sharing photos that i took during

647
00:37:05,530 --> 00:37:09,920
a protest with new york times so that they can publish it on their newspaper that doesn't mean

648
00:37:10,220 --> 00:37:14,598
anything like i work for the new york times it could just be that let's say you henry went to

649
00:37:14,620 --> 00:37:18,780
protest somewhere in the United States, took a bunch of videos and want to share it with a bunch

650
00:37:18,780 --> 00:37:22,440
of New York news organizations. And you shared it with Wall Street Journal and New York Times and

651
00:37:22,560 --> 00:37:27,580
whatnot. Doesn't mean anything. So it doesn't mean you work for those organizations and we won't know

652
00:37:27,800 --> 00:37:33,780
who you work for. All we know is that one person shared another, shared an album with another

653
00:37:34,000 --> 00:37:39,598
person. But we just want to make sure that there's only two persons. And the only way to ensure that

654
00:37:39,600 --> 00:37:41,200
that we could come up with is using email.

655
00:37:41,810 --> 00:37:43,480
Because if we use IP addresses,

656
00:37:44,220 --> 00:37:45,560
again, VPNs could be a problem.

657
00:37:46,200 --> 00:37:48,620
If we use only email addresses

658
00:37:48,860 --> 00:37:50,440
and don't bind them to IP addresses,

659
00:37:51,180 --> 00:37:52,780
then what happens is that

660
00:37:53,210 --> 00:37:54,940
what if you just create one email inbox

661
00:37:55,130 --> 00:37:57,220
and thousands of people just use that email inbox?

662
00:37:57,500 --> 00:37:58,380
That's a problem.

663
00:37:58,940 --> 00:38:00,520
So from our perspective,

664
00:38:00,610 --> 00:38:03,820
if we want to prevent the spread of CSAM and MAS

665
00:38:04,100 --> 00:38:07,200
while still respecting your privacy and your security

666
00:38:07,380 --> 00:38:09,080
and keeping your content and to end encrypted,

667
00:38:09,140 --> 00:38:14,640
this seemed like a fair compromise to say hey if the people you are who you are sharing these images

668
00:38:14,880 --> 00:38:20,580
with are your family that is completely fine if they if they use their email addresses so 90 of

669
00:38:20,680 --> 00:38:25,920
the users who will share the photos with their families it's good if you're a war photographer

670
00:38:26,200 --> 00:38:31,040
and we had calls with various different war photographers one of which is from finland and

671
00:38:31,040 --> 00:38:35,118
i'm very happy to say that he's been an avid supporter of everything that we're doing he was

672
00:38:35,140 --> 00:38:40,060
in South Sudan and he went to Ukraine and he was telling us like, hey, this is how we think you

673
00:38:40,180 --> 00:38:44,000
should do it. He seemed to be okay with this idea. We fact-checked it with a bunch of other war

674
00:38:44,180 --> 00:38:47,840
photographers with the United States just to be on the safe side that this is not just a European

675
00:38:48,210 --> 00:38:53,060
concept and maybe the Americans are more sensitive. But no, it seems that they understand and they say,

676
00:38:53,120 --> 00:38:56,840
hey, look, this is completely reasonable that my editor would need to put in their email address.

677
00:38:57,440 --> 00:39:01,278
If they're so concerned that we would find out, then the editor could just put in a random Gmail

678
00:39:01,280 --> 00:39:02,780
address and we write it in the New York Times.

679
00:39:03,820 --> 00:39:04,580
It's very simple.

680
00:39:05,280 --> 00:39:06,820
If they want to use a burner email,

681
00:39:07,060 --> 00:39:09,120
bind by me. The whole idea here is that

682
00:39:09,620 --> 00:39:11,400
it's binded to an IP address and we know

683
00:39:11,620 --> 00:39:13,400
it's one person accessing it, not multiple people.

684
00:39:14,300 --> 00:39:15,220
I think the

685
00:39:15,360 --> 00:39:17,240
clarification here, and actually we do the same thing

686
00:39:17,780 --> 00:39:19,200
in our Plexus app,

687
00:39:19,340 --> 00:39:20,940
the open source app that you can get from F-Droid.

688
00:39:20,980 --> 00:39:23,400
It's how you score de-Google

689
00:39:23,460 --> 00:39:25,220
apps and see which ones work so you

690
00:39:25,320 --> 00:39:27,160
know before you move to a custom ROM what apps work.

691
00:39:27,800 --> 00:39:29,240
But we also require an email

692
00:39:29,500 --> 00:39:30,198
and pretty much

693
00:39:31,100 --> 00:39:32,440
we don't care what the email is.

694
00:39:32,690 --> 00:39:33,740
You can use an alias email.

695
00:39:33,850 --> 00:39:34,700
It literally doesn't matter.

696
00:39:34,750 --> 00:39:36,360
The whole point is that there is at least

697
00:39:36,500 --> 00:39:39,280
a unique identifier to mark yourself down

698
00:39:39,390 --> 00:39:41,440
so that we can prevent people from spinning up

699
00:39:41,970 --> 00:39:43,140
tons of bot accounts

700
00:39:43,960 --> 00:39:46,180
and then just spamming ratings onto this platform.

701
00:39:46,320 --> 00:39:48,540
We have to keep the ratings legit.

702
00:39:49,380 --> 00:39:50,740
It's a very different implication,

703
00:39:50,870 --> 00:39:52,360
but it's kind of the same problem, actually.

704
00:39:52,720 --> 00:39:52,960
Exactly.

705
00:39:53,250 --> 00:39:55,740
I don't love it, but we don't have a better way

706
00:39:55,860 --> 00:39:57,560
of identifying people on the internet.

707
00:39:57,870 --> 00:39:59,420
So for as long as email lives,

708
00:39:59,580 --> 00:40:01,220
And I hope email lives forever.

709
00:40:01,320 --> 00:40:02,260
I think email is amazing.

710
00:40:02,480 --> 00:40:03,200
It's interoperable.

711
00:40:03,300 --> 00:40:04,500
We don't have to worry about,

712
00:40:04,920 --> 00:40:07,360
oh, I want to send you an email from Gmail to Hotmail,

713
00:40:07,580 --> 00:40:09,160
but sorry, they don't interoperate.

714
00:40:09,520 --> 00:40:11,900
I can't email you from my Gmail to your Hotmail.

715
00:40:12,080 --> 00:40:12,940
There is no such thing.

716
00:40:13,200 --> 00:40:14,480
It just works across providers.

717
00:40:15,140 --> 00:40:16,660
So for as long as email lives,

718
00:40:17,200 --> 00:40:18,420
I think this is going to be a great solution.

719
00:40:18,820 --> 00:40:21,120
And if you're very concerned about your privacy,

720
00:40:21,640 --> 00:40:24,680
that, hey, I cannot share my email with you

721
00:40:25,000 --> 00:40:27,540
for no fucking reason, never, ever,

722
00:40:28,440 --> 00:40:29,420
then get a burner email.

723
00:40:29,480 --> 00:40:32,580
Geez, if that's your threat model, you should have a burner email anyway.

724
00:40:33,040 --> 00:40:37,780
I was going to ask you about aliases, because I assume privacy.com would work,

725
00:40:37,850 --> 00:40:40,120
like alias cards, alias emails.

726
00:40:40,840 --> 00:40:45,140
The point isn't, and this is again the same thing with Plexus on our ends,

727
00:40:45,240 --> 00:40:49,420
the point isn't to try to identify people, we just need some way to verify

728
00:40:49,570 --> 00:40:51,760
that 10 ratings came from one person.

729
00:40:53,360 --> 00:40:57,280
So that way we know if one person is spamming, we can just shut that one down.

730
00:40:57,820 --> 00:41:00,020
which I don't think we've had many problems with.

731
00:41:00,060 --> 00:41:02,440
I think that system has actually prevented a lot of the issues.

732
00:41:02,440 --> 00:41:05,100
I mean, look, privacy.com, as great as it is,

733
00:41:05,440 --> 00:41:06,500
they also have limitations.

734
00:41:06,800 --> 00:41:09,280
They stop you from creating more than five accounts

735
00:41:09,360 --> 00:41:11,180
on the same platform, if I remember correctly.

736
00:41:11,240 --> 00:41:13,380
So they have limitations for the exact same reasons.

737
00:41:13,500 --> 00:41:15,180
They don't want their platform to be abused.

738
00:41:15,660 --> 00:41:17,320
So everyone's trying to do the right thing,

739
00:41:17,420 --> 00:41:19,300
and everyone's trying to put some artificial limits

740
00:41:19,400 --> 00:41:21,820
to prevent abuse, and there is a lot of abuse.

741
00:41:22,580 --> 00:41:25,620
So this is the best solution that we could come up with.

742
00:41:25,700 --> 00:41:26,480
I don't love it.

743
00:41:26,720 --> 00:41:29,900
I would have rather made it so that you can just make one link

744
00:41:30,060 --> 00:41:32,540
and anyone on the other end can just access it freely.

745
00:41:32,840 --> 00:41:34,220
But the world is a horrible place.

746
00:41:34,390 --> 00:41:38,440
And unfortunately, there's a lot of people trying to use these types of platforms for awful things.

747
00:41:38,530 --> 00:41:39,980
And I don't want to take part of it.

748
00:41:40,240 --> 00:41:45,940
So the way we decided to go about it is it may not be the most privacy-sensitive way,

749
00:41:46,090 --> 00:41:47,980
perhaps, if you have to give a payment method in an email.

750
00:41:48,250 --> 00:41:50,320
But then if you want to store your photos on Cryptee,

751
00:41:50,660 --> 00:41:53,200
and if you want to store more than the free plan, which is 100 megabytes,

752
00:41:53,300 --> 00:41:54,780
you're going to have to give a payment method anyway.

753
00:41:55,480 --> 00:42:00,260
And if the recipients want to receive anything, be it on WeTransfer or on WhatsApp, they're

754
00:42:00,280 --> 00:42:01,720
going to have to give a phone number or an email.

755
00:42:02,080 --> 00:42:04,100
And I would rather that it's an email than a phone number.

756
00:42:04,320 --> 00:42:05,280
So I think it's a lot nicer.

757
00:42:06,380 --> 00:42:10,640
Yeah, maybe, you know, before we zoom out a little bit here, just threat modeling in

758
00:42:10,760 --> 00:42:15,580
general, I feel like this, I wasn't going to name call anybody, but I think it's quite

759
00:42:15,840 --> 00:42:20,280
relevant because Proton had the recent activist who they were forced to hand over.

760
00:42:20,460 --> 00:42:21,460
I don't even think it was direct.

761
00:42:21,720 --> 00:42:22,540
I think it was indirect.

762
00:42:22,820 --> 00:42:24,980
I think the FBI tapped on the Swiss authorities.

763
00:42:25,060 --> 00:42:28,700
The Swiss authorities demanded a specific user account from Proton,

764
00:42:28,700 --> 00:42:31,840
and they were able to get some basic information about their payment method,

765
00:42:32,500 --> 00:42:34,740
and then sent that to the Swiss authorities, sent it to the FBI.

766
00:42:35,000 --> 00:42:36,120
But of course, the headline was,

767
00:42:37,540 --> 00:42:40,060
Proton just gave information to the FBI, which is willy-nilly.

768
00:42:40,220 --> 00:42:41,560
But I think this is the same thing.

769
00:42:42,280 --> 00:42:45,960
There was a really good article put out from Freedom of the Press about this incident,

770
00:42:46,140 --> 00:42:48,740
which pretty much said, Proton is not anonymous.

771
00:42:49,260 --> 00:42:51,780
Here is the actual threat model that they're accounting for.

772
00:42:52,300 --> 00:42:54,300
And it talks about how Proton has never jeopardized

773
00:42:54,800 --> 00:42:55,340
internet encrypted emails.

774
00:42:55,800 --> 00:42:57,200
That is always something that they've stuck to.

775
00:42:57,320 --> 00:42:59,900
They've never handed over email content.

776
00:43:00,580 --> 00:43:01,840
But they've handed over an IP address.

777
00:43:02,260 --> 00:43:05,060
They've handed over recovery email in the past.

778
00:43:05,120 --> 00:43:06,860
They've handed over payment information.

779
00:43:07,920 --> 00:43:09,580
These aren't things that they promise to keep safe,

780
00:43:09,820 --> 00:43:10,920
but it's a threat modeling question.

781
00:43:11,020 --> 00:43:13,280
And so I don't know how this ties into you.

782
00:43:13,380 --> 00:43:15,640
How do you view the threat model of a Crypteee user

783
00:43:16,280 --> 00:43:18,960
and what realistically a user can and cannot expect?

784
00:43:19,020 --> 00:43:21,560
Because I think those are healthy expectations to set for users.

785
00:43:22,060 --> 00:43:24,920
We have a lot of this documented on our security page.

786
00:43:25,090 --> 00:43:27,180
I highly encourage everyone to go take a look at that.

787
00:43:27,360 --> 00:43:28,860
And we have a privacy policy.

788
00:43:29,720 --> 00:43:32,080
I think that is so simple that we have two chapters.

789
00:43:32,360 --> 00:43:33,980
There's like one for lawyers and one for humans.

790
00:43:34,180 --> 00:43:37,420
So you can go very simply read what we keep.

791
00:43:37,470 --> 00:43:37,960
They're not humans.

792
00:43:39,740 --> 00:43:41,200
Lawyers are not humans, trust me.

793
00:43:44,800 --> 00:43:48,060
So basically you can just go read our privacy policy

794
00:43:48,420 --> 00:43:51,058
and hopefully get a better idea about what we keep

795
00:43:51,380 --> 00:43:53,740
and what we promise to encrypt and keep safe

796
00:43:54,380 --> 00:43:56,440
and what we cannot keep encrypted.

797
00:43:56,540 --> 00:43:59,660
Like we can't keep your payment name encrypted

798
00:43:59,900 --> 00:44:02,700
because that's not something we keep in the first place.

799
00:44:02,920 --> 00:44:05,080
That's our payment processor Stripe, for example.

800
00:44:05,340 --> 00:44:08,320
So there are things, and the same goes for Proton, right?

801
00:44:08,480 --> 00:44:09,820
Like they use Stripe under the hood.

802
00:44:09,940 --> 00:44:10,840
So at the end of the day,

803
00:44:10,880 --> 00:44:12,940
their payment processor also has to have these names.

804
00:44:13,280 --> 00:44:15,380
And I know from having spoken to Andy

805
00:44:15,640 --> 00:44:17,000
and we do exactly the same thing.

806
00:44:17,140 --> 00:44:19,298
So if you essentially work with us or with Proton

807
00:44:19,300 --> 00:44:21,700
or I think most privacy respecting services,

808
00:44:22,340 --> 00:44:24,780
we essentially don't give you a real email to Stripe

809
00:44:25,080 --> 00:44:28,340
and we give an alias email to Stripe if we can,

810
00:44:28,740 --> 00:44:30,440
where we can and where we're able to,

811
00:44:30,560 --> 00:44:32,240
unless you want like a special billing address

812
00:44:32,480 --> 00:44:33,680
to receive your emails.

813
00:44:34,040 --> 00:44:35,240
But the whole point here is that

814
00:44:35,620 --> 00:44:37,740
that's not necessarily a vector in and of itself

815
00:44:38,220 --> 00:44:39,220
where you could be identified.

816
00:44:39,900 --> 00:44:45,720
And no one, this includes Proton, us, Signal, anyone, right?

817
00:44:45,920 --> 00:44:47,498
We are legal entities

818
00:44:47,500 --> 00:44:49,860
and we operate within the framework of law.

819
00:44:50,380 --> 00:44:51,840
And there is this thing called

820
00:44:52,350 --> 00:44:53,640
mutual legal assistant treaties.

821
00:44:54,120 --> 00:44:55,740
And we have to obey this

822
00:44:56,100 --> 00:44:58,660
if we receive an international court order legally.

823
00:44:59,180 --> 00:45:01,280
And it would have to go through multiple different courts.

824
00:45:01,330 --> 00:45:02,940
And that's literally the reason

825
00:45:03,100 --> 00:45:04,800
why I moved to Estonia to start crypto

826
00:45:05,000 --> 00:45:06,400
because it's got some of the strongest

827
00:45:06,620 --> 00:45:07,560
privacy protections in place.

828
00:45:08,280 --> 00:45:09,980
We do not have key disclosure laws

829
00:45:10,290 --> 00:45:11,620
unlike most other European countries

830
00:45:11,880 --> 00:45:13,680
like Germany has got a more sensitive situation.

831
00:45:14,080 --> 00:45:15,280
No one can come to Estonia,

832
00:45:15,620 --> 00:45:19,160
pass a mutual legal assistance treaty paper and say,

833
00:45:19,280 --> 00:45:23,640
hey, we want your courts to issue an order to Crypteed to give us access.

834
00:45:24,460 --> 00:45:27,120
And essentially, Estonia can only enforce us to say,

835
00:45:27,240 --> 00:45:29,760
hey, we just need you to give the most basic,

836
00:45:30,720 --> 00:45:32,860
whatever unencrypted stuff that you have access to,

837
00:45:32,910 --> 00:45:34,680
and that's it because we don't hold your keys.

838
00:45:35,300 --> 00:45:36,660
We cannot give anything else.

839
00:45:37,020 --> 00:45:38,680
So we're in the same position as Proton, right?

840
00:45:38,980 --> 00:45:42,460
If someone comes to us and says, hey, you have to legally obey this.

841
00:45:42,840 --> 00:45:43,880
It's a criminal court order.

842
00:45:44,640 --> 00:45:47,460
There's a criminal order issued, so we have to obey it.

843
00:45:47,940 --> 00:45:53,040
And to put an indifference perspective, if someone gave you three euros a month and said,

844
00:45:53,120 --> 00:45:55,720
hey, I want you to break the law for me, would you do that?

845
00:45:56,020 --> 00:45:56,420
You wouldn't.

846
00:45:57,320 --> 00:46:02,300
So no one's going to, neither us nor Proton nor any other company is going to break the

847
00:46:02,300 --> 00:46:03,640
law for you for three euros a month.

848
00:46:03,880 --> 00:46:05,880
If you're Apple, you have billions at your disposal.

849
00:46:06,260 --> 00:46:09,500
Maybe you can break the law and get away with it because you can afford to pay for it.

850
00:46:10,120 --> 00:46:10,580
And they do.

851
00:46:10,880 --> 00:46:11,620
Same with Meta.

852
00:46:11,770 --> 00:46:14,480
They just get a fine, a slap in the wrist for like five million.

853
00:46:15,000 --> 00:46:15,920
and they can get away with that.

854
00:46:16,200 --> 00:46:17,420
But in our case, that's not the case.

855
00:46:17,720 --> 00:46:19,400
So if someone comes to us and says,

856
00:46:19,450 --> 00:46:21,820
hey, we need you to give access to everything you have,

857
00:46:22,260 --> 00:46:23,620
that's going to be your email address.

858
00:46:24,480 --> 00:46:27,040
That's going to be your account meta information,

859
00:46:27,620 --> 00:46:30,120
which is like the last time you logged in, for example,

860
00:46:30,740 --> 00:46:32,840
which is something we have to keep for your security

861
00:46:32,890 --> 00:46:34,820
to make sure that we can keep your account safe

862
00:46:35,460 --> 00:46:36,760
or things like the IP address

863
00:46:37,260 --> 00:46:38,520
so that we can make sure that

864
00:46:39,040 --> 00:46:40,620
while you're logged in in the United States,

865
00:46:40,870 --> 00:46:42,640
someone doesn't try to suspiciously log in

866
00:46:42,770 --> 00:46:44,500
to your account from Czech Republic

867
00:46:44,560 --> 00:46:45,780
that's not cool either.

868
00:46:46,000 --> 00:46:47,940
Someone needs to be able to take a look at it digitally

869
00:46:48,060 --> 00:46:50,060
and see, hey, this doesn't look right.

870
00:46:50,260 --> 00:46:51,840
Henley normally accesses from the US.

871
00:46:52,340 --> 00:46:54,080
All of a sudden, he popped in Czech Republic

872
00:46:54,360 --> 00:46:56,340
and something's fishy here.

873
00:46:56,980 --> 00:46:58,200
Maybe your phone is stolen.

874
00:46:58,440 --> 00:47:00,200
But there needs to be a system

875
00:47:00,380 --> 00:47:02,360
that can safe keep your account as well.

876
00:47:02,700 --> 00:47:04,280
And these are all compromises, right?

877
00:47:04,360 --> 00:47:06,160
We also want the security aspect of things.

878
00:47:06,460 --> 00:47:09,120
So we tried to find a middle ground

879
00:47:09,260 --> 00:47:10,420
and our promise to our users

880
00:47:10,560 --> 00:47:12,820
is that your data is always encrypted.

881
00:47:13,180 --> 00:47:14,060
We don't hold your keys.

882
00:47:14,480 --> 00:47:18,440
you hold your keys, and we're going to do everything we can to keep your account safe.

883
00:47:18,660 --> 00:47:21,220
That's the best we can do, that's the best Proton can do,

884
00:47:21,970 --> 00:47:24,940
and under any law framework, under any legal framework,

885
00:47:25,180 --> 00:47:26,500
that's what any company can do.

886
00:47:27,360 --> 00:47:30,300
Unless you live somewhere on a pirate boat

887
00:47:30,410 --> 00:47:33,820
and you just land yourself completely lawless.

888
00:47:34,560 --> 00:47:36,600
I don't think you should trust them to keep your data safe.

889
00:47:36,880 --> 00:47:39,500
Yeah, I'd be curious how their servers work on the ocean.

890
00:47:40,280 --> 00:47:40,920
Beats me, Matt.

891
00:47:41,000 --> 00:47:41,840
Maybe easier to cool.

892
00:47:42,100 --> 00:47:44,740
You can strap the servers on the bottom of the boat

893
00:47:45,040 --> 00:47:49,100
and water-cooled servers.

894
00:47:49,500 --> 00:47:51,180
I have a very tinfoil hat theory

895
00:47:51,260 --> 00:47:53,060
that I haven't verified or talked to anyone about.

896
00:47:53,160 --> 00:47:55,100
But I think the hype around,

897
00:47:55,600 --> 00:47:58,120
let's send data centers to the satellites,

898
00:47:58,720 --> 00:48:00,100
to space, and all that hype

899
00:48:00,540 --> 00:48:01,480
somewhat has to do with that

900
00:48:01,560 --> 00:48:03,020
because it would technically be lawless.

901
00:48:05,600 --> 00:48:07,000
Okay, that's very possible.

902
00:48:07,780 --> 00:48:09,440
Any company on planet Earth

903
00:48:09,680 --> 00:48:11,300
would still need to operate those satellites

904
00:48:11,300 --> 00:48:13,640
and they would still need to be under some legal jurisdiction.

905
00:48:13,770 --> 00:48:14,960
So there's some gray area there.

906
00:48:15,559 --> 00:48:16,580
Yeah, I'm sorry.

907
00:48:16,770 --> 00:48:17,800
More questions keep coming up.

908
00:48:18,220 --> 00:48:20,120
One thing that came up, what are your feelings?

909
00:48:20,460 --> 00:48:22,120
I think I briefly mentioned this at one point.

910
00:48:22,500 --> 00:48:25,120
Mega, the end-to-end encrypted cloud provider,

911
00:48:25,440 --> 00:48:27,000
which also has pretty good photo support.

912
00:48:27,180 --> 00:48:28,580
I use them for photos for quite a while.

913
00:48:28,800 --> 00:48:32,340
They have end-to-end encrypted sharing via a public link.

914
00:48:32,770 --> 00:48:34,140
So I'm curious what your thoughts are on that.

915
00:48:34,260 --> 00:48:38,380
And actually, on this note, it's funny because Mega was criticized in the past.

916
00:48:38,470 --> 00:48:40,600
I think it was like a semi-viral Reddit thread.

917
00:48:41,300 --> 00:48:43,420
accusing Mega of not having end-to-end encryption

918
00:48:43,900 --> 00:48:45,420
because they were banning accounts

919
00:48:45,470 --> 00:48:47,540
that were spreading CSAM using their platform.

920
00:48:48,620 --> 00:48:50,020
When really, when Mega cleared up,

921
00:48:50,050 --> 00:48:51,860
they said, well, these are public URLs.

922
00:48:53,660 --> 00:48:56,340
Like, if you're sharing a public URL,

923
00:48:56,760 --> 00:48:57,680
even if there's an encryption,

924
00:48:57,930 --> 00:48:59,840
like the encryption key is literally in the URL.

925
00:48:59,900 --> 00:49:01,140
In the URL, exactly.

926
00:49:01,460 --> 00:49:03,380
So it's not really end-to-end encrypted

927
00:49:03,640 --> 00:49:04,740
once you have a public link.

928
00:49:04,980 --> 00:49:06,020
So I don't know if you have thoughts

929
00:49:06,050 --> 00:49:07,460
on kind of the way they approach that

930
00:49:08,500 --> 00:49:09,580
and what that looks like.

931
00:49:10,280 --> 00:49:10,360
Yeah.

932
00:49:10,740 --> 00:49:12,360
There is no public sharing with us.

933
00:49:12,760 --> 00:49:16,140
So I can at least say that with us, it wouldn't be the same

934
00:49:16,360 --> 00:49:18,680
because the idea is that you shared the link with someone else

935
00:49:19,040 --> 00:49:20,600
and only that other person can access it.

936
00:49:20,820 --> 00:49:23,400
That being said, if you grab a Cryptee photo sharing link,

937
00:49:23,600 --> 00:49:25,020
which has a key in the URL also,

938
00:49:25,400 --> 00:49:28,480
and then you just go and post that link to your MySpace account

939
00:49:28,780 --> 00:49:29,940
because that's what's fashionable,

940
00:49:30,760 --> 00:49:33,020
then if we end up running into that link,

941
00:49:33,460 --> 00:49:35,340
yeah, of course we can see the contents.

942
00:49:35,600 --> 00:49:38,400
If we do see it within the first 48 hours,

943
00:49:38,600 --> 00:49:40,760
because then after that, the invite will expire

944
00:49:40,820 --> 00:49:42,100
and we wouldn't be able to access it either.

945
00:49:42,600 --> 00:49:44,340
Only the person who verified their email

946
00:49:44,620 --> 00:49:46,040
and added it into the system will be able to.

947
00:49:48,420 --> 00:49:49,380
I guess a better question is,

948
00:49:49,440 --> 00:49:50,860
why didn't you go with mega system?

949
00:49:51,020 --> 00:49:52,500
What issues do you have with that system

950
00:49:52,640 --> 00:49:54,060
and why doesn't it align with what you want?

951
00:49:55,040 --> 00:49:55,580
Mass spread.

952
00:49:55,900 --> 00:49:58,900
We just didn't want our system to be used

953
00:49:59,140 --> 00:50:03,200
for potential use of mass spread of CSAM.

954
00:50:03,680 --> 00:50:05,600
Because if you can spread images with a link,

955
00:50:06,080 --> 00:50:08,820
it could potentially lead to unlimited number of users

956
00:50:08,910 --> 00:50:10,200
being able to access those images.

957
00:50:10,370 --> 00:50:13,840
Then all it would take is to share a CSAM album.

958
00:50:13,890 --> 00:50:16,920
You just grab a link, post it on a form, and that's it.

959
00:50:17,500 --> 00:50:18,900
Thousands of people can access it.

960
00:50:19,210 --> 00:50:22,020
That sort of defeats the purpose of trying to prevent mass spread.

961
00:50:22,500 --> 00:50:25,800
So our goal here is to clamp down on that

962
00:50:25,960 --> 00:50:28,760
and limit it to maximum five number of users,

963
00:50:28,810 --> 00:50:30,040
and that's the whole email thing.

964
00:50:30,360 --> 00:50:33,260
Yeah, and I think if I was running a service,

965
00:50:34,360 --> 00:50:35,940
Even if, you know, because I'm thinking,

966
00:50:36,170 --> 00:50:37,860
like if I'm running Henry Photos

967
00:50:38,720 --> 00:50:40,080
and an encrypted photo provider,

968
00:50:40,320 --> 00:50:41,880
if I did the public sharing method,

969
00:50:42,740 --> 00:50:43,120
I'm good.

970
00:50:43,900 --> 00:50:45,360
That sounds like a lot of work.

971
00:50:45,530 --> 00:50:46,800
And I mean, yeah, we have our PeerTube server.

972
00:50:47,130 --> 00:50:48,040
We host videos, man.

973
00:50:48,900 --> 00:50:51,200
But it's completely, you know, my moderation platform

974
00:50:51,540 --> 00:50:55,300
or my moderation protocol for our PeerTube?

975
00:50:55,620 --> 00:50:56,340
No public account.

976
00:50:56,410 --> 00:50:57,820
Like it's just for our content

977
00:50:58,130 --> 00:50:59,820
because I'm not going to deal with moderating that

978
00:51:00,000 --> 00:51:01,180
for similar reasons as you guys.

979
00:51:01,460 --> 00:51:04,200
But if I ran this service and, you know,

980
00:51:04,280 --> 00:51:05,460
I started getting reports.

981
00:51:05,880 --> 00:51:07,300
Let's say we had a reporting system.

982
00:51:07,940 --> 00:51:08,920
We maxed it out.

983
00:51:09,060 --> 00:51:11,400
So if we noticed more than 100 downloads on a file,

984
00:51:11,920 --> 00:51:12,780
we can set up algorithms.

985
00:51:12,890 --> 00:51:13,960
We can try to automate this.

986
00:51:13,970 --> 00:51:15,920
We can do a lot to try to prevent this kind of problem.

987
00:51:15,990 --> 00:51:16,900
But at the end of the day,

988
00:51:17,400 --> 00:51:21,200
you're not going to get around you and your team needing.

989
00:51:21,600 --> 00:51:24,240
I would need to have either myself or a team member

990
00:51:24,520 --> 00:51:27,720
or a whole team dedicated to having to access those links

991
00:51:28,380 --> 00:51:31,620
to see if it's actually appropriate content or not.

992
00:51:32,020 --> 00:51:34,640
And so now you have to have the moderation teams,

993
00:51:34,800 --> 00:51:37,980
and we've seen the mental health issues that come from these big tech companies

994
00:51:38,100 --> 00:51:44,860
that outsource this to third-world countries and people there

995
00:51:44,960 --> 00:51:47,720
who already aren't in a great situation,

996
00:51:48,040 --> 00:51:49,060
and they're having to look at this,

997
00:51:49,060 --> 00:51:50,420
and they have done studies on these people,

998
00:51:50,480 --> 00:51:53,640
and it's disastrous to their mental health.

999
00:51:53,820 --> 00:51:54,680
It's permanent damage.

1000
00:51:55,400 --> 00:51:56,920
It's just a nasty problem to deal with.

1001
00:51:57,020 --> 00:51:58,940
So that's why I would probably avoid that.

1002
00:51:59,260 --> 00:52:00,100
But I don't know how to handle that.

1003
00:52:00,460 --> 00:52:01,160
I'd like to ask them.

1004
00:52:01,240 --> 00:52:05,920
huge reason why we don't want to ever deal with that either and i don't think anyone wants to

1005
00:52:06,220 --> 00:52:11,180
and look we're based in northern europe we have extremely high quality of life standards i'm like

1006
00:52:11,400 --> 00:52:16,800
in a situation where i would not want to put any of my co-workers or any of my team members in a

1007
00:52:16,900 --> 00:52:22,320
situation of hey someone needs to look at c sam like that's just horrible nobody should ever have

1008
00:52:22,440 --> 00:52:27,640
to do this but if there ever comes a point where someone needs to it should be trained professionals

1009
00:52:27,840 --> 00:52:31,800
It shouldn't be us who have no training, no psychological training or support.

1010
00:52:32,100 --> 00:52:33,400
It should not be us.

1011
00:52:34,460 --> 00:52:38,460
And if there ever comes a day where we have to have a team, then I would also have to

1012
00:52:38,560 --> 00:52:41,960
hire a bunch of psychologists to train and support the staff.

1013
00:52:42,340 --> 00:52:44,260
It's not going to be just the staff who does this.

1014
00:52:44,880 --> 00:52:48,900
I would want for each person who does the support to be a dedicated psychologist or a

1015
00:52:49,060 --> 00:52:50,020
therapist to be able to assist him.

1016
00:52:50,060 --> 00:52:51,140
And it's not an easy situation.

1017
00:52:51,840 --> 00:52:54,540
And I don't think anybody should ever have to do this.

1018
00:52:54,620 --> 00:52:55,320
It's pretty horrible.

1019
00:52:55,340 --> 00:52:58,960
And I think the solution to this is to try to prevent mass spread of this.

1020
00:52:59,280 --> 00:53:01,400
And that's a big start.

1021
00:53:02,260 --> 00:53:05,500
And if we can all do our best to try to clamp down on that,

1022
00:53:05,660 --> 00:53:07,500
the world would be a much better place in the first place.

1023
00:53:08,700 --> 00:53:11,640
Yeah, so, you know, finally zooming out,

1024
00:53:12,000 --> 00:53:14,660
as I was promising we'd get to at some point.

1025
00:53:15,960 --> 00:53:16,520
I'm doing good.

1026
00:53:16,740 --> 00:53:19,460
What liability, and maybe this is more philosophical,

1027
00:53:19,780 --> 00:53:21,300
more of your personal views here,

1028
00:53:22,160 --> 00:53:24,360
do encrypted services have liability?

1029
00:53:24,960 --> 00:53:29,240
Is there anything wrong with building a tool?

1030
00:53:29,940 --> 00:53:34,140
I mean, I think Tornado Cash is maybe a decent example here.

1031
00:53:34,900 --> 00:53:37,920
So Tornado Cash was the, I might get some of the details wrong.

1032
00:53:38,140 --> 00:53:41,980
My understanding, it was a mixing service on the Ethereum blockchain, I believe.

1033
00:53:43,020 --> 00:53:50,160
And the idea is you can send funds into Tornado Cash and it spits it out and it essentially mixes those funds so you don't know where it came from.

1034
00:53:50,300 --> 00:53:53,760
It's a privacy service to make cryptocurrency a little bit more private.

1035
00:53:54,460 --> 00:53:58,580
Those guys, both of them, I believe they got thrown in jail, or at least they were tried.

1036
00:53:58,640 --> 00:54:01,900
But I think they were arrested, and that's still an unfolding case.

1037
00:54:02,160 --> 00:54:04,160
But did they have liability, right?

1038
00:54:04,380 --> 00:54:09,020
Like, if it was, let's say, just 1%, I don't know what the exact amount was.

1039
00:54:09,140 --> 00:54:12,860
They cite different numbers, but it's just for the example, right?

1040
00:54:13,220 --> 00:54:16,780
What liability do developers have when it comes to the software they're building?

1041
00:54:16,920 --> 00:54:18,040
And is that their decision?

1042
00:54:18,340 --> 00:54:20,220
Is that their problem?

1043
00:54:20,940 --> 00:54:23,660
Like, do you see a personal liability building a service?

1044
00:54:24,260 --> 00:54:25,340
How do you grapple with this?

1045
00:54:25,390 --> 00:54:27,120
How do you think other people might grapple with this?

1046
00:54:27,980 --> 00:54:29,280
Let me reframe the question.

1047
00:54:29,890 --> 00:54:33,160
Because I think liability and responsibility are two different things.

1048
00:54:34,840 --> 00:54:37,620
Being responsible and legally liable are two different things.

1049
00:54:37,920 --> 00:54:40,700
Maybe there is no law that regulates me to do certain things,

1050
00:54:41,100 --> 00:54:43,900
but I should still have a moral responsibility to do the right thing.

1051
00:54:44,160 --> 00:54:46,520
And I think this is the most important distinction here,

1052
00:54:46,680 --> 00:54:50,820
is that there are countries where the age of consent is so fucking low

1053
00:54:51,180 --> 00:54:53,940
that it almost makes you question the meaning of the word consent.

1054
00:54:54,340 --> 00:54:58,300
and there are countries where the age of consent is high and they respect people's consent.

1055
00:54:58,860 --> 00:55:03,500
And what I'm trying to get to by saying this is that I don't want to be the arbiter of what the

1056
00:55:03,620 --> 00:55:08,560
legal age of consent is because every country has a different number, right? But it's very important

1057
00:55:08,780 --> 00:55:13,540
that whatever the law says doesn't matter. I have a moral responsibility aside from the legal

1058
00:55:13,800 --> 00:55:19,559
liability. And it is that I want to do the morally right thing. I don't want to do just what's legally

1059
00:55:19,580 --> 00:55:24,640
okay and just move on and just wash my hands and think this is fine. The best way to look at this,

1060
00:55:24,760 --> 00:55:29,160
I think, is to perhaps use a physical analogy rather than a digital one. I think encryption

1061
00:55:29,800 --> 00:55:35,060
is a dual-use technology, right? You can use it to keep personal data safe, but you can also use it

1062
00:55:35,080 --> 00:55:41,360
in militaries to keep military communications safe, etc., in the context of war or to harm someone

1063
00:55:41,500 --> 00:55:47,640
with it, to distribute bad content with it. So encryption is dual-use. But guns, for example,

1064
00:55:47,700 --> 00:55:54,860
are not exactly dual use. It's just built to kill or to harm people. A similar analogy is knives,

1065
00:55:55,240 --> 00:56:00,880
right? There are kitchen knives you can buy from Ikea to cook and feed your family, and you can

1066
00:56:01,200 --> 00:56:06,760
exactly go to Ikea, buy a large kitchen knife, and hurt someone with it. Does this mean Ikea should

1067
00:56:06,790 --> 00:56:11,700
be responsible for anyone who ever attempts a knife murder? Like, I don't think so. I don't think

1068
00:56:11,700 --> 00:56:16,119
it's Ikea's responsibility if someone buys a kitchen knife and kills someone with it. But this doesn't

1069
00:56:16,140 --> 00:56:20,260
To solve this problem, we just go around banning all sales of kitchen knives.

1070
00:56:21,240 --> 00:56:25,900
We're also not saying, hey, we're going to ID every single person who buys kitchen knives at IKEA either.

1071
00:56:26,060 --> 00:56:27,440
That's not a solution either here.

1072
00:56:29,200 --> 00:56:32,960
Maybe if we age verified kitchen knives, it would solve the problem.

1073
00:56:33,240 --> 00:56:35,460
You know what I saw this morning in my RSS feed?

1074
00:56:35,480 --> 00:56:38,080
A company wants to age verify vapes.

1075
00:56:38,540 --> 00:56:44,200
And I'm like, oh my god, you know the meme where men will do X before going to therapy?

1076
00:56:44,460 --> 00:56:48,060
You know this meme? Men will literally, and then it's something ridiculous before going to therapy.

1077
00:56:48,880 --> 00:56:52,200
It's kind of the same thing. It's like, we will literally age verify

1078
00:56:52,680 --> 00:56:56,340
a vape before holding vape companies accountable. We will literally

1079
00:56:57,040 --> 00:57:00,200
age verify every platform in the world before holding big tech companies

1080
00:57:00,440 --> 00:57:04,100
accountable. It is so, yeah, sorry, I didn't mean to steal your thunder.

1081
00:57:04,430 --> 00:57:07,760
That just came to mind. Totally, man. But that's exactly what I'm getting at, is that

1082
00:57:08,560 --> 00:57:12,380
I don't think IKEA is responsible for selling a kitchen knife. In the same

1083
00:57:12,400 --> 00:57:16,700
way I don't think we're responsible for selling another dual-use technology like encryption.

1084
00:57:17,720 --> 00:57:24,120
So what we do is we store your photos and secure them with encryption and someone can use it for

1085
00:57:24,300 --> 00:57:28,660
good or bad in the same way someone can use a kitchen knife they bought from Ikea for good or

1086
00:57:28,820 --> 00:57:34,320
bad. So that's the distinction. If we were selling guns it would be inherently bad and I think we

1087
00:57:34,320 --> 00:57:39,099
should ID everyone and raise the age limit for guns and maybe actually do a mental sanity check

1088
00:57:39,120 --> 00:57:42,980
at least in the United States, like in Europe, we don't seem to have this problem, but just saying.

1089
00:57:43,600 --> 00:57:50,920
So what I'm trying to get to here is I don't think beyond that moral responsibility to do the right

1090
00:57:51,040 --> 00:57:57,280
thing, to make sure that we do everything within our power to make sure that our platform is not

1091
00:57:57,380 --> 00:58:03,040
used to spread bad content and put restrictions in place to make sure we have the reasonable

1092
00:58:03,300 --> 00:58:07,600
pipelines to comply with the law. If someone comes to us and says, hey, this needs to be taken down,

1093
00:58:07,680 --> 00:58:14,800
we can take it down but do so in a responsible manner where we tried our very fucking best to

1094
00:58:14,800 --> 00:58:19,660
make sure that someone can't spread bad content on our platform that is the best we can do and

1095
00:58:19,700 --> 00:58:25,220
unfortunately that's not even the tip of the iceberg for just how bad the situation is for

1096
00:58:25,600 --> 00:58:29,760
the opposite of the spectrum which is like facebook or google photos and whatnot there is

1097
00:58:30,350 --> 00:58:36,440
millions and billions of csm material being shared on these platforms precisely because it's open for

1098
00:58:37,080 --> 00:58:42,860
anyone and everyone to upload content. And they're not doing this type of due diligence. So it's all

1099
00:58:42,980 --> 00:58:49,200
about doing what we can, the very fucking best we can to hold that moral responsibility bar as high

1100
00:58:49,360 --> 00:58:55,580
as we can, even if the legal or, I don't know, lawful frameworks of Europe or Estonia doesn't

1101
00:58:55,760 --> 00:59:00,840
require us to. They don't tell us, hey, you need to have this, this, this, this, this, and this type

1102
00:59:00,840 --> 00:59:05,160
of checks in place to be able to do end-to-end encrypted photo sharing. They don't. There is no

1103
00:59:05,180 --> 00:59:09,080
a law that really requires us to do that. So we don't really need to do any of those things,

1104
00:59:09,490 --> 00:59:15,640
but we want to do this because this is the right thing to do. And to that extent, yes, we are

1105
00:59:15,780 --> 00:59:21,840
liable. And I think we should, and all platforms should be held to a higher standard so that they

1106
00:59:22,080 --> 00:59:26,100
do the right thing. And not because the law requires them to, but because it is our moral

1107
00:59:26,420 --> 00:59:32,740
responsibility. So if I make kitchen knives that also have fucking lasers and laser guided aiming

1108
00:59:32,760 --> 00:59:38,000
and whatever. And it's very clearly a murder weapon and not just for cutting cucumbers, man,

1109
00:59:38,000 --> 00:59:42,000
I don't know. Like that's the line where it gets really blurry. So I just want to be very

1110
00:59:42,260 --> 00:59:47,300
responsible with what we do. And I would like to think after having spoken to enough people that

1111
00:59:47,440 --> 00:59:53,320
we're trying our best and I think we're doing good, but I do think the bar should be a lot higher.

1112
00:59:53,940 --> 00:59:58,940
And I think platforms should be held to a greater standard where all these civil rights groups and

1113
00:59:59,040 --> 01:00:02,680
women's rights organizations and children's rights organizations should come to us and say,

1114
01:00:02,740 --> 01:00:06,120
Say, hey, we need you to add this other thing to make this a bit better.

1115
01:00:06,620 --> 01:00:10,300
You don't have to compromise on encryption, but we would love if you could just do this too.

1116
01:00:10,720 --> 01:00:13,900
And that's the kind of dialogue that we're open to and we want to have.

1117
01:00:13,900 --> 01:00:17,740
And we've spoken to as many as we could, and we want to speak to as many more as we can.

1118
01:00:18,220 --> 01:00:23,360
So that we try to cover all our bases and hope to be able to do the right thing here.

1119
01:00:25,140 --> 01:00:29,180
Yeah, so I don't remember exactly what the thread was called.

1120
01:00:29,320 --> 01:00:32,500
I just remember there being a pretty hot thread back on our forum.

1121
01:00:32,600 --> 01:00:33,640
before it was archived.

1122
01:00:33,860 --> 01:00:36,360
Someone came on our forum and pretty much asked,

1123
01:00:36,500 --> 01:00:39,740
hey, how do we actually talk about the CSAM problem?

1124
01:00:40,200 --> 01:00:41,300
How do we all feel about it?

1125
01:00:41,720 --> 01:00:43,040
It was kind of just a vibe check,

1126
01:00:43,120 --> 01:00:44,460
and someone was kind of expressing

1127
01:00:44,540 --> 01:00:46,300
and having some of these thoughts of like,

1128
01:00:46,440 --> 01:00:48,500
hey, what if these encrypted services

1129
01:00:48,780 --> 01:00:50,860
do have this negative side to them?

1130
01:00:51,080 --> 01:00:53,420
Is there something that we should be doing about this?

1131
01:00:53,500 --> 01:00:54,980
Is there something we can do about this?

1132
01:00:55,240 --> 01:00:58,280
AKA, is there some legitimacy to the whole idea

1133
01:00:58,360 --> 01:00:59,900
that encryption can harm children?

1134
01:01:00,300 --> 01:01:03,240
And it wasn't a productive thread, I think.

1135
01:01:03,450 --> 01:01:06,440
Like, a lot of people were very absolute in saying,

1136
01:01:06,510 --> 01:01:07,940
like, no, we cannot fight encryption.

1137
01:01:08,320 --> 01:01:09,160
This is not our fault.

1138
01:01:09,320 --> 01:01:09,980
This is not our problem.

1139
01:01:10,860 --> 01:01:12,780
Bad things will be done.

1140
01:01:13,400 --> 01:01:16,000
You know, I'm sure you can imagine how that went.

1141
01:01:16,740 --> 01:01:19,180
Yeah, and just to pause you for a second,

1142
01:01:19,650 --> 01:01:21,320
encryption is not what harms children.

1143
01:01:21,680 --> 01:01:24,000
It's the people who use encryption that harms children.

1144
01:01:24,440 --> 01:01:27,120
And in the same way, it's not knives that harm children.

1145
01:01:27,500 --> 01:01:29,640
It's people who use the knives that harm the children.

1146
01:01:30,160 --> 01:01:33,640
We don't put knives into jail, we put the people who use the knives into jail.

1147
01:01:33,970 --> 01:01:38,220
And yeah, I think it's an education problem, not exactly a technical problem.

1148
01:01:39,120 --> 01:01:45,760
So how can people who are listening to this, who are, maybe they've already been thinking about this,

1149
01:01:46,030 --> 01:01:48,740
or maybe they're listening to this and going, wow, I disagree with all of this.

1150
01:01:48,890 --> 01:01:53,780
But either way, how can people engage in a discussion from here on out in a way that's less awkward

1151
01:01:54,240 --> 01:01:57,560
and that appreciates the complexity of the situation?

1152
01:01:57,660 --> 01:02:05,540
How can people, because I would like to think after this is recorded, because I'm learning a lot from this as well.

1153
01:02:05,540 --> 01:02:10,160
I'm trying to, again, we had this 10-minute spiel and now we've been recording for over an hour now.

1154
01:02:11,180 --> 01:02:16,120
If our forum was still live and someone made this thread, what would you suggest that discussion looks like now?

1155
01:02:16,280 --> 01:02:21,200
Both for someone who's posting the thread and somebody who's responding to it to have a productive discussion about this.

1156
01:02:21,960 --> 01:02:43,660
Look, I think the hardest, harshest thing I've learned is after having gone to Brussels and having spoken to the regulators and having spoken to these experts and women's rights orgs and children's rights orgs, one thing became clear that wasn't clear for me in the past, which is just how bad the situation is.

1157
01:02:44,120 --> 01:02:52,080
it's not as light or as simple as we think it is the the severity of CSAM on the internet is a lot

1158
01:02:52,280 --> 01:02:57,800
worse than we realize and it's getting worse progressively these rights groups are monitoring

1159
01:02:57,830 --> 01:03:01,920
the situation and I'm not talking about law enforcement I'm talking about women's rights

1160
01:03:01,920 --> 01:03:08,100
and children's rights organizations these don't have the inherent instinct to hey we want to do

1161
01:03:08,360 --> 01:03:12,459
widespread surveillance their goal is not to do surveillance they're not law enforcement they're

1162
01:03:12,480 --> 01:03:19,500
another government. They're just people trying to help kids and women. It's bad. It's really,

1163
01:03:19,640 --> 01:03:27,140
really bad. And we don't wake up every day and have to deal with the reality of a horribly broken

1164
01:03:27,440 --> 01:03:32,860
internet where people share CSAM and mess with millions of people, but they do. Every morning,

1165
01:03:33,360 --> 01:03:37,360
in the same way you and I wake up and open up YouTube and maybe grab a cup of coffee and watch

1166
01:03:37,360 --> 01:03:42,440
a home renovation video, they have to go to work and spend eight hours having to deal with CSAM

1167
01:03:42,720 --> 01:03:49,680
at scale of millions and they just see how bad it is we don't so understandably so there is this gap

1168
01:03:50,700 --> 01:03:55,000
between people who don't see that side of the internet and people who do see that side of the

1169
01:03:55,180 --> 01:03:59,860
internet and it's a massive communication gap because nobody wants to talk about this topic

1170
01:04:00,100 --> 01:04:03,980
nobody really wants to bring this topic up it's one of those things where it's like it's dark we

1171
01:04:04,100 --> 01:04:11,800
know it we just hide it under the the rug and just move on and i think it's it's horrible that

1172
01:04:12,340 --> 01:04:19,640
the internet is this broken and we don't have this conversation especially as tech founders and

1173
01:04:21,340 --> 01:04:25,640
when i sat down to try to talk with the folks in the european commission when i sat down to talk

1174
01:04:25,650 --> 01:04:31,260
with the folks with digital rights orgs and others and especially during the heat of the chat control

1175
01:04:31,520 --> 01:04:36,980
debate there was an inherent eyebrow raise moment where they were like why are you asking these like

1176
01:04:37,000 --> 01:04:43,660
what is it that you're trying to sort of gauge here like it's bad but i was like look guys i

1177
01:04:44,140 --> 01:04:49,040
understand it's bad but just how bad like because that matters if it's one person in the whole world

1178
01:04:49,180 --> 01:04:52,740
that we're worried about like maybe in the grand scheme of things it's not that bad but if it's

1179
01:04:53,160 --> 01:04:58,160
millions of people it's bad and if it's billions of people it's fucking out of control right so it's

1180
01:04:58,240 --> 01:05:04,760
there's got to be a scale of just how bad it's severity and and i think it depends on that if

1181
01:05:04,780 --> 01:05:09,520
we realize, hey, there's 8 billion people on the planet and 4 billion of them are abusing children,

1182
01:05:09,800 --> 01:05:14,520
well, we have to do something more drastic about it. But if it's, I don't know, 100,000, a million,

1183
01:05:14,780 --> 01:05:18,280
whatnot, then we should think about compromises and figure out how to solve this problem in

1184
01:05:18,390 --> 01:05:21,500
different ways, perhaps because maybe encryption itself isn't the right way to solve this.

1185
01:05:22,080 --> 01:05:26,960
So once I went to talk to these people, to regulators, to digital rights organizations,

1186
01:05:27,300 --> 01:05:33,120
to civil rights organizations, it was then I understood just how bad. And that was sort of

1187
01:05:33,140 --> 01:05:37,980
the awakening moment for me because as a tech founder who's in his own tech bubble and in his

1188
01:05:38,160 --> 01:05:42,720
own friend group where things are jolly and we don't have to worry about these types of things

1189
01:05:42,900 --> 01:05:48,220
day to day i don't know just how bad it was i don't get the cds every day so i had no idea and i don't

1190
01:05:48,280 --> 01:05:53,560
think any one of us do hopefully none of us have to do this unless you're a content moderator then

1191
01:05:53,940 --> 01:05:58,360
the more power to you thank you for keeping the internet clean at the the cost of your personal

1192
01:05:58,360 --> 01:06:04,660
sanity. It's pretty horrible out there. And this is the type of thing that I think I didn't realize.

1193
01:06:05,020 --> 01:06:11,740
And to address your question, I don't know what the right set of questions to ask are or the right

1194
01:06:11,840 --> 01:06:16,380
kinds of conversations that we should have around this topic are. But I think it has to do with

1195
01:06:16,700 --> 01:06:21,580
education first. We all have to first be educated to understand just how bad it is and go from there.

1196
01:06:22,240 --> 01:06:28,340
In the same way we started to, thankfully, in our generation, have the conversation around

1197
01:06:28,360 --> 01:06:33,220
mental health issues and slowly people are becoming more and more aware of mental health

1198
01:06:33,440 --> 01:06:38,460
issues and perhaps our parents' generation didn't have as much in-depth of an understanding of

1199
01:06:38,570 --> 01:06:43,540
mental health issues and they're now becoming more aware of the situation. I think that this is the

1200
01:06:43,700 --> 01:06:48,900
same kind of conversation that we need to start to have to say, hey, this is an ongoing issue,

1201
01:06:49,210 --> 01:06:53,780
this is a widespread problem and we need to do something about this. And only then we can maybe

1202
01:06:54,020 --> 01:06:58,320
start to understand, okay, these are the types of conversations we should have. These are the types of

1203
01:06:58,340 --> 01:07:03,540
compromises that we should have. Maybe 15 years ago, 20 years ago, I don't know, ADHD wasn't as

1204
01:07:03,560 --> 01:07:08,400
well understood, right? We didn't know how to accommodate people with ADHD. And now we understand

1205
01:07:08,460 --> 01:07:12,680
it a lot better. We have a lot better dialogues around it, and we have a lot better solutions

1206
01:07:13,000 --> 01:07:19,220
around it as a result of those conversations. So the way I see this is the same. It's an issue that

1207
01:07:19,420 --> 01:07:24,180
we need to talk about, ideally at scale, and everyone needs to be better educated about it.

1208
01:07:24,340 --> 01:07:39,220
Everyone needs to have a better understanding about what causes this and how we can solve this and how we can address this in the interim in the meantime so that we don't have to worry about a much worse problem coming into the next 10, 20 years until we can actually find a solution to address this problem.

1209
01:07:39,840 --> 01:07:43,220
So, yeah, I would say that education is perhaps the first step.

1210
01:07:43,270 --> 01:07:48,900
In the same way we manage to educate our societies, not to go around stabbing people with kitchen knives and just use them to cook.

1211
01:07:49,090 --> 01:07:51,420
I mean, sure, there is a number of people who probably do that.

1212
01:07:51,550 --> 01:07:54,220
We need to do the same exact type of education with this.

1213
01:07:54,620 --> 01:07:57,780
then it's going to start somewhere, as dark as it is.

1214
01:07:58,340 --> 01:08:00,680
It reminds me of just privacy concerns in general.

1215
01:08:02,100 --> 01:08:05,340
If somebody was trying to have a discussion about privacy concerns,

1216
01:08:05,470 --> 01:08:07,680
but both people aren't aware of how bad it really is

1217
01:08:07,970 --> 01:08:09,360
and what the implications are,

1218
01:08:09,370 --> 01:08:11,000
then you can't really have a discussion about it.

1219
01:08:11,240 --> 01:08:12,640
I don't think it's any different from that.

1220
01:08:13,860 --> 01:08:15,660
If anyone here cares about privacy,

1221
01:08:15,860 --> 01:08:18,000
but you're talking to someone and they don't seem to care at all,

1222
01:08:18,140 --> 01:08:20,200
what's missing there is their understanding

1223
01:08:20,450 --> 01:08:22,240
of what's actually happening behind the scenes.

1224
01:08:22,319 --> 01:08:26,319
And once you can explain that, then you can have an actual discussion about how to solve the problem.

1225
01:08:26,580 --> 01:08:32,120
So it almost sounds like that would be a good starting point of like, well, actually citing how bad the problem is.

1226
01:08:32,420 --> 01:08:35,120
So everyone in the room can at least acknowledge, okay, yeah, this is a problem.

1227
01:08:35,960 --> 01:08:38,160
And now we can talk more about what a solution looks like.

1228
01:08:38,880 --> 01:08:42,440
And to add to that point, I think the difficulty here is with privacy.

1229
01:08:42,799 --> 01:08:45,480
It's a bit easier to actually be able to express this to people, right?

1230
01:08:45,600 --> 01:08:49,240
You can show people, see, this is how they take your data out of your phone.

1231
01:08:49,440 --> 01:08:56,940
Or they could be like, hey, see, this is exactly the kind of thing that happens where you go to that link and see this network inspector in your browser.

1232
01:08:57,200 --> 01:08:58,880
This is the kinds of data that they upload.

1233
01:08:59,080 --> 01:08:59,520
It's bad.

1234
01:08:59,640 --> 01:09:01,960
Like you can actually visually show some of the stuff.

1235
01:09:02,560 --> 01:09:05,700
But with CSAM, you kind of can't.

1236
01:09:06,140 --> 01:09:09,600
That's sort of what makes this very difficult is that nobody wants to see this.

1237
01:09:09,740 --> 01:09:10,140
It's horrible.

1238
01:09:11,060 --> 01:09:12,940
You can just cite numbers, I suppose.

1239
01:09:13,400 --> 01:09:14,339
You can cite numbers.

1240
01:09:14,520 --> 01:09:24,740
But when you get to the point of, hey, there is millions and billions on the internet, then that's a number so great that it's incomprehensible to our reptile brains that unfortunately we don't understand.

1241
01:09:24,859 --> 01:09:26,380
And I think that's where it gets really difficult.

1242
01:09:26,740 --> 01:09:29,819
So that's why I think it's a dialogue like mental health.

1243
01:09:29,880 --> 01:09:45,000
We just have to talk to everyone and understand and be able to sort of have these difficult conversations and hopefully read more news articles that are authored by these women's rights and children's rights organizations and hopefully get a better understanding of the situation.

1244
01:09:45,319 --> 01:09:50,500
I guess on this note, I guess the regulation side of it, chat control has been a big thing.

1245
01:09:50,980 --> 01:09:55,980
And you can expand on maybe what happened last week and why it's kind of died down quite a bit.

1246
01:09:56,020 --> 01:09:58,860
I am asking this question all the time because I've been doing coverage for chat control.

1247
01:09:58,940 --> 01:10:00,260
I'm doing coverage for age verification.

1248
01:10:00,420 --> 01:10:03,900
I'm doing coverage for all these different issues happening globally right now on the channel.

1249
01:10:04,300 --> 01:10:07,840
And I just am so lost on some of this regulation.

1250
01:10:08,460 --> 01:10:09,680
Some of it just seems stupid.

1251
01:10:10,060 --> 01:10:16,220
Like a lot of the stuff that comes out of these U.S. states, I'm just like, okay, this is clearly some person who's never used a computer.

1252
01:10:16,920 --> 01:10:20,040
And this just sounded like a good idea on paper and they signed off on it.

1253
01:10:20,220 --> 01:10:21,960
And they just do not understand even where.

1254
01:10:22,400 --> 01:10:28,500
Because even if that was their goal, if their goal was to ban VPNs, if their goal was to be repressive and do all these crazy things,

1255
01:10:29,600 --> 01:10:33,500
the bills themselves and how they're written aren't actually going to accomplish those things.

1256
01:10:33,970 --> 01:10:35,740
And anyone would be able to tell them that.

1257
01:10:36,690 --> 01:10:38,660
But on the other hand, I'm still conflicted.

1258
01:10:38,980 --> 01:10:44,740
People who are trying to pass chat control, people who are trying to break encryption, do they have ulterior motives?

1259
01:10:45,340 --> 01:10:49,560
Are they using this child thing to get what they want, which is something different?

1260
01:10:49,900 --> 01:10:52,460
Or do they just genuinely care about this topic?

1261
01:10:52,720 --> 01:10:55,160
They're all in the room and they agree on this child problem.

1262
01:10:55,660 --> 01:10:59,000
They just don't understand the implications of what it would mean to ban an encryption.

1263
01:10:59,120 --> 01:11:04,340
And we, everyone, just kind of disagrees on whether or not banning encryption is part of that solution.

1264
01:11:04,860 --> 01:11:08,660
I don't know if you can speak to a litmus test based on especially your time speaking to a lot of these people.

1265
01:11:09,880 --> 01:11:11,180
Okay, it's very complex.

1266
01:11:11,500 --> 01:11:14,800
So I'll start by saying that I think the answer is yes and no.

1267
01:11:15,080 --> 01:11:22,860
Because when we talk about groups like European Commission, European Parliament, they represent a very vast spectrum of political parties.

1268
01:11:23,100 --> 01:11:28,660
For those of you living in the United States, Europe has a multi-party system where it's not just two parties.

1269
01:11:29,040 --> 01:11:35,360
We have multiple parties, all forming coalitions and sort of representing everyone all at the same time.

1270
01:11:35,560 --> 01:11:38,180
There is no two-party battles going on, so to speak.

1271
01:11:38,420 --> 01:11:41,280
And I think this is very important because when we talk about the European Commission,

1272
01:11:41,350 --> 01:11:46,000
we sort of sum everyone up into this tiny little narrow slice of an organization thinking,

1273
01:11:46,150 --> 01:11:47,480
oh, that's everyone.

1274
01:11:47,860 --> 01:11:51,080
That's what everyone in the European Commission must be thinking, but it's not.

1275
01:11:51,380 --> 01:11:55,800
And European Commission consists of some incredibly talented and smart people,

1276
01:11:56,070 --> 01:12:00,620
at least based on my personal experience having interacted with them at the Digital Markets Act team.

1277
01:12:00,690 --> 01:12:04,100
And they have experts with PhDs in computer science.

1278
01:12:04,380 --> 01:12:05,300
They are not politicians.

1279
01:12:05,610 --> 01:12:07,460
They are experts in their own domain.

1280
01:12:07,820 --> 01:12:11,040
law, computer science, whatever have you, whatever it is that they're going to regulate,

1281
01:12:11,090 --> 01:12:12,160
they have experts in these topics.

1282
01:12:12,800 --> 01:12:17,420
So that straight off the bat is very different from a bunch of lawyers sitting around a table

1283
01:12:17,640 --> 01:12:18,920
trying to decide what it is.

1284
01:12:18,920 --> 01:12:21,140
They actually have people who understand what's going on.

1285
01:12:21,300 --> 01:12:24,020
And they consult others who understand what's going on.

1286
01:12:24,380 --> 01:12:29,300
They consult the civil society and human rights organizations and women's rights organizations.

1287
01:12:29,650 --> 01:12:33,400
And if they're in doubt, they talk to digital rights organizations like EDRI.

1288
01:12:33,800 --> 01:12:36,140
If they're in doubt, they talk to Free Software Foundation Europe,

1289
01:12:36,380 --> 01:12:37,920
like folks like Lucas that you interviewed.

1290
01:12:38,210 --> 01:12:40,020
And I don't remember if you had the chance to speak

1291
01:12:40,050 --> 01:12:41,780
the Yann Penfrat from EDRI.

1292
01:12:43,180 --> 01:12:45,740
Amazing folks that they basically bring into the room

1293
01:12:45,770 --> 01:12:46,620
to have these conversations.

1294
01:12:47,700 --> 01:12:51,000
So long story short, this is a very divisive topic.

1295
01:12:51,560 --> 01:12:54,100
And the way it was packaged when it was presented

1296
01:12:54,620 --> 01:12:57,980
was literally, if you look up the name of chat control,

1297
01:12:58,180 --> 01:12:59,880
which is just an alias that we've given

1298
01:13:01,480 --> 01:13:03,280
amongst privacy conscious folks,

1299
01:13:03,560 --> 01:13:05,520
but the actual name of the legislation

1300
01:13:05,520 --> 01:13:08,880
is to prevent the distribution of child sexual abuse or whatever.

1301
01:13:08,970 --> 01:13:13,700
Like, they just named it in such a way that if you say no,

1302
01:13:13,790 --> 01:13:17,060
it just makes you sound like you are one of the bad people.

1303
01:13:17,210 --> 01:13:19,920
Like, it's just the way it was packaged and wrapped already

1304
01:13:20,140 --> 01:13:24,080
from the name and the get-go is just a little bit hostile,

1305
01:13:24,600 --> 01:13:30,200
which should, rightfully so, sort of raise the sort of, like,

1306
01:13:30,400 --> 01:13:32,580
awareness and maybe ring the alarm bells a little bit

1307
01:13:32,770 --> 01:13:34,740
to wonder who it is that proposed this.

1308
01:13:34,880 --> 01:13:46,280
If I recall correctly, it's been such a fucking saga with roller coasters of ups and downs and who pitched it and then who shapeshifted and then who added some new segments to that thing.

1309
01:13:46,360 --> 01:13:47,540
It's like, it's a mess.

1310
01:13:47,720 --> 01:13:55,900
But if I remember correctly, the initial party from Denmark who proposed this said, hey, we should do something like this and pitch this idea.

1311
01:13:56,120 --> 01:14:03,160
one of their key political members of the party who pitched this actually got arrested for child

1312
01:14:03,320 --> 01:14:09,400
sexual abuse related issues. So that should tell you everything you really need to know about

1313
01:14:10,060 --> 01:14:14,100
what's going on. So someone should look this up and find the links or I can just try to send you a

1314
01:14:14,200 --> 01:14:19,360
link and give you a better idea of what's actually going on here. It's quite ironic. So with that in

1315
01:14:19,380 --> 01:14:27,000
mind, I don't know if the key goal here was to, in fact, keep kids safe. At least maybe for that

1316
01:14:27,200 --> 01:14:32,460
person, it wasn't. But it's very important to acknowledge that there are so many people,

1317
01:14:33,460 --> 01:14:35,980
as a part of the European Commission, as a part of the European Parliament,

1318
01:14:36,560 --> 01:14:43,340
as regulators, as civil society orgs, as women's rights orgs, whose goal 100% was to keep kids safe.

1319
01:14:43,420 --> 01:14:51,140
That is undeniable. That being said, when chat control took its new evolved Pikachu to whatever

1320
01:14:51,600 --> 01:14:57,380
crazy monster form it took, it turned into a bit of an interesting surveillance legislation that

1321
01:14:57,580 --> 01:15:02,700
took a really strange turn in a really strange shape where initially it was pushed out by

1322
01:15:03,500 --> 01:15:09,540
religious and conservative parties, coalitions in Europe, which sort of made it a bit easier to

1323
01:15:09,560 --> 01:15:16,820
track down who it is that actually lobbied for it. And to spare everyone the legal complications of

1324
01:15:16,900 --> 01:15:22,000
how lobbying actually works in Europe, there are some protections in place to make sure that there

1325
01:15:22,100 --> 01:15:27,580
is no corporate influence in European politics. But needless to say, it is not sufficient to the

1326
01:15:27,820 --> 01:15:31,440
extent that you would expect. And I can give you one personal example of this from the Digital

1327
01:15:31,640 --> 01:15:36,940
Markets Act, where every single Digital Markets Act hearing that I went to, to spend seven hours in

1328
01:15:36,960 --> 01:15:42,040
front of a camera talking and yelling at Apple lawyers. There were a bunch of rights organizations,

1329
01:15:42,500 --> 01:15:51,740
air quotes, who were there to essentially just repeat Apple's marketing taglines. And we realized

1330
01:15:51,860 --> 01:15:57,640
that those rights organizations were actually entirely funded by Apple. So sure. I remember

1331
01:15:57,860 --> 01:16:04,280
seeing one of these that they put out these like really formal looking PDF reports on why

1332
01:16:04,280 --> 01:16:07,440
European legislation was harmful to Apple users.

1333
01:16:07,870 --> 01:16:10,340
I think it was maybe in the context of Google search.

1334
01:16:10,960 --> 01:16:12,620
They were showing Google search results

1335
01:16:12,840 --> 01:16:15,700
and how they have to have these pop-ups in Europe

1336
01:16:15,810 --> 01:16:16,880
that don't exist in the US.

1337
01:16:17,200 --> 01:16:19,680
And we found that users were 30% more dissatisfied

1338
01:16:20,030 --> 01:16:21,520
and we care a lot about this topic

1339
01:16:21,670 --> 01:16:23,140
and we think that Europe should back off.

1340
01:16:23,780 --> 01:16:25,160
But when you looked at who funded it,

1341
01:16:25,190 --> 01:16:26,300
it was big tech companies.

1342
01:16:26,970 --> 01:16:28,000
Sorry, I didn't mean to cut you off.

1343
01:16:28,520 --> 01:16:29,760
It's good, but that's the thing.

1344
01:16:29,900 --> 01:16:31,840
Essentially, that's how it creeps in.

1345
01:16:32,240 --> 01:16:38,160
And it has the shape and the face of a right organization, XYZ, fill the blank.

1346
01:16:38,410 --> 01:16:41,540
But it is, in fact, behind the scenes funded by a tech company.

1347
01:16:42,040 --> 01:16:50,360
When we looked at, out of curiosity, who it is that is actually funding the organizations pushing for these chat control-like issues,

1348
01:16:50,700 --> 01:16:53,700
one of the key names that showed up was Thorn.

1349
01:16:54,040 --> 01:16:57,040
And it keeps coming up, this name, I'm sure you've heard of it.

1350
01:16:57,160 --> 01:17:07,440
It's a company that was initially funded by Ashton Kutcher, and it was a startup that specializes in utilizing AI tools to scan images in video to detect CSAM.

1351
01:17:07,720 --> 01:17:10,040
And they are essentially a CSAM scanning company.

1352
01:17:10,600 --> 01:17:23,980
And if they could theoretically lobby to make this happen in Europe, and messengers were required to scan these messages, there are no bigger service providers out there who could theoretically handle the volume.

1353
01:17:24,460 --> 01:17:28,300
So it would be Thorne who would be handling this and they would be making millions from this.

1354
01:17:28,580 --> 01:17:35,760
So needless to say, Thorne was quite often and very frequently and vocally involved in lobbying for things like this.

1355
01:17:36,280 --> 01:17:45,680
There are other organizations that I don't think fully understood what was going on and lobbied for it without fully actually understanding the context of what it would entail.

1356
01:17:46,240 --> 01:17:49,360
But I think, again, it boils down to this communication issue, right?

1357
01:17:49,520 --> 01:17:58,260
Like there weren't enough technical people in the room to sometimes explain, hey, in order to do that, you have to just break the whole world's encryption and just make everyone insecure.

1358
01:17:58,390 --> 01:17:59,260
And that's not a solution.

1359
01:17:59,620 --> 01:18:04,820
That's just to give you a little sample taster of how some of this came to be.

1360
01:18:05,040 --> 01:18:18,600
And eventually, when you talk to enough people and name the legislation in a specific way and talk about the topic with the framework of protecting kids, I think it makes it very difficult for people to have these conversations at scale.

1361
01:18:18,780 --> 01:18:25,140
until everyone's in the room and they can anonymously and maybe secretly push the vote no but

1362
01:18:25,360 --> 01:18:29,280
if you go to the corridor and the halls of european commission and want to have these conversations

1363
01:18:29,600 --> 01:18:33,280
with people then they're a little more hesitant or the european parliament they're a bit more

1364
01:18:33,440 --> 01:18:39,040
hesitant because they are like hey like i don't want to say no fuck protecting kids we shouldn't

1365
01:18:39,180 --> 01:18:43,300
break encryption like that's not any what anyone wants to say and that's not their view either

1366
01:18:43,600 --> 01:18:47,860
everybody wants to protect kids the point is not everyone wants to break encryption it's that

1367
01:18:47,880 --> 01:18:53,920
conversation the nuance of it was lost in a lot of these conversations and it's it's a very nuanced

1368
01:18:53,920 --> 01:18:58,520
and very difficult problem of how it came to be and it took a very different shape and eventually

1369
01:18:58,780 --> 01:19:04,160
it got to a point so serious that in some countries the parties that were secretly pushing to make

1370
01:19:04,230 --> 01:19:09,360
this happen and the conversations that were secretly happening behind the scenes not so secret

1371
01:19:09,560 --> 01:19:13,739
but secret as in the context of behind the scenes public doesn't directly engage with

1372
01:19:14,440 --> 01:19:18,480
it took really interesting shapes like defense ministries of certain countries got involved

1373
01:19:18,940 --> 01:19:24,660
they were like hey maybe this is not a good idea like months before this happened the swedish

1374
01:19:24,880 --> 01:19:28,800
military and the defense ministry said hey we're going to start using signal and then a few weeks

1375
01:19:28,980 --> 01:19:33,960
later someone else from the swedish ministry said hey we will like the support check control and

1376
01:19:34,080 --> 01:19:39,440
then the swedish defense was like uh what no like we just said we're gonna use it we don't want

1377
01:19:39,680 --> 01:19:43,720
wait hold on a second and it just really complicated situations and that should go to

1378
01:19:43,740 --> 01:19:50,140
how complex the issue is. So all this is to say, it took a very interesting shape. And I ended up

1379
01:19:50,280 --> 01:19:56,120
running around trying to talk to as many regulators as I could within my reach in the languages that I

1380
01:19:56,240 --> 01:20:01,520
could speak in the fluency that I could communicate with. And it's a very difficult topic that I don't

1381
01:20:01,600 --> 01:20:05,740
want to make mistakes and have it get lost in translation. It was very difficult. And there

1382
01:20:05,900 --> 01:20:09,580
aren't that many experts who can talk about these things or have the connections. Maybe there are the

1383
01:20:09,820 --> 01:20:13,160
experts, but they just don't know how to reach out to their regulators. So it's very difficult,

1384
01:20:13,360 --> 01:20:19,000
very nuanced. And it took a while, man. I mean, Meredith was in Brussels. I think she was in

1385
01:20:19,040 --> 01:20:23,880
Amsterdam talking to all these people as much as she could. She was going to all the seminars. Like

1386
01:20:24,000 --> 01:20:28,200
I was talking to as many people as I could. I'm sure folks at Proton were doing the same,

1387
01:20:29,100 --> 01:20:34,820
Tutai were doing the same. We were all engaged at what capacity or another trying to do the right

1388
01:20:34,920 --> 01:20:40,159
thing. So it took a while, but I think we managed to convince the right people to have the right

1389
01:20:40,180 --> 01:20:46,080
conversations to sit down and say maybe this was a bad idea in the first place and i think there are

1390
01:20:46,200 --> 01:20:52,800
very few people who could take more credit than people like patrick or people like marcel from the

1391
01:20:52,920 --> 01:20:59,600
pirate party who did their very best to try to explain the reality of what this would entail and

1392
01:20:59,900 --> 01:21:04,860
made magic happen so and there's tons of people who i don't even know their names of someone created

1393
01:21:04,880 --> 01:21:09,720
the Stop Chat Control website, I think, from Denmark. And everybody used that to send these

1394
01:21:09,880 --> 01:21:14,280
emails to the European Parliament and European Commission. Whoever you are, you're an amazing

1395
01:21:14,580 --> 01:21:18,420
anonymous person. Thank you for making that website, because that basically was the tipping

1396
01:21:18,640 --> 01:21:24,840
point, I think. And one person single-handedly made it possible to fight this battle. So it's

1397
01:21:24,890 --> 01:21:29,760
very complex. And I think it's just the fact that it's a very difficult topic to talk about

1398
01:21:29,780 --> 01:21:35,200
made it very difficult for everyone to engage in it also. Yeah, so last week, it was about one or

1399
01:21:35,200 --> 01:21:40,720
two weeks ago, Patrick kind of cited it as mostly dead. I'm still trying to wrap my head around it.

1400
01:21:40,730 --> 01:21:46,960
So can you explain the most recent update very briefly? So essentially this last week, what

1401
01:21:47,240 --> 01:21:53,920
happened is there were two votes held. And to the best of my understanding, the one on Thursday was

1402
01:21:54,090 --> 01:21:58,720
final and the European Commission essentially, sorry, European Parliament essentially killed it.

1403
01:21:59,000 --> 01:22:00,440
That's to the best of my understanding.

1404
01:22:01,180 --> 01:22:05,860
And yeah, that's at least the best of my understanding.

1405
01:22:05,960 --> 01:22:10,060
But again, this is one of those topics that it will come back.

1406
01:22:10,740 --> 01:22:14,080
I can tell you already, this is not a final win.

1407
01:22:14,540 --> 01:22:23,180
This will need to be drafted into every single independent country's laws to say you cannot compromise on people's privacy.

1408
01:22:23,560 --> 01:22:25,040
And it is their human right.

1409
01:22:25,320 --> 01:22:28,760
It is their national right to respect their privacy.

1410
01:22:28,840 --> 01:22:32,080
and it cannot be altered with.

1411
01:22:32,520 --> 01:22:34,240
It should be in law, codified in law,

1412
01:22:34,460 --> 01:22:36,720
for it to never be possibly brought up again.

1413
01:22:37,100 --> 01:22:38,880
Even then, the European Parliament might try.

1414
01:22:39,400 --> 01:22:41,180
Certain parties might try to push their agendas.

1415
01:22:41,560 --> 01:22:42,220
It could happen.

1416
01:22:42,380 --> 01:22:43,960
It's been, in this particular case,

1417
01:22:43,980 --> 01:22:45,720
the right-wing and the Christian Democrat parties,

1418
01:22:45,960 --> 01:22:47,480
and tomorrow it could be someone else.

1419
01:22:47,920 --> 01:22:49,880
But point being, for now, we're good,

1420
01:22:50,000 --> 01:22:50,920
to the best of my understanding.

1421
01:22:51,320 --> 01:22:53,300
And on the verge of learning that,

1422
01:22:53,420 --> 01:22:55,180
we thought maybe now is the right time

1423
01:22:55,220 --> 01:22:56,760
to launch our sharing features

1424
01:22:56,780 --> 01:22:59,160
because we've been waiting very long and thought,

1425
01:22:59,240 --> 01:23:00,420
okay, this is the right time.

1426
01:23:00,640 --> 01:23:04,040
If there's ever a better time, it's Thursday when they announced it.

1427
01:23:04,660 --> 01:23:07,540
They announced that it's dead, and an hour later we released sharing.

1428
01:23:07,900 --> 01:23:09,740
So we were literally waiting for it.

1429
01:23:10,780 --> 01:23:11,720
Yeah, one quick note.

1430
01:23:12,060 --> 01:23:16,080
I've seen, on my end, it might swing a little bit, plus or minus 5%,

1431
01:23:16,600 --> 01:23:19,940
but I've seen chat control, encryption bans,

1432
01:23:20,100 --> 01:23:21,520
all this stuff proposed from both sides.

1433
01:23:21,640 --> 01:23:24,200
I know that you mentioned right quite a bit in yours,

1434
01:23:24,520 --> 01:23:28,600
But in the U.S., I've seen left-wing politicians try to pass it.

1435
01:23:28,650 --> 01:23:32,880
I've also seen in Europe some people on the left also try to pass this kind of stuff.

1436
01:23:32,930 --> 01:23:36,440
So I just wanted to make that clear because, again, another thing, back in the forum days,

1437
01:23:36,650 --> 01:23:39,440
there was a lot of pointing fingers of right versus left.

1438
01:23:39,510 --> 01:23:42,700
And someone who was right-wing would point at the left-wing people and say,

1439
01:23:42,780 --> 01:23:46,980
well, actually, this other bill in the neighboring country is from the other side.

1440
01:23:47,080 --> 01:23:48,780
This is a pretty bipartisan thing.

1441
01:23:49,830 --> 01:23:52,580
The important thing, though, I think it's worth mentioning these things

1442
01:23:52,600 --> 01:23:56,660
because if you voted for the Christian Democrats or any one of the right-wing parties, first of all,

1443
01:23:56,760 --> 01:24:01,720
I have utmost respect for you. Vote for whichever party you want to. That's the beauty of the

1444
01:24:02,060 --> 01:24:07,000
European governing system is that we are all represented, right? But the important thing here

1445
01:24:07,160 --> 01:24:11,800
is you should also know if you want your privacy rights to be respected. And if it is your

1446
01:24:12,040 --> 01:24:16,340
representatives who tried to push this, then I think you should know, be it left or right,

1447
01:24:16,720 --> 01:24:20,920
who pushed for it. And there are amazing websites where you can go check out who pushed for it,

1448
01:24:20,960 --> 01:24:22,880
and you can log in and take a look at it

1449
01:24:22,940 --> 01:24:24,880
and you'll find out that this was mostly the case

1450
01:24:25,140 --> 01:24:27,020
in most countries, at least in the Nordics.

1451
01:24:27,400 --> 01:24:28,600
And it's very easy to see.

1452
01:24:29,200 --> 01:24:30,900
Yeah, and I liked also Patrick.

1453
01:24:31,180 --> 01:24:33,860
He put out his blog and then shortly after he said,

1454
01:24:33,940 --> 01:24:36,220
now we can actually focus on keeping children safe.

1455
01:24:36,400 --> 01:24:38,380
I think his overall point he was trying to make

1456
01:24:38,660 --> 01:24:40,560
was so much of the air in the room

1457
01:24:40,760 --> 01:24:42,400
was taken up by chat control.

1458
01:24:42,980 --> 01:24:44,360
And a lot of the technologies

1459
01:24:44,520 --> 01:24:45,540
that they were trying to use for chat control

1460
01:24:45,640 --> 01:24:46,440
weren't even evidence-based.

1461
01:24:46,740 --> 01:24:49,640
Like 50% of the times, or even more than 50% of the times,

1462
01:24:49,740 --> 01:24:54,400
they were just positive yeah i mean your client's eye scanning it just doesn't fucking work like so

1463
01:24:54,580 --> 01:25:01,520
yeah so it's just it's cool how now he's saying like you know now we actually have that that air

1464
01:25:01,860 --> 01:25:06,400
we can all kind of reflect on this problem and so maybe it's good timing as well for this podcast

1465
01:25:06,800 --> 01:25:11,480
for us to kind of think about this i mean i think so too if if you're a regulator if you're one of

1466
01:25:11,480 --> 01:25:15,140
those political parties and if you want to have this conversation please reach out to us i'd be

1467
01:25:15,280 --> 01:25:18,901
more than happy to sit down and talk about the reality of encryption and how that keeps you safe

1468
01:25:18,920 --> 01:25:21,000
and everyone's safe and your constituents safe

1469
01:25:21,360 --> 01:25:24,820
and how revoking encryption is not going to help you

1470
01:25:24,840 --> 01:25:26,800
or your constituents or kids for that matter.

1471
01:25:26,980 --> 01:25:28,360
People are still going to keep doing what they're doing.

1472
01:25:28,780 --> 01:25:30,260
They're still going to find ways to encrypt this stuff

1473
01:25:30,340 --> 01:25:31,020
if they want to do that.

1474
01:25:32,460 --> 01:25:32,860
Yeah.

1475
01:25:34,300 --> 01:25:36,620
Because when I first started getting to know you,

1476
01:25:37,100 --> 01:25:39,160
you weren't really active in the legal side of things

1477
01:25:39,440 --> 01:25:40,320
as far as I know back then.

1478
01:25:42,280 --> 01:25:44,040
So I'd say the last few years,

1479
01:25:44,880 --> 01:25:46,480
you've been getting a lot more active

1480
01:25:46,940 --> 01:25:48,000
on the political side of things.

1481
01:25:48,100 --> 01:25:52,200
I know we have casual conversations all the time,

1482
01:25:52,280 --> 01:25:54,040
but I've never had to directly ask,

1483
01:25:56,660 --> 01:25:58,420
did it just happen organically?

1484
01:25:58,620 --> 01:25:59,640
Did you start getting into this?

1485
01:26:00,020 --> 01:26:01,560
What's the value you see in that?

1486
01:26:02,180 --> 01:26:05,480
Can you shed a little bit of light of what it looks like to do that?

1487
01:26:05,640 --> 01:26:06,380
Yeah, absolutely.

1488
01:26:08,040 --> 01:26:09,260
I'm not even saying what you do,

1489
01:26:09,520 --> 01:26:11,360
so you should probably expand on what you've done.

1490
01:26:12,580 --> 01:26:14,280
I don't love it. I'll start with that.

1491
01:26:14,540 --> 01:26:24,040
I've never seen myself as someone who would ever wear a suit and sit in a room, and yet I found myself, caught myself doing that in a way that I'm like, fuck, what am I doing?

1492
01:26:24,220 --> 01:26:25,160
Why am I wearing a blazer?

1493
01:26:25,260 --> 01:26:26,120
Why am I wearing a suit?

1494
01:26:26,360 --> 01:26:27,760
And I'm just a nerd, guys.

1495
01:26:28,640 --> 01:26:31,160
I'm a coder, and I come from a technical background.

1496
01:26:31,520 --> 01:26:32,460
This was all new to me.

1497
01:26:33,340 --> 01:26:41,060
And I think what started all this was two years ago when the whole Digital Markets Act-related issue started to happen.

1498
01:26:41,240 --> 01:26:42,760
And you and I had this video, actually.

1499
01:26:43,020 --> 01:26:46,240
I think I can attribute the day that I got dragged into all this shit.

1500
01:26:46,540 --> 01:26:50,560
Pretty much that video that you and I made about Apple trying to block web apps in Europe

1501
01:26:51,140 --> 01:26:52,020
under some false premises.

1502
01:26:52,640 --> 01:26:56,080
Little backstories, Apple said, hey, we're going to kill all web apps in Europe.

1503
01:26:56,300 --> 01:26:58,460
Basically, Apple said, we're going to kill web apps in the EU.

1504
01:26:59,360 --> 01:27:01,320
And they tried to use...

1505
01:27:01,320 --> 01:27:03,800
They didn't even say, they just pushed out an update that killed it.

1506
01:27:03,920 --> 01:27:05,200
Yeah, they just said that they just killed it.

1507
01:27:06,120 --> 01:27:08,880
They pushed down beta and they were going to kill the companies

1508
01:27:09,100 --> 01:27:11,080
and they were like, hey, you know what, we're just going to do this.

1509
01:27:11,180 --> 01:27:18,820
And the false excuse they used for that was that they were saying, hey, in order to comply with the Digital Markets Act, we have to do this.

1510
01:27:19,260 --> 01:27:21,340
And I was like, what kind of a bullshit act is that?

1511
01:27:21,360 --> 01:27:22,320
Let me go read that shit.

1512
01:27:22,480 --> 01:27:23,680
They're going to kill my company, you know.

1513
01:27:24,800 --> 01:27:26,940
And I had to sit down and read.

1514
01:27:26,940 --> 01:27:30,980
And I was like, nothing in the Digital Markets Act says that they need to do that.

1515
01:27:31,060 --> 01:27:31,880
What the fuck is that about?

1516
01:27:32,360 --> 01:27:41,120
And then I started reaching out to a bunch of organizations and by some amazing miracle, I connected with two incredible dudes from Australia who started Open Web Adversions.

1517
01:27:41,120 --> 01:27:47,920
who were trying to advocate for the open web and for essentially to get more browsers and browser engines in iOS.

1518
01:27:48,580 --> 01:27:54,320
That's apparently what triggered Apple to say, hey, if you want to have other browser engines on iOS, then we're going to kill web apps.

1519
01:27:54,450 --> 01:27:59,660
Even though that is a complete bullshit argument and there was no security benefit, no nothing.

1520
01:28:00,390 --> 01:28:06,240
And open web advocacy and myself involved, we published an open letter to Tim Cook saying, don't do this.

1521
01:28:06,560 --> 01:28:15,760
And within 48 hours, we got 5,000 signatures from every single big tech organization you can imagine, to every single engineer from every company you can imagine, small and big.

1522
01:28:16,160 --> 01:28:17,100
Everyone said no.

1523
01:28:17,500 --> 01:28:21,500
And Apple immediately backed down and said, maybe we're not going to do that.

1524
01:28:21,860 --> 01:28:25,360
But only Safari should be able to install web apps and stuff like this.

1525
01:28:26,620 --> 01:28:26,840
Whatever.

1526
01:28:27,100 --> 01:28:32,880
That was my first unexpected foray into the legal side of things with Digital Markets Act.

1527
01:28:33,100 --> 01:28:42,240
And that sort of got me directly, that affected me directly and got me involved directly because I have a small company that was going to be very directly affected by this.

1528
01:28:42,720 --> 01:28:47,860
And effectively, that act would have made or break my company, so to speak, in a very specific manner.

1529
01:28:48,680 --> 01:28:55,520
And I reached out via the Open Web Advocacy to the European Commission to say like, hey, like, what's going on?

1530
01:28:55,700 --> 01:28:56,640
Can we have this conversation?

1531
01:28:57,000 --> 01:28:57,900
Can we talk a little bit?

1532
01:28:57,960 --> 01:29:01,000
And they were incredibly kind and incredibly amazing.

1533
01:29:01,260 --> 01:29:08,700
And they've actually invited open web advocacy in every capacity that they can to engage with us at every capacity that they can.

1534
01:29:08,890 --> 01:29:14,720
Almost weekly, we had calls with the European Commission trying to understand what we can do, how we can fix things, how we can remedy things.

1535
01:29:15,000 --> 01:29:16,400
What would be a good compromise?

1536
01:29:16,700 --> 01:29:17,680
What would be a good solution?

1537
01:29:17,900 --> 01:29:19,820
What are some things we're happy about and not happy about?

1538
01:29:19,850 --> 01:29:23,900
And we had years of these conversations almost every few weeks with the European Commission.

1539
01:29:24,140 --> 01:29:28,100
And also with the UK competition market authorities and with the Japanese ones.

1540
01:29:28,920 --> 01:29:38,940
And it slowly became a thing where at first I started helping out the folks at Open Web Advocacy because I was directly impacted and my livelihood was going to be directly impacted by it.

1541
01:29:39,160 --> 01:29:42,700
And in the process of doing so, I ended up making a lot of trips to Brussels.

1542
01:29:43,520 --> 01:29:57,300
And over the course of the years now, I've been there enough times that I got to meet all these incredible people from these amazing organizations like Lucas from Free Software Foundation or Jan Penfrott from EDRI or Bruce from Vivaldi.

1543
01:29:57,340 --> 01:30:00,620
and Kush representing Mozilla, their policy person.

1544
01:30:01,020 --> 01:30:02,880
There's been incredible people that we got to meet

1545
01:30:02,930 --> 01:30:03,920
over the course of these years

1546
01:30:04,420 --> 01:30:07,240
who have been working their butts off every fucking day

1547
01:30:07,360 --> 01:30:08,980
trying to do the right thing for our rights.

1548
01:30:09,520 --> 01:30:11,300
And it's just incredible.

1549
01:30:11,350 --> 01:30:15,340
The atmosphere is just really uplifting and energizing

1550
01:30:15,520 --> 01:30:16,680
because you walk into that room

1551
01:30:16,780 --> 01:30:18,740
and you realize you're surrounded by all these other people

1552
01:30:18,750 --> 01:30:20,080
who are trying to protect your rights.

1553
01:30:20,660 --> 01:30:22,460
And everyone's got your back.

1554
01:30:22,740 --> 01:30:25,840
And that sort of got me, that dragged me into this a little bit.

1555
01:30:26,120 --> 01:30:32,680
And I've realized over the course of the years in doing this, I don't think encryption alone is going to keep us safe.

1556
01:30:33,520 --> 01:30:36,460
I can say as much that it's not a magic bullet.

1557
01:30:37,080 --> 01:30:39,840
I mean, it's a dual use technology, but it is not.

1558
01:30:40,080 --> 01:30:49,060
And it's a controlled export good, but it's still not going to protect us in that you can have the best, best piece of mathematical formulas and encryption that you want.

1559
01:30:49,220 --> 01:30:53,840
If someone says you're not legally allowed to use it, good fucking luck.

1560
01:30:53,980 --> 01:30:55,340
No company is going to be able to offer it.

1561
01:30:55,560 --> 01:30:59,880
And maybe you can try to do it open source and GitHub will take it down or someone else will take your website down.

1562
01:30:59,900 --> 01:31:03,980
I mean, if someone says this is not cool legally, you're not going to be able to do this.

1563
01:31:04,280 --> 01:31:11,080
And I realized, OK, so encryption is a big piece of the puzzle, but not the only piece of the puzzle that we need to focus on.

1564
01:31:11,360 --> 01:31:21,500
And I realized, OK, like web is another big piece of this puzzle, because if we lose the web, if we lose the openness and interoperability of the web, we're also fucked.

1565
01:31:21,820 --> 01:31:27,340
Because if we depend only on the app stores, you've seen what happened to Ice Report apps in the United States.

1566
01:31:27,580 --> 01:31:30,360
You've seen what happened to, I mean, it's super weird, right?

1567
01:31:30,470 --> 01:31:31,200
In the United States.

1568
01:31:31,220 --> 01:31:32,060
VPNs in Russia, China.

1569
01:31:32,100 --> 01:31:34,040
VPNs in Russia and China, exactly.

1570
01:31:34,750 --> 01:31:40,920
And the irony of the situation is in the United States, if you ask Apple, they would say, hey, we want to protect the kids.

1571
01:31:41,660 --> 01:31:47,140
And if you look at their app store guidelines, they will say that they will take down anything that harms people.

1572
01:31:47,660 --> 01:31:52,660
Yet, they've continued hosting Rock, which literally is used for creating CSAM.

1573
01:31:52,800 --> 01:31:53,720
And they didn't take it down.

1574
01:31:53,880 --> 01:31:55,460
There is an amazing Verge article.

1575
01:31:55,640 --> 01:31:56,080
I know.

1576
01:31:57,480 --> 01:32:01,300
Oh, I thought you were going to reference what the Alt Store guys posted on Mastodon.

1577
01:32:01,400 --> 01:32:01,520
Exactly.

1578
01:32:01,800 --> 01:32:02,840
I was getting to that.

1579
01:32:03,540 --> 01:32:04,100
Yeah, yeah, yeah.

1580
01:32:04,220 --> 01:32:04,940
You can explain it.

1581
01:32:05,560 --> 01:32:05,660
Yeah.

1582
01:32:06,480 --> 01:32:10,560
Over the course of going to Brussels, I met these incredible folks, Riley and Shane, from the Alt Store.

1583
01:32:10,700 --> 01:32:14,940
And they've created this beautiful piece of alternative app store for iOS devices.

1584
01:32:15,540 --> 01:32:19,080
And they made an amazing app that only shows you a yes or no, basically.

1585
01:32:19,440 --> 01:32:25,980
And it says, which of the app stores have CSAM on it and allows apps that creation of CSAM.

1586
01:32:26,180 --> 01:32:29,180
And believe it or not, it is the Apple's app store and not the Alt Store.

1587
01:32:29,820 --> 01:32:35,340
Yet, those were exactly the kinds of shit that they've accused Alt Store of distributing potentially.

1588
01:32:35,520 --> 01:32:43,200
And that's why they wanted to have some notarization process so they can take a look at which apps are being distributed through the Alt Store, yada, yada, yada.

1589
01:32:43,580 --> 01:32:45,400
And the irony here is, it's so weird.

1590
01:32:45,500 --> 01:32:49,740
right? Apple says, hey, we want to protect minorities. We want to protect kids. That's why

1591
01:32:49,740 --> 01:32:55,580
we have all these policies. And they don't take down Grok, which is used for creation and distribution

1592
01:32:55,740 --> 01:33:01,320
of CSAM. But then they declared people who work for ICE as minorities who should be protected.

1593
01:33:01,760 --> 01:33:07,400
And then they took down ICE reporting apps. The irony here is these are grown up men with fucking

1594
01:33:07,620 --> 01:33:11,361
guns. I don't think they need your digital protection. I think they can protect themselves

1595
01:33:11,380 --> 01:33:14,180
versus kids who are being abused, on the other hand,

1596
01:33:14,700 --> 01:33:16,600
whose sexual abuse is being distributed

1597
01:33:16,760 --> 01:33:17,980
on your goddamn app store

1598
01:33:18,550 --> 01:33:19,860
cannot protect themselves with guns,

1599
01:33:20,100 --> 01:33:22,180
so maybe you should take down Grok, but they don't.

1600
01:33:22,380 --> 01:33:25,200
And I think the open web is very important for this reason

1601
01:33:25,480 --> 01:33:26,960
because we don't depend on these app stores

1602
01:33:27,200 --> 01:33:28,580
and what they decide is okay

1603
01:33:28,740 --> 01:33:29,900
and what they decide isn't okay.

1604
01:33:30,320 --> 01:33:33,560
And I don't use the word decentralization lightly here

1605
01:33:33,720 --> 01:33:36,200
because I think web isn't exactly decentralized either

1606
01:33:36,460 --> 01:33:38,140
depending on where you host your website, right?

1607
01:33:38,520 --> 01:33:43,220
But web is still not necessarily part of these big corporations.

1608
01:33:43,720 --> 01:33:49,140
Like you can choose to host your website somewhere else where that corporation cannot decide who gets to have a website.

1609
01:33:49,760 --> 01:33:51,180
So I think web is a big component.

1610
01:33:51,560 --> 01:33:54,200
That's why I think it hit the right nerve for me.

1611
01:33:54,220 --> 01:33:59,840
And I said I need to get involved and do something to protect the future of the web so that it's not just these fucking app stores.

1612
01:34:00,360 --> 01:34:01,800
I don't want Apple or Google killing the web.

1613
01:34:02,080 --> 01:34:03,200
encryption is a big portion of it.

1614
01:34:03,440 --> 01:34:04,760
I want to make sure that we protect it.

1615
01:34:04,800 --> 01:34:09,540
And I think in order to be able to do all those, you kind of need to have these legal conversations.

1616
01:34:09,920 --> 01:34:14,360
And that's sort of how I became more involved into all this legal situation.

1617
01:34:14,920 --> 01:34:17,180
How I got dragged into this legal situation, let's say.

1618
01:34:17,200 --> 01:34:26,720
And I realized I can hide behind my keyboard and just code or go there and try to do some change as much as I can by talking to people and show my face.

1619
01:34:26,980 --> 01:34:29,120
And I had to get a good haircut for that.

1620
01:34:30,360 --> 01:34:32,120
And the old one wasn't doing it.

1621
01:34:35,360 --> 01:34:37,880
and now I have to put blazers and suits on

1622
01:34:39,020 --> 01:34:41,460
and I have to sweat for seven, eight hours

1623
01:34:42,040 --> 01:34:43,180
but that's okay, it's worth it

1624
01:34:44,900 --> 01:34:49,380
I think it's a pretty frequent realization for some people

1625
01:34:49,540 --> 01:34:51,560
a person who just comes to mind right now

1626
01:34:51,760 --> 01:34:53,140
just on the fly is Louis Rossman, right?

1627
01:34:53,280 --> 01:34:54,200
Louis, you know Louis?

1628
01:34:56,740 --> 01:34:57,940
Yeah, yeah, not me neither

1629
01:34:58,260 --> 01:35:01,441
but he started just kind of making these tutorials

1630
01:35:01,440 --> 01:35:03,800
about something he knew a lot about on YouTube,

1631
01:35:04,480 --> 01:35:05,600
which is just repairing masks.

1632
01:35:06,200 --> 01:35:07,580
I watched Louis Rossman

1633
01:35:07,880 --> 01:35:09,740
back when he had a very early channel.

1634
01:35:09,800 --> 01:35:12,040
Back when he was smaller than we are right now

1635
01:35:12,140 --> 01:35:13,240
because we just hit 300K.

1636
01:35:13,820 --> 01:35:15,080
So I think I remember watching him

1637
01:35:15,100 --> 01:35:16,740
when he was at 200K.

1638
01:35:16,880 --> 01:35:17,500
Thank you, thank you.

1639
01:35:17,640 --> 01:35:19,200
Yeah, we were actually in person for it,

1640
01:35:19,520 --> 01:35:20,040
recording a project,

1641
01:35:20,400 --> 01:35:21,300
and then we hit 300K,

1642
01:35:21,420 --> 01:35:23,440
so we made a little cake.

1643
01:35:23,820 --> 01:35:24,140
Nice!

1644
01:35:24,820 --> 01:35:26,760
Did you see the photo on Mastodon by any chance?

1645
01:35:27,400 --> 01:35:28,140
No, I should look.

1646
01:35:28,520 --> 01:35:29,340
Okay, yeah, check it out.

1647
01:35:29,460 --> 01:35:30,320
We made a little cake.

1648
01:35:30,460 --> 01:35:33,820
We wrote 300K on it, and it was a great, great cake.

1649
01:35:34,080 --> 01:35:34,740
Well-deserved.

1650
01:35:35,060 --> 01:35:35,500
Amazing content.

1651
01:35:35,740 --> 01:35:36,660
Yes, thank you, thank you.

1652
01:35:36,780 --> 01:35:37,980
But back then, it was just tutorials.

1653
01:35:38,180 --> 01:35:41,780
I was just fascinated, because he would show how to repair this one component on a Mac.

1654
01:35:41,910 --> 01:35:46,780
He would explain the specific things that Apple didn't want to acknowledge as real problems.

1655
01:35:47,000 --> 01:35:54,820
He's like, but I've had this repair at my repair shop in New York come up on literally every single person when they hit year seven on this specific MacBook model.

1656
01:35:55,460 --> 01:35:57,540
And so he was kind of sharing this detail.

1657
01:35:57,900 --> 01:36:01,800
But he never talked about right to repair in the way that he might do now.

1658
01:36:02,110 --> 01:36:06,700
So, you know, he went from just making these kind of like interesting videos to then talking.

1659
01:36:06,770 --> 01:36:10,820
And he started doing more like talk show style videos where he started, you know, talking.

1660
01:36:10,960 --> 01:36:13,160
He's like, here's why you need to get active in this bill.

1661
01:36:13,340 --> 01:36:14,320
I'm so tired of Apple.

1662
01:36:14,530 --> 01:36:17,060
It was becoming a little bit more on the advocacy side of things.

1663
01:36:17,620 --> 01:36:27,700
But now, like, I don't know how much of air quotes right to repair was, you know, started by or even advanced by Louis Rossman single handedly.

1664
01:36:27,800 --> 01:36:29,140
I'd say 90%.

1665
01:36:29,760 --> 01:36:30,880
I'd say a huge amount of it was.

1666
01:36:32,380 --> 01:36:35,760
He was the person who exposed me to the concept of right to repair.

1667
01:36:35,920 --> 01:36:38,440
He was the one who started introducing me to these bills.

1668
01:36:39,080 --> 01:36:42,080
And then his audience starts following this as well.

1669
01:36:42,240 --> 01:36:47,480
I like to think that what we've been doing the last year at Techlore is quite similar.

1670
01:36:48,040 --> 01:36:50,180
We started with a lot of privacy tools.

1671
01:36:50,360 --> 01:36:52,100
Like, oh, this is more private than this.

1672
01:36:52,120 --> 01:36:55,040
But it's like, oh wait, they're actually just trying to ban encrypted tools.

1673
01:36:55,880 --> 01:36:57,880
So we can still have those discussions.

1674
01:36:58,240 --> 01:36:58,960
That's quite fascinating.

1675
01:36:59,240 --> 01:37:00,000
I think it's fascinating.

1676
01:37:00,080 --> 01:37:01,160
I think it's really fun to invite.

1677
01:37:01,380 --> 01:37:02,620
I mean, I've had Session on this podcast.

1678
01:37:02,800 --> 01:37:03,900
I've had SimpleX on this podcast.

1679
01:37:04,240 --> 01:37:05,660
They both think the other one sucks.

1680
01:37:05,980 --> 01:37:08,200
It would be really interesting to get a roundtable discussion

1681
01:37:08,420 --> 01:37:10,460
between all of them and see what they think.

1682
01:37:10,520 --> 01:37:12,660
I think it would be a well-moderated discussion.

1683
01:37:12,840 --> 01:37:14,020
And they're both really nice guys,

1684
01:37:14,100 --> 01:37:15,660
so I know that they would be very professional.

1685
01:37:15,840 --> 01:37:16,720
I'll just say this already.

1686
01:37:17,080 --> 01:37:17,620
Proton's awesome.

1687
01:37:22,640 --> 01:37:23,320
It would be good.

1688
01:37:23,520 --> 01:37:25,460
I need to get Proton on the podcast, actually.

1689
01:37:25,540 --> 01:37:27,500
But I'm like, this doesn't matter.

1690
01:37:27,840 --> 01:37:30,040
It does matter because I find it technically interesting.

1691
01:37:30,210 --> 01:37:32,600
And I'm sure that still deserves a place.

1692
01:37:32,970 --> 01:37:34,580
But I was putting so much emphasis on that

1693
01:37:34,670 --> 01:37:37,060
when really we were missing the whole picture of like,

1694
01:37:37,060 --> 01:37:38,660
well, these tools are only allowed to exist.

1695
01:37:39,060 --> 01:37:42,540
And people already talk about how hard it is to just move to Signal.

1696
01:37:43,140 --> 01:37:46,300
Signal is, feature parity-wise, quite similar to WhatsApp.

1697
01:37:46,660 --> 01:37:47,920
You have stories on Signal.

1698
01:37:48,840 --> 01:37:49,860
Yeah, I would argue better too.

1699
01:37:50,220 --> 01:37:53,700
And now we're talking about getting people over to Session or SimpleX,

1700
01:37:53,880 --> 01:37:55,760
which the usability is far worse than Signal.

1701
01:37:55,840 --> 01:37:56,780
I have no issue saying that.

1702
01:37:57,040 --> 01:37:57,960
So that's already kind of an issue.

1703
01:37:58,100 --> 01:38:00,140
But then, okay, let's set up a situation

1704
01:38:00,560 --> 01:38:01,860
where encryption is just banned.

1705
01:38:02,080 --> 01:38:04,100
So now you have to convince your friends and family

1706
01:38:04,240 --> 01:38:05,260
not just to move to Signal,

1707
01:38:05,360 --> 01:38:07,200
which apparently is hard enough to do for a lot of people.

1708
01:38:07,280 --> 01:38:08,720
And it is, I have this problem as well.

1709
01:38:08,900 --> 01:38:10,680
You now have to convince them to use an illegal tool

1710
01:38:11,680 --> 01:38:12,340
along the way.

1711
01:38:12,700 --> 01:38:14,280
You really think that's a healthy environment?

1712
01:38:15,120 --> 01:38:16,560
And maybe compile it themselves

1713
01:38:17,000 --> 01:38:18,240
and not get it from the app store,

1714
01:38:18,520 --> 01:38:20,100
load it as they call it,

1715
01:38:20,100 --> 01:38:21,060
and find ways to install it.

1716
01:38:21,060 --> 01:38:22,040
What do you mean, you idiot?

1717
01:38:22,260 --> 01:38:26,320
You didn't build your own signal source code for Linux Arm,

1718
01:38:26,540 --> 01:38:27,740
which they still haven't done, by the way.

1719
01:38:27,820 --> 01:38:28,860
I'm very frustrated about that.

1720
01:38:29,220 --> 01:38:30,780
I am extremely frustrated as well.

1721
01:38:30,820 --> 01:38:35,160
I don't know if you've seen, but I switched all Chromebooks with great pleasure.

1722
01:38:35,420 --> 01:38:40,920
With Android apps running in Linux apps on there as well?

1723
01:38:40,920 --> 01:38:43,460
Yeah, you can run Linux apps, you can run progressive web apps,

1724
01:38:43,520 --> 01:38:44,380
you can run Android apps.

1725
01:38:44,560 --> 01:38:46,880
It's actually, if you think about it, the most accepting platform.

1726
01:38:47,380 --> 01:38:47,980
Everything runs on it.

1727
01:38:48,280 --> 01:38:49,080
In some ways, yeah.

1728
01:38:49,360 --> 01:38:51,040
That's interesting, except you have to deal with Linux Arm.

1729
01:38:52,120 --> 01:38:54,240
So far it hasn't been that bad.

1730
01:38:54,480 --> 01:38:56,660
Even our development tools are all working great.

1731
01:38:57,460 --> 01:39:00,560
I can safely say that we shipped it with Chrome OS

1732
01:39:00,660 --> 01:39:02,860
and I'm really happy about it as a web app developer

1733
01:39:02,920 --> 01:39:03,780
who makes web apps.

1734
01:39:04,600 --> 01:39:08,380
And I don't have to use Apple products for stuff anymore.

1735
01:39:08,820 --> 01:39:09,360
It's really good.

1736
01:39:09,600 --> 01:39:12,140
After having shit-talked them so many times in these hearings,

1737
01:39:12,380 --> 01:39:15,320
I feel like it's going to be really nice to flash my Lenovo.

1738
01:39:15,460 --> 01:39:18,740
I saw one of the new Microsoft laptops.

1739
01:39:19,680 --> 01:39:20,960
I don't know their lineup confuses.

1740
01:39:21,020 --> 01:39:22,820
because I still associate the Surface

1741
01:39:23,020 --> 01:39:24,680
as the crappy little piece of crap

1742
01:39:24,720 --> 01:39:25,700
they released the first time.

1743
01:39:26,260 --> 01:39:26,960
But it's grown a lot.

1744
01:39:27,120 --> 01:39:28,580
It's genuinely really nice hardware.

1745
01:39:28,660 --> 01:39:29,860
I think it's the Surface laptop.

1746
01:39:30,620 --> 01:39:32,700
It is kind of like MacBook Pro build quality.

1747
01:39:32,900 --> 01:39:34,480
It runs an ARM chip, the Snapdragon X.

1748
01:39:34,520 --> 01:39:36,960
I don't think it's as good, but it's still okay.

1749
01:39:37,340 --> 01:39:38,720
So the hardware is getting there,

1750
01:39:38,980 --> 01:39:40,460
but dude, when the hardware's out,

1751
01:39:40,920 --> 01:39:41,680
Linux still isn't there.

1752
01:39:42,860 --> 01:39:43,700
This is very off topic.

1753
01:39:44,320 --> 01:39:46,060
So let's get back on track.

1754
01:39:46,060 --> 01:39:48,940
On topic in that I think all this was made possible

1755
01:39:49,100 --> 01:39:50,800
by incredible folks like Louis Russman,

1756
01:39:50,840 --> 01:39:54,800
initially taking their time to engage with the legal process

1757
01:39:55,200 --> 01:39:57,680
and making it so that the right to repair movement

1758
01:39:57,880 --> 01:39:59,820
catch the right steam and caught the right momentum

1759
01:39:59,980 --> 01:40:01,080
and basically said, hey, you know what,

1760
01:40:01,170 --> 01:40:02,300
maybe we should do something about this.

1761
01:40:02,400 --> 01:40:05,120
And then we started getting computer manufacturers

1762
01:40:05,340 --> 01:40:07,760
like Framework and System 74 or 76.

1763
01:40:07,850 --> 01:40:09,440
I forget their name because I'm bad with numbers.

1764
01:40:09,820 --> 01:40:14,000
Yeah, honestly, it's all made possible by folks like this,

1765
01:40:14,840 --> 01:40:17,240
pushing the limits of what's possible with regulation.

1766
01:40:17,450 --> 01:40:19,880
And I realize just how important that is.

1767
01:40:20,080 --> 01:40:25,840
I'm trying my best in the best way I can, spending our funds as much as I can.

1768
01:40:25,870 --> 01:40:29,660
So if you're spending money on Krypton, you have a subscription, all of it is going to this shit.

1769
01:40:31,460 --> 01:40:42,540
And essentially, we're trying to push the boundaries of what we can do with our legal team, our tiny, small engineering team, and me making trips to Brussels, trying to talk to the regulators to defend your rights.

1770
01:40:43,380 --> 01:40:45,940
And I can tell you the same about Vivaldi.

1771
01:40:45,940 --> 01:40:51,540
If you use a Vivaldi browser, they're trying their very fucking best to do this in the same exact way.

1772
01:40:51,880 --> 01:40:53,840
And this is really important, is what I'm trying to say.

1773
01:40:53,980 --> 01:40:55,400
And Signal's doing the same.

1774
01:40:55,660 --> 01:41:02,840
Meredith is flying out every now and then trying to represent Signal and encryption and our conversations, privacy as a whole.

1775
01:41:03,480 --> 01:41:07,660
I have immense levels of respect to these people for doing what they do,

1776
01:41:07,760 --> 01:41:12,500
because I know firsthand just how difficult it is to actually do and live out of luggages in hotels.

1777
01:41:12,760 --> 01:41:18,300
And aside from the inconveniences, if you have to keep doing it and grabbing drinks with these people, it gets really tiring.

1778
01:41:18,600 --> 01:41:23,260
You know, at our age, it's not that easy to keep getting tipsy and talk about privacy all day long.

1779
01:41:26,180 --> 01:41:30,580
So not that I don't dislike it, but it's hard to convince the right people.

1780
01:41:31,300 --> 01:41:33,960
Yeah. Well, that's all I have.

1781
01:41:33,960 --> 01:41:35,700
This is a very no-filter podcast.

1782
01:41:36,220 --> 01:41:39,400
Yeah, but it was a little more filtered than the last one, I think.

1783
01:41:39,700 --> 01:41:42,300
The last one was quite all over the place.

1784
01:41:42,980 --> 01:41:44,020
unhinged. That's a good way to put it.

1785
01:41:44,200 --> 01:41:46,580
Great. Well, I'd appreciate anybody's thoughts in the comments.

1786
01:41:47,120 --> 01:41:48,660
Obviously, keep in mind

1787
01:41:48,820 --> 01:41:50,320
the sensitive nature of what we're discussing.

1788
01:41:50,680 --> 01:41:52,700
Even if you firmly disagree and you're

1789
01:41:52,820 --> 01:41:54,540
like, I think this child stuff is nonsense

1790
01:41:55,100 --> 01:41:56,500
and I think encryption is

1791
01:41:56,960 --> 01:41:58,440
be respectful about it because

1792
01:41:58,900 --> 01:42:00,580
keep in mind the context of how sensitive

1793
01:42:00,840 --> 01:42:02,780
this conversation is. People in the comments

1794
01:42:02,980 --> 01:42:04,640
might have people personally affected by this.

1795
01:42:04,820 --> 01:42:06,840
I tried my best to treat this as sensitively

1796
01:42:06,840 --> 01:42:08,740
as possible. John clearly did as well.

1797
01:42:09,020 --> 01:42:10,841
So I just ask a little bit of compassion

1798
01:42:10,840 --> 01:42:11,600
and kindness as well.

1799
01:42:13,840 --> 01:42:16,600
To this day, our intro video on our channel

1800
01:42:17,020 --> 01:42:18,800
doesn't have comments anymore because it's an intro video.

1801
01:42:19,540 --> 01:42:22,180
I want to keep the comments open to still have people discuss this

1802
01:42:22,280 --> 01:42:23,460
in a productive way.

1803
01:42:24,280 --> 01:42:24,760
That's all I ask.

1804
01:42:24,960 --> 01:42:27,840
I think the most important thing we can do right now

1805
01:42:28,200 --> 01:42:31,680
is to engage in civil conversations about this topic

1806
01:42:32,200 --> 01:42:35,640
because just ignoring what all these people

1807
01:42:35,720 --> 01:42:38,380
trying to protect kids are trying to say is not going to work.

1808
01:42:38,580 --> 01:42:41,060
We can't just close our ears and say, I can't hear you.

1809
01:42:41,380 --> 01:42:43,340
I know you're trying to protect kids, but I can't hear you.

1810
01:42:43,520 --> 01:42:44,860
It's just not going to work out.

1811
01:42:45,120 --> 01:42:47,260
We have to have very civil conversations around this.

1812
01:42:47,540 --> 01:42:49,220
And we can't ignore what they're saying.

1813
01:42:49,940 --> 01:42:52,180
And we're not going to make compromises on encryption.

1814
01:42:52,600 --> 01:42:53,320
I promise you that.

1815
01:42:53,500 --> 01:42:54,320
No one is going to.

1816
01:42:54,420 --> 01:42:59,620
None of the privacy-respecting, security-respecting companies that you hold dear to your heart,

1817
01:42:59,760 --> 01:43:02,700
like Proton or Tuta or Signal, we're not going to make compromise.

1818
01:43:03,300 --> 01:43:05,160
But we also have to have this conversation.

1819
01:43:05,540 --> 01:43:06,940
And it's going to be a difficult conversation.

1820
01:43:07,320 --> 01:43:12,760
And I hope that you would be as civil and as open-hearted and open-minded about it,

1821
01:43:12,850 --> 01:43:16,380
because there will be people in the comments, statistically, there will be people in the

1822
01:43:16,600 --> 01:43:20,880
comments section who are going to be personally affected, who were going to be personally affected,

1823
01:43:21,440 --> 01:43:23,120
who may be actively affected.

1824
01:43:23,590 --> 01:43:25,800
And I sincerely hope that you will respect them.

1825
01:43:26,160 --> 01:43:29,560
And I sincerely hope that you will understand that it's a very nuanced topic.

1826
01:43:29,700 --> 01:43:31,080
It's a very difficult one for them.

1827
01:43:31,520 --> 01:43:35,120
And these people who have these conversations in public

1828
01:43:35,440 --> 01:43:38,440
and talk about their incredibly dark traumas and experiences,

1829
01:43:38,720 --> 01:43:40,860
they make incredibly difficult sacrifices

1830
01:43:40,950 --> 01:43:43,780
to have these conversations and make these possible.

1831
01:43:44,070 --> 01:43:46,920
So I hope you can respect that as well.

1832
01:43:47,660 --> 01:43:47,860
Yeah.

1833
01:43:48,450 --> 01:43:49,620
Well, thank you, John, for your time.

1834
01:43:50,540 --> 01:43:52,160
I know that I told you beforehand

1835
01:43:52,230 --> 01:43:54,800
this would be 80, 90% informational.

1836
01:43:55,880 --> 01:43:56,420
But it's good.

1837
01:43:59,560 --> 01:44:00,980
But congrats on the release.

1838
01:44:01,260 --> 01:44:02,900
I know this is something you've been cooking for a long time.

1839
01:44:03,140 --> 01:44:03,720
Thank you so much.

1840
01:44:03,940 --> 01:44:06,260
Yeah, I guess if anyone wants to check that out,

1841
01:44:07,800 --> 01:44:08,920
I guess you have to be a paid user.

1842
01:44:10,060 --> 01:44:11,800
But there is a Crypteec.

1843
01:44:12,940 --> 01:44:14,200
I don't know if there's cards.

1844
01:44:14,540 --> 01:44:15,980
I think there are cards for external links.

1845
01:44:16,120 --> 01:44:17,220
But we'll leave it in the show notes.

1846
01:44:20,120 --> 01:44:20,300
Yeah.

1847
01:44:21,280 --> 01:44:21,760
Yeah, there you go.

1848
01:44:22,560 --> 01:44:24,360
It's funny, too, because what you see here

1849
01:44:24,600 --> 01:44:28,060
isn't the format even that is done in post in Riverside.

1850
01:44:28,090 --> 01:44:28,220
Oh, shit.

1851
01:44:28,230 --> 01:44:28,680
It's quite funny.

1852
01:44:29,820 --> 01:44:30,580
God damn it.

1853
01:44:30,800 --> 01:44:32,020
I've scratched all sorts of body parts.

1854
01:44:32,130 --> 01:44:33,220
We'll find out how that looks like.

1855
01:44:35,140 --> 01:44:35,820
Well, thank you.

1856
01:44:36,290 --> 01:44:38,080
And I'm sure we'll be back on at some point.

1857
01:44:38,460 --> 01:44:40,680
And until then, keep up the great work.

1858
01:44:41,560 --> 01:44:41,900
Sounds good.

1859
01:44:42,180 --> 01:44:42,780
Thank you so much, Henry.

1860
01:44:42,940 --> 01:44:43,700
Deeply appreciate you all.

1861
01:44:44,540 --> 01:44:44,600
Yeah.

1862
01:44:45,260 --> 01:44:46,540
And there you have it, everybody.

1863
01:44:46,840 --> 01:44:48,900
I hope that we all learned something today.

1864
01:44:49,180 --> 01:44:49,920
I know I did.

1865
01:44:50,060 --> 01:44:51,620
This is a very new territory for me.

1866
01:44:51,860 --> 01:44:54,020
And I didn't really have the full terminology

1867
01:44:54,440 --> 01:44:57,420
nor understanding of this topic going into the interview.

1868
01:44:57,630 --> 01:44:59,881
And it's clear that John has done a lot of his research

1869
01:44:59,880 --> 01:45:01,960
and has talked to a lot of very experienced people

1870
01:45:02,060 --> 01:45:03,800
to be able to share what he knows with us today.

1871
01:45:03,980 --> 01:45:04,840
If you enjoyed this interview,

1872
01:45:05,220 --> 01:45:06,880
definitely leave a rating down in the description.

1873
01:45:07,220 --> 01:45:08,460
I'd love to hear any comments you have

1874
01:45:08,600 --> 01:45:10,400
as long as they please are sensitive

1875
01:45:10,660 --> 01:45:12,280
given the nature of what we discussed today.

1876
01:45:12,460 --> 01:45:14,020
Thank you all for watching this Techlore Talks

1877
01:45:14,140 --> 01:45:15,360
and I'll see you next time.

1878
01:45:42,540 --> 01:45:45,340
*music*