1
00:00:00,119 --> 00:00:02,671
I am. I nap they them pronouns, and

2
00:00:02,671 --> 00:00:05,224
you are listening to firewalls don't stop dragons.

3
00:00:13,620 --> 00:00:16,257
Hello, everybody Welcome back to firewalls don't stop

4
00:00:16,417 --> 00:00:18,669
Dragons. I'm your host Kerry Parker. Today we

5
00:00:18,669 --> 00:00:20,108
have episode 384

6
00:00:20,108 --> 00:00:21,466
for 07/08/2024.

7
00:00:21,866 --> 00:00:23,544
And we've got a great interview for you,

8
00:00:23,704 --> 00:00:26,341
we'll be talking with Io nap shortly. Before

9
00:00:26,341 --> 00:00:27,540
we get into that, though, a couple quick

10
00:00:27,540 --> 00:00:27,700
things.

11
00:00:28,431 --> 00:00:30,420
First of all, I did confirm my Andy

12
00:00:30,420 --> 00:00:31,533
and in of interview, I said that might

13
00:00:31,533 --> 00:00:31,851
be a thing.

14
00:00:32,646 --> 00:00:35,350
They have recently made some interesting moves, including,

15
00:00:35,748 --> 00:00:37,852
new... Google Docs replacement

16
00:00:38,226 --> 00:00:40,450
and also changing kind of the way their

17
00:00:40,450 --> 00:00:43,230
company has structured to to to try to

18
00:00:43,230 --> 00:00:44,604
protect it actually from

19
00:00:45,154 --> 00:00:46,752
External influence. So we're gonna talk about that

20
00:00:46,752 --> 00:00:47,892
with Andy. Probably

21
00:00:48,351 --> 00:00:50,109
probably in August and you might not actually

22
00:00:50,109 --> 00:00:51,867
hear it till September. But it's coming. I

23
00:00:51,867 --> 00:00:54,438
promise that will be a great interview. Also,

24
00:00:54,677 --> 00:00:56,372
the book surge is happening.

25
00:00:57,306 --> 00:00:58,979
I've been talking about this for several weeks.

26
00:00:59,377 --> 00:01:00,812
I finally got cut my...

27
00:01:01,544 --> 00:01:02,980
My act together on this and figure out

28
00:01:02,980 --> 00:01:04,097
what I was going to do.

29
00:01:04,735 --> 00:01:06,012
I'm not gonna get into it all before

30
00:01:06,012 --> 00:01:07,368
the interview or I don't wanna take away

31
00:01:07,368 --> 00:01:09,931
from that. But Stay tuned, please. After the

32
00:01:09,931 --> 00:01:11,998
interview, and I will give you the full

33
00:01:11,998 --> 00:01:12,475
low down.

34
00:01:13,270 --> 00:01:15,735
If nothing else, go to FDSD dot me

35
00:01:15,735 --> 00:01:18,929
slash book surge, all 1 word. I changed

36
00:01:18,929 --> 00:01:20,757
the name from book blitz. I thought blitz

37
00:01:20,757 --> 00:01:23,459
sound a little bit like carpet bombing. Maybe.

38
00:01:23,936 --> 00:01:25,287
So anyway, I went with book surge.

39
00:01:26,336 --> 00:01:28,564
Fd s dot me slash book surge, where...

40
00:01:28,803 --> 00:01:30,076
Or, of course, you'll find the link of

41
00:01:30,076 --> 00:01:32,065
the show notes for the full details. But

42
00:01:32,145 --> 00:01:33,895
I'm fully planning to do this for 1

43
00:01:33,895 --> 00:01:34,793
week. So

44
00:01:35,342 --> 00:01:36,691
if you wanna help out with that, and

45
00:01:36,770 --> 00:01:37,326
I hope you do,

46
00:01:38,120 --> 00:01:39,707
you definitely need check that out right away.

47
00:01:40,342 --> 00:01:42,565
Alright. So let me preface the interview with

48
00:01:42,565 --> 00:01:43,438
with Irene a little bit.

49
00:01:44,327 --> 00:01:46,154
They used to work at Google in the

50
00:01:46,154 --> 00:01:47,823
privacy department at Google.

51
00:01:48,459 --> 00:01:50,525
And so I had to asked them a

52
00:01:50,525 --> 00:01:52,368
lot of questions that. In this interview is

53
00:01:52,368 --> 00:01:53,719
almost like a 2 interviews in 1.

54
00:01:54,435 --> 00:01:55,786
The first part of this interview, we talk

55
00:01:55,786 --> 00:01:57,455
a lot about what it's like to work

56
00:01:57,455 --> 00:01:58,886
in the privacy department at Google.

57
00:01:59,617 --> 00:02:02,482
Irene was there for 5 years in the

58
00:02:02,482 --> 00:02:04,711
upper echelon of the privacy department and got

59
00:02:04,711 --> 00:02:06,166
to see firsthand

60
00:02:07,434 --> 00:02:09,192
what really goes on there, and I had

61
00:02:09,192 --> 00:02:09,751
to ask,

62
00:02:10,390 --> 00:02:12,148
I had ask about that. And it turns

63
00:02:12,148 --> 00:02:14,166
out that there was a lot of turmoil

64
00:02:15,195 --> 00:02:17,575
in that area of the company at the

65
00:02:17,575 --> 00:02:19,875
time, a lot of strife with the employees

66
00:02:19,875 --> 00:02:21,565
and upper management, which

67
00:02:22,349 --> 00:02:25,385
affected Irene ability to actually properly

68
00:02:25,836 --> 00:02:28,610
advocate for Google's users. In other words, probably

69
00:02:28,610 --> 00:02:30,960
you. And Are eventually found that

70
00:02:31,340 --> 00:02:34,219
they believed to be more effective that they

71
00:02:34,219 --> 00:02:35,500
had to leave Google,

72
00:02:36,219 --> 00:02:37,990
which is what they did. And then the

73
00:02:37,990 --> 00:02:40,056
second part of the interview, we actually talk

74
00:02:40,056 --> 00:02:40,556
about

75
00:02:41,010 --> 00:02:44,052
privacy enhancing technologies or or Pet.

76
00:02:44,918 --> 00:02:47,144
There's some really interesting technology out there where

77
00:02:47,144 --> 00:02:49,926
a researcher basically trying to find ways to

78
00:02:49,926 --> 00:02:50,426
process

79
00:02:51,118 --> 00:02:53,764
a lot of potentially very personal data

80
00:02:54,155 --> 00:02:56,014
in a way that doesn't violate

81
00:02:56,314 --> 00:02:57,854
the privacy of the people

82
00:02:58,715 --> 00:03:01,194
whose data that is. And that may seem

83
00:03:01,194 --> 00:03:03,525
like an oxy moron. But there actually are

84
00:03:03,525 --> 00:03:05,778
some interesting ways to do this. And Irene

85
00:03:06,075 --> 00:03:07,212
and I discuss

86
00:03:07,907 --> 00:03:10,058
several of these promising technologies and kinda where

87
00:03:10,058 --> 00:03:11,413
they are now and where they might have

88
00:03:11,413 --> 00:03:14,697
some use. I will also say that since

89
00:03:14,697 --> 00:03:15,893
we had our interview,

90
00:03:16,611 --> 00:03:18,148
Google's privacy chief

91
00:03:19,084 --> 00:03:21,397
left, I don't know if they were fired,

92
00:03:21,636 --> 00:03:24,203
but after 13 years with Google,

93
00:03:24,841 --> 00:03:26,676
some of which time must have,

94
00:03:27,314 --> 00:03:29,388
overlap I think with irene, has left in,

95
00:03:29,946 --> 00:03:31,940
according to Google will not be replaced.

96
00:03:32,914 --> 00:03:33,414
So

97
00:03:33,793 --> 00:03:35,231
I don't know what to read into that,

98
00:03:35,551 --> 00:03:37,788
but it it doesn't sound like it's a

99
00:03:37,788 --> 00:03:40,265
good thing. So again, I've got some information

100
00:03:40,265 --> 00:03:41,797
for you about the book surge that I'm

101
00:03:41,797 --> 00:03:43,151
doing. I really love to get your help

102
00:03:43,151 --> 00:03:45,722
on that. Stay tuned after the interview for

103
00:03:45,860 --> 00:03:48,170
important information about that, but let's not waste

104
00:03:48,170 --> 00:03:49,445
more time. Let's get to my interview with

105
00:03:49,525 --> 00:03:50,401
Irene nap.

106
00:03:56,547 --> 00:03:59,015
Our aim is an ex Google information privacy

107
00:03:59,015 --> 00:04:01,509
x Burton now in the profit world and

108
00:04:01,509 --> 00:04:03,830
an activist around topics including trans rights and

109
00:04:03,830 --> 00:04:06,150
mental health welcome Irene. Thank you so much.

110
00:04:06,310 --> 00:04:07,763
I'm glad to be here. Oh, I cannot

111
00:04:07,763 --> 00:04:09,918
wait for this. So I bash Google pretty

112
00:04:09,918 --> 00:04:11,914
much on the reg and you know, over

113
00:04:11,914 --> 00:04:14,627
privacy in particular. And so I'm really interested

114
00:04:14,627 --> 00:04:16,556
to talk to somebody who's actually work there

115
00:04:16,556 --> 00:04:17,748
in the privacy space no.

116
00:04:18,464 --> 00:04:20,133
So this this is going gonna be great.

117
00:04:20,292 --> 00:04:22,677
But ironically, even though that you're from Google

118
00:04:22,677 --> 00:04:24,346
and I'd bash. Them I'm I'm hoping today,

119
00:04:24,505 --> 00:04:26,831
we can actually talk about perhaps ways that

120
00:04:26,831 --> 00:04:29,145
we can use data in a positive way.

121
00:04:29,305 --> 00:04:30,981
So a little bit of a little twist

122
00:04:30,981 --> 00:04:32,178
on that. Why don't we start by you

123
00:04:32,178 --> 00:04:33,556
telling us a little bit more about

124
00:04:33,947 --> 00:04:35,615
Internet safety labs and the kind of work

125
00:04:35,615 --> 00:04:36,171
you guys do there?

126
00:04:37,045 --> 00:04:39,031
Sure. So we believe, you know, I...

127
00:04:39,826 --> 00:04:41,255
In my career at Google, I felt that

128
00:04:41,335 --> 00:04:43,416
I was holding the company. Accountable from inside.

129
00:04:43,972 --> 00:04:45,005
I felt that that was what I was

130
00:04:45,005 --> 00:04:46,833
doing was making things better for everybody,

131
00:04:47,468 --> 00:04:50,187
But I no longer believe that change can

132
00:04:50,187 --> 00:04:51,942
come solely from inside the system. I believe

133
00:04:51,942 --> 00:04:52,921
there needs to be

134
00:04:53,458 --> 00:04:53,856
external,

135
00:04:54,415 --> 00:04:54,915
independent

136
00:04:55,452 --> 00:04:59,142
oversight, and I believe that independent safety testing

137
00:04:59,773 --> 00:05:02,341
is necessary. And we say safety, not privacy

138
00:05:02,954 --> 00:05:05,897
because privacy is the company's framing. Right? They

139
00:05:05,897 --> 00:05:07,908
want us to view this as an exciting

140
00:05:08,377 --> 00:05:10,125
new thing. And I'm here to tell you

141
00:05:10,125 --> 00:05:11,022
actually safety

142
00:05:11,396 --> 00:05:13,381
is a boring thing that we should all

143
00:05:13,381 --> 00:05:14,731
be routinely demanding.

144
00:05:15,367 --> 00:05:15,867
Every

145
00:05:16,399 --> 00:05:17,455
child's toy

146
00:05:17,844 --> 00:05:19,912
you know, if you buy a toy teddy

147
00:05:19,912 --> 00:05:22,379
bear, somebody has made sure that it's fire

148
00:05:22,379 --> 00:05:25,084
proof material. Mh. There is nothing like that

149
00:05:25,084 --> 00:05:26,755
for software. Nothing what.

150
00:05:27,564 --> 00:05:30,341
Yeah. Except these like... Yeah. And and and

151
00:05:30,341 --> 00:05:31,611
there needs to be. We need to live

152
00:05:31,611 --> 00:05:33,357
in a world, you know, it it should

153
00:05:33,357 --> 00:05:34,961
be clear to the... And to the public

154
00:05:34,961 --> 00:05:36,711
these days that software can harm people.

155
00:05:37,268 --> 00:05:37,768
And,

156
00:05:38,381 --> 00:05:40,290
you know, we need the same mechanisms that

157
00:05:40,290 --> 00:05:42,699
we have in every other area of society

158
00:05:42,915 --> 00:05:44,448
to the companies want us to believe it's

159
00:05:44,448 --> 00:05:46,606
a big ass, but it's not. Well, I

160
00:05:46,606 --> 00:05:47,964
just read that point recently on a podcast

161
00:05:47,964 --> 00:05:49,104
right where we've got liability

162
00:05:49,642 --> 00:05:51,560
for every other industry I can think of.

163
00:05:51,800 --> 00:05:53,895
You can't ship an unsafe car or unsafe

164
00:05:53,895 --> 00:05:57,014
food or unsafe drugs or unsafe toys and

165
00:05:57,014 --> 00:05:59,175
yet, for some reason, software has gotten away

166
00:05:59,175 --> 00:06:00,868
with this terms of use where they say,

167
00:06:01,027 --> 00:06:02,785
you know, use as is at your own

168
00:06:02,785 --> 00:06:04,383
risk and how that ever came to, you

169
00:06:04,383 --> 00:06:06,060
know, I why that's still the case? I

170
00:06:06,060 --> 00:06:07,754
don't know. I think a lot of the

171
00:06:07,754 --> 00:06:09,754
white has got traction has been, you know,

172
00:06:10,875 --> 00:06:13,615
ideology that people perhaps accepted with

173
00:06:14,245 --> 00:06:16,547
less critical of a critical mind, then they

174
00:06:16,547 --> 00:06:18,927
should have, people including me in my past.

175
00:06:19,483 --> 00:06:20,991
You know, I I want to stay on

176
00:06:20,991 --> 00:06:22,394
partisan in here so I won't

177
00:06:23,070 --> 00:06:24,500
you know, go into depth on that. But

178
00:06:24,659 --> 00:06:27,519
I think that there's this tech ethos that

179
00:06:27,519 --> 00:06:29,506
we all kind of bought into, like, Yeah.

180
00:06:29,744 --> 00:06:31,928
We will make the world better through technology

181
00:06:31,985 --> 00:06:34,689
whether you like it or not. But society

182
00:06:34,689 --> 00:06:36,757
is a stakeholder if you're seeking to change

183
00:06:36,757 --> 00:06:37,951
society. Right? Yeah.

184
00:06:38,762 --> 00:06:40,995
Yeah. And and, you know, I think early

185
00:06:40,995 --> 00:06:42,829
on, it was you know, kinda like the

186
00:06:42,829 --> 00:06:44,264
the old Facebook, thing, you know, move fast

187
00:06:44,264 --> 00:06:45,380
and break things, and I think they were

188
00:06:45,380 --> 00:06:46,949
like, hands off at tool, you know, this

189
00:06:47,069 --> 00:06:49,300
too Nascent. This is too early. You gonna

190
00:06:49,300 --> 00:06:51,053
s this amazing thing that we're creating here,

191
00:06:51,212 --> 00:06:52,168
You know, it hands off for a little

192
00:06:52,168 --> 00:06:53,363
while, and I think we just never got

193
00:06:53,363 --> 00:06:53,841
hands back on.

194
00:06:54,652 --> 00:06:56,164
I tell my team at work moves slow

195
00:06:56,164 --> 00:06:58,471
and be careful. I tell them do not

196
00:06:58,471 --> 00:07:00,221
listen to Facebook. They are wrong.

197
00:07:00,937 --> 00:07:01,096
Yes.

198
00:07:02,788 --> 00:07:04,866
Alright. So let's get into this. How did

199
00:07:04,866 --> 00:07:06,624
you come to work at Google? You know,

200
00:07:06,784 --> 00:07:08,382
and then tell us, what was your experience

201
00:07:08,382 --> 00:07:10,060
there and, you know, why did you ultimately

202
00:07:10,060 --> 00:07:10,220
leave?

203
00:07:11,353 --> 00:07:13,103
Yeah. So I had what is kind of

204
00:07:13,103 --> 00:07:15,650
a traditional life path, but a pretty traditional

205
00:07:15,650 --> 00:07:18,298
tech path in that I'm self educated

206
00:07:18,689 --> 00:07:20,370
and I moved to the company that I

207
00:07:20,370 --> 00:07:22,709
saw as the 1 doing the most interesting

208
00:07:23,649 --> 00:07:25,569
work to truly advance the state of the

209
00:07:25,569 --> 00:07:25,810
art.

210
00:07:26,383 --> 00:07:28,212
Because I believe that I still believe in

211
00:07:28,212 --> 00:07:29,883
making a better world by building things,

212
00:07:30,519 --> 00:07:33,008
and I believed in the company's stated ideals

213
00:07:33,318 --> 00:07:34,829
at the time I thought they did too.

214
00:07:35,306 --> 00:07:36,897
What... If you don't mind asking, what what

215
00:07:36,897 --> 00:07:37,851
what kind of time were are we talking

216
00:07:37,851 --> 00:07:39,759
about here? I was there for 5 years.

217
00:07:39,918 --> 00:07:41,191
What, like, when did you spend how long

218
00:07:41,191 --> 00:07:42,543
it ago did you start? Oh, yeah. I'm

219
00:07:42,543 --> 00:07:43,907
sorry. I I was there.

220
00:07:44,542 --> 00:07:46,842
I joined in, wow. It's been so long

221
00:07:46,842 --> 00:07:49,063
that I need to properly remember this. I

222
00:07:49,063 --> 00:07:50,515
joined in 20 14.

223
00:07:50,888 --> 00:07:53,319
Okay. I left in 20 19

224
00:07:54,339 --> 00:07:56,180
through a sequence of events bound up with

225
00:07:56,180 --> 00:07:57,779
what had become a labor movement at that

226
00:07:57,779 --> 00:07:58,020
point.

227
00:07:58,913 --> 00:08:00,902
So you got in there with ideals? What

228
00:08:00,902 --> 00:08:02,333
did you find when you got in there

229
00:08:02,333 --> 00:08:03,924
and what was it that ultimately lip caused

230
00:08:03,924 --> 00:08:06,169
you leave? Well, at at first, I felt

231
00:08:06,250 --> 00:08:07,529
I found a lot of people who seem

232
00:08:07,529 --> 00:08:08,569
to share my ideals,

233
00:08:09,290 --> 00:08:12,410
You know, I'm transgender right, and we are

234
00:08:12,410 --> 00:08:15,379
disproportionately represented in the tech industry. More so

235
00:08:15,379 --> 00:08:17,920
than any other industry, partly because many of

236
00:08:17,920 --> 00:08:19,826
us screw up online, you know, for people

237
00:08:19,826 --> 00:08:21,255
in my age. The Internet was the first

238
00:08:21,255 --> 00:08:23,578
place we could be ourselves yeah without fear

239
00:08:23,578 --> 00:08:26,052
of persecution. And so many of us became

240
00:08:26,052 --> 00:08:29,266
very technically knowledgeable, and actually, many of us

241
00:08:29,405 --> 00:08:31,480
specifically went into information privacy.

242
00:08:32,054 --> 00:08:34,550
Which makes sense when you hear it, but

243
00:08:34,848 --> 00:08:35,487
right. Right.

244
00:08:36,604 --> 00:08:38,041
So, you know, I found a lot of,

245
00:08:38,121 --> 00:08:39,877
like, minded people, some of whom I still

246
00:08:39,877 --> 00:08:40,835
very much admire.

247
00:08:41,888 --> 00:08:43,564
And, you know, who I kind of learned

248
00:08:43,564 --> 00:08:46,278
my own activism skills from, like, I had

249
00:08:46,278 --> 00:08:48,274
had events in my life, which had led

250
00:08:48,274 --> 00:08:50,442
me down this path of feeling that we

251
00:08:50,442 --> 00:08:51,340
should not passively

252
00:08:52,112 --> 00:08:53,623
accept the world we are given. We should

253
00:08:53,623 --> 00:08:54,578
seek to make it better.

254
00:08:55,293 --> 00:08:56,963
But it was what... It was during those

255
00:08:56,963 --> 00:08:58,020
first 2

256
00:08:58,807 --> 00:09:00,314
years that I was at Google that that

257
00:09:00,314 --> 00:09:03,409
really jailed for me. And at first, the

258
00:09:03,409 --> 00:09:05,258
company seemed to welcome this criticism.

259
00:09:05,965 --> 00:09:08,205
It had a culture of appearing to welcome

260
00:09:08,205 --> 00:09:09,264
bottom up grassroots

261
00:09:09,884 --> 00:09:12,365
leadership. Mh. And people I knew had made

262
00:09:12,365 --> 00:09:14,580
change of various sorts through that sort of

263
00:09:14,699 --> 00:09:15,358
bottom up

264
00:09:15,892 --> 00:09:17,722
before, and they were always told,

265
00:09:18,756 --> 00:09:20,824
this is good. The company welcomes this.

266
00:09:21,540 --> 00:09:23,779
Bottom I I still believe that many of

267
00:09:23,779 --> 00:09:26,344
us, you know, we did bring important perspective

268
00:09:26,479 --> 00:09:27,297
to the privacy

269
00:09:27,750 --> 00:09:29,497
conversations. And I'm sure we'll talk more later.

270
00:09:30,229 --> 00:09:31,825
About the corporate narrative there and why I

271
00:09:31,825 --> 00:09:34,379
don't recommend accepting it un critically. But I

272
00:09:34,379 --> 00:09:36,933
believe that marginalized experiences are necessary part of

273
00:09:36,933 --> 00:09:38,311
that because we're thinking

274
00:09:38,623 --> 00:09:40,446
of dangers that other people aren't because of

275
00:09:40,446 --> 00:09:42,824
our moved experience. It's it's a job qualification.

276
00:09:45,360 --> 00:09:47,277
And not just to ourselves, but to other

277
00:09:47,277 --> 00:09:49,178
marginalized groups who we talked to on the

278
00:09:49,178 --> 00:09:51,634
regular and, you know, you know, I felt

279
00:09:51,634 --> 00:09:52,847
rewarded for my privacy

280
00:09:53,472 --> 00:09:56,276
work there. I have come to understand

281
00:09:56,967 --> 00:09:58,000
that, like,

282
00:09:58,794 --> 00:10:00,859
upper management, and that by that, I mean,

283
00:10:01,097 --> 00:10:01,256
like,

284
00:10:01,986 --> 00:10:03,281
Vpn level and above

285
00:10:03,735 --> 00:10:05,587
views privacy as,

286
00:10:05,962 --> 00:10:07,950
you know, it's revenue negative, but it's intended

287
00:10:07,950 --> 00:10:09,483
to be, but they view it

288
00:10:09,794 --> 00:10:12,351
through a much more cynical lens than what

289
00:10:12,351 --> 00:10:14,268
the people I directly reported through,

290
00:10:15,227 --> 00:10:16,846
led us to believe and

291
00:10:17,478 --> 00:10:18,993
some of those people, I believe that they

292
00:10:18,993 --> 00:10:20,508
believed it and some of them. I believe

293
00:10:20,508 --> 00:10:21,465
they lied to my face.

294
00:10:22,581 --> 00:10:25,133
And that's been quite a thing to unpack

295
00:10:25,133 --> 00:10:25,850
over the years.

296
00:10:26,504 --> 00:10:28,662
So did you find... I I don't really

297
00:10:28,662 --> 00:10:31,618
honestly follow the other... When Larry Sergey we're

298
00:10:31,618 --> 00:10:34,335
were kinda descending. III don't really follow it.

299
00:10:34,495 --> 00:10:36,424
So was there Did you notice these changes

300
00:10:36,424 --> 00:10:39,134
were kind of tied to upper management changes

301
00:10:39,134 --> 00:10:40,329
or does it... Was it a company shift?

302
00:10:40,488 --> 00:10:42,002
Or It was. Okay.

303
00:10:42,814 --> 00:10:44,572
It it was, around the time of the

304
00:10:44,652 --> 00:10:47,289
Alphabet restructuring was when things really started to

305
00:10:47,289 --> 00:10:47,848
go south.

306
00:10:48,567 --> 00:10:50,896
Now I have... It's not just that, like,

307
00:10:51,055 --> 00:10:53,039
Sun is a different style of leader than

308
00:10:53,197 --> 00:10:54,308
Larry and Sergey were.

309
00:10:54,943 --> 00:10:57,006
He is. He's... You know, his background is

310
00:10:57,006 --> 00:10:58,776
from the Kin, the consulting firm

311
00:10:59,483 --> 00:11:02,109
which I don't think Highly of either the

312
00:11:02,109 --> 00:11:02,848
kind of

313
00:11:03,701 --> 00:11:04,679
very bleak

314
00:11:05,054 --> 00:11:07,681
bureaucratic worldview that it em amaze or the

315
00:11:07,681 --> 00:11:08,818
decisions that he made

316
00:11:09,128 --> 00:11:10,639
and how he seemed to be a total

317
00:11:10,639 --> 00:11:13,263
non entity from the from my perspective. And

318
00:11:13,263 --> 00:11:14,479
remember, I had more visibility

319
00:11:14,933 --> 00:11:17,080
up the Org chart than most people did

320
00:11:17,080 --> 00:11:18,614
because privacy work is horizontal.

321
00:11:19,164 --> 00:11:21,741
Meaning it touches everything the company does

322
00:11:22,440 --> 00:11:22,919
necessarily.

323
00:11:23,398 --> 00:11:24,916
So it was my job to know all

324
00:11:24,916 --> 00:11:26,035
the company's damn secrets.

325
00:11:28,755 --> 00:11:31,380
So what is the privacy organization like? Like,

326
00:11:31,459 --> 00:11:33,944
how what what what is that structure like

327
00:11:33,944 --> 00:11:36,101
within a company like Google that's... Who who's

328
00:11:36,101 --> 00:11:38,818
an ad company. Right? Yeah. So 1 thing

329
00:11:38,818 --> 00:11:40,656
to keep in mind, the reason these companies

330
00:11:40,656 --> 00:11:42,587
have privacy teams is not out goodness of

331
00:11:42,587 --> 00:11:44,736
their hearts. It's because they have at various

332
00:11:44,736 --> 00:11:46,487
times in their history been sued by the

333
00:11:46,726 --> 00:11:47,123
Ftc,

334
00:11:47,601 --> 00:11:49,750
and in that consent decree resulting from that

335
00:11:49,750 --> 00:11:51,421
lawsuit because they effectively lost.

336
00:11:51,993 --> 00:11:54,219
They were required by law to hire a

337
00:11:54,219 --> 00:11:57,160
privacy expert, And they chose, of course, to

338
00:11:57,160 --> 00:11:59,068
hire more than 1 privacy expert.

339
00:11:59,879 --> 00:12:01,315
Because they don't wish to be seen as

340
00:12:01,315 --> 00:12:02,353
doing the bare minimum.

341
00:12:03,390 --> 00:12:05,466
But, you know, they never actually moved back

342
00:12:05,466 --> 00:12:07,801
and told the Ftc much later

343
00:12:08,514 --> 00:12:10,034
when I was doing a bit of worker

344
00:12:10,034 --> 00:12:11,714
shareholder On my, I was actually...

345
00:12:12,434 --> 00:12:13,794
I briefly spoke to,

346
00:12:14,754 --> 00:12:16,835
somebody from the New York State attorney general's

347
00:12:16,835 --> 00:12:19,086
office who who informed me that as far

348
00:12:19,086 --> 00:12:20,763
as they knew they thought they might get

349
00:12:20,763 --> 00:12:23,157
Google the hire 1 privacy person. And as

350
00:12:23,157 --> 00:12:25,073
far as they knew that must have happened,

351
00:12:25,232 --> 00:12:26,190
but they never heard back.

352
00:12:26,921 --> 00:12:28,987
Right. Mh. So they thought perhaps that 1

353
00:12:28,987 --> 00:12:29,861
person was me.

354
00:12:30,815 --> 00:12:31,133
Right?

355
00:12:32,722 --> 00:12:33,119
Quite a thing.

356
00:12:34,168 --> 00:12:34,327
Yeah.

357
00:12:35,441 --> 00:12:36,338
Alright. So,

358
00:12:36,873 --> 00:12:38,384
yeah know, what are some of the key

359
00:12:38,384 --> 00:12:41,408
priorities or or principles guiding Google's approach to

360
00:12:41,408 --> 00:12:42,544
privacy during

361
00:12:43,014 --> 00:12:45,556
know, your time there, that you saw? And

362
00:12:45,556 --> 00:12:46,985
maybe how did that evolve over the years?

363
00:12:47,144 --> 00:12:48,892
And then, you know, we've touched us a

364
00:12:48,892 --> 00:12:50,083
little bit, but how did that comp port

365
00:12:50,083 --> 00:12:51,616
with your your

366
00:12:51,990 --> 00:12:52,228
expectations?

367
00:12:53,595 --> 00:12:53,993
Yeah.

368
00:12:55,027 --> 00:12:57,333
At first, I saw the company willing to

369
00:12:57,333 --> 00:12:59,774
take a real financial hit for what a

370
00:12:59,894 --> 00:13:02,775
appeared to be, you know, moral reasons, and

371
00:13:03,069 --> 00:13:05,292
I was happy about that. I no longer

372
00:13:05,292 --> 00:13:06,959
really believe that the moral reasons were the

373
00:13:06,959 --> 00:13:08,919
primary ones, and I can't talk about the

374
00:13:09,039 --> 00:13:11,194
tales of this. I... This is something I

375
00:13:11,194 --> 00:13:12,791
practiced over the years knowing what I can

376
00:13:12,791 --> 00:13:13,589
and cannot say.

377
00:13:14,946 --> 00:13:17,180
At first, I was relatively happy with that,

378
00:13:17,754 --> 00:13:20,148
as time passed, I came to see that

379
00:13:20,148 --> 00:13:20,967
they were

380
00:13:21,584 --> 00:13:23,738
accepting pushback only to the 4... Only to

381
00:13:23,738 --> 00:13:25,036
the extent that it didn't

382
00:13:25,508 --> 00:13:27,660
challenged the stuff they had already decided to

383
00:13:27,660 --> 00:13:28,160
do

384
00:13:28,856 --> 00:13:30,928
effectively or that they had decided they needed

385
00:13:30,928 --> 00:13:32,761
or that they had decided they felt entitled

386
00:13:32,761 --> 00:13:33,001
too.

387
00:13:33,652 --> 00:13:36,116
I also came to understand, and I told

388
00:13:36,116 --> 00:13:38,183
my teammates this, but even as somebody who

389
00:13:38,183 --> 00:13:40,090
knows it and is trying to correct for

390
00:13:40,090 --> 00:13:41,918
it. It's hard. We're all biased,

391
00:13:43,045 --> 00:13:45,188
when we're within an incentive structure like that

392
00:13:45,188 --> 00:13:45,926
to view

393
00:13:46,457 --> 00:13:46,957
management

394
00:13:47,727 --> 00:13:50,362
dictates as laws as of nature. They are

395
00:13:50,362 --> 00:13:52,671
not. They are things that humans decided.

396
00:13:53,388 --> 00:13:55,856
But especially if it comes down from a

397
00:13:55,856 --> 00:13:57,369
way up the chain as just a yes

398
00:13:57,369 --> 00:13:57,926
or a no.

399
00:13:58,419 --> 00:13:59,860
And you have no context on why it

400
00:13:59,860 --> 00:14:01,720
was a yes or no. You start thinking

401
00:14:01,860 --> 00:14:03,620
or what can we do within that constraint.

402
00:14:03,860 --> 00:14:06,100
Right? And you start justifying it to yourself.

403
00:14:06,339 --> 00:14:07,240
You start thinking

404
00:14:07,874 --> 00:14:09,549
They must have had good reason for that.

405
00:14:10,586 --> 00:14:12,021
The what the thing that really changed my

406
00:14:12,021 --> 00:14:15,291
mind was a, like, highly politicized and reported

407
00:14:15,291 --> 00:14:16,886
on thing. Is it okay to get into

408
00:14:16,886 --> 00:14:17,206
that year?

409
00:14:17,859 --> 00:14:20,096
Sure if you'd like? Yeah. Sure. It was

410
00:14:20,096 --> 00:14:21,774
the James D incident.

411
00:14:22,254 --> 00:14:24,571
And this was a this was a person,

412
00:14:24,971 --> 00:14:26,424
by the way, I'm on it I'm a

413
00:14:26,424 --> 00:14:29,301
autistic non binary person. This was a... This

414
00:14:29,301 --> 00:14:31,858
was an autistic man from what I know.

415
00:14:32,577 --> 00:14:33,936
I believe that was is how you would

416
00:14:33,936 --> 00:14:35,029
self describe, but I don't

417
00:14:35,549 --> 00:14:36,049
who

418
00:14:36,828 --> 00:14:39,305
used his neuro divergence as an excuse to

419
00:14:39,305 --> 00:14:40,285
write this

420
00:14:40,983 --> 00:14:43,381
political piece expressing the view that women should

421
00:14:43,381 --> 00:14:44,260
not work at Google.

422
00:14:44,993 --> 00:14:45,493
And

423
00:14:46,268 --> 00:14:48,579
so a bunch of us were quite upset

424
00:14:48,579 --> 00:14:51,527
by this. Sure. And we looked into it

425
00:14:51,527 --> 00:14:53,631
and what we found was that this document

426
00:14:53,845 --> 00:14:55,060
appeared to have been written

427
00:14:55,512 --> 00:14:57,972
to tow a very specific legal line.

428
00:14:58,687 --> 00:14:58,925
Like,

429
00:14:59,971 --> 00:15:01,422
things that get lit

430
00:15:01,794 --> 00:15:04,568
that form test cases were always constructed. You

431
00:15:04,568 --> 00:15:05,836
know, I know this speaks become an activist

432
00:15:05,836 --> 00:15:06,629
in other context.

433
00:15:07,438 --> 00:15:09,906
You never happen on those fact patterns by

434
00:15:09,906 --> 00:15:12,158
accident unless you sort through thousands of people

435
00:15:12,295 --> 00:15:14,047
who've all done something like this. It's always

436
00:15:14,047 --> 00:15:15,343
like a plane who

437
00:15:15,813 --> 00:15:17,325
went out of their way to create a

438
00:15:17,325 --> 00:15:20,748
good test case by not having extraneous facts.

439
00:15:21,783 --> 00:15:23,158
And and this document

440
00:15:23,613 --> 00:15:25,698
appeared according to people we talked to who

441
00:15:25,858 --> 00:15:28,011
I can't name to have been constructed to

442
00:15:28,011 --> 00:15:29,846
be a test case for a legal principle,

443
00:15:30,005 --> 00:15:31,521
which they appear to be going for the

444
00:15:31,521 --> 00:15:32,021
argument

445
00:15:32,478 --> 00:15:34,972
that saying that women shouldn't work at Google

446
00:15:35,444 --> 00:15:37,381
is protected labor

447
00:15:38,158 --> 00:15:38,318
organization.

448
00:15:39,515 --> 00:15:39,754
Okay.

449
00:15:40,632 --> 00:15:42,787
And we were deeply upset by this, and

450
00:15:42,787 --> 00:15:44,795
it prompted us to read up on what

451
00:15:44,795 --> 00:15:45,908
our labor rights are,

452
00:15:46,624 --> 00:15:48,454
which we had never done before. In all

453
00:15:48,454 --> 00:15:48,954
this

454
00:15:49,488 --> 00:15:51,809
activism up to now, we had been perceiving

455
00:15:51,809 --> 00:15:53,876
it as something welcomed to buy the company,

456
00:15:54,274 --> 00:15:55,943
you know, which I no longer think it

457
00:15:55,943 --> 00:15:57,453
really was. I think they had a variety

458
00:15:57,453 --> 00:15:58,860
of strategies to make us

459
00:15:59,296 --> 00:16:01,441
feel that way while diff confusing any real

460
00:16:01,441 --> 00:16:03,507
change by making promises that will never be

461
00:16:03,507 --> 00:16:04,484
followed up on

462
00:16:04,937 --> 00:16:05,675
by finding

463
00:16:06,464 --> 00:16:08,300
exactly what they can get away with by

464
00:16:08,300 --> 00:16:10,615
breaking every moral decision down into a million

465
00:16:10,615 --> 00:16:11,913
fine grained choices

466
00:16:12,850 --> 00:16:14,367
so that you moves your concept of right

467
00:16:14,367 --> 00:16:15,379
and wrong. It's

468
00:16:15,739 --> 00:16:18,053
lost in the paperwork. So we we learned

469
00:16:18,053 --> 00:16:19,888
that we actually had a right to talk

470
00:16:19,888 --> 00:16:22,043
about politics as long as we did so

471
00:16:22,043 --> 00:16:23,022
within the context,

472
00:16:23,734 --> 00:16:26,206
of how we're working conditions. And certainly,

473
00:16:26,605 --> 00:16:28,200
if somebody is saying that I should not

474
00:16:28,200 --> 00:16:31,310
be working here, that is a working condition

475
00:16:31,310 --> 00:16:32,745
that it's a hostile work environment.

476
00:16:33,320 --> 00:16:35,159
And and so we we read up on

477
00:16:35,159 --> 00:16:37,079
our rights and we realized that more broadly,

478
00:16:37,320 --> 00:16:39,320
we actually even had a right to talk

479
00:16:39,320 --> 00:16:41,159
to the press about our working conditions.

480
00:16:42,053 --> 00:16:44,203
And that Google had, you know, through a

481
00:16:44,203 --> 00:16:47,251
variety of internal policies designed to mislead without

482
00:16:47,309 --> 00:16:49,459
saying any prohibit anything that they're not legally

483
00:16:49,459 --> 00:16:50,256
allowed to prohibit.

484
00:16:51,068 --> 00:16:52,345
They had met us to the belief that

485
00:16:52,345 --> 00:16:54,339
it's very dangerous to talk to press. And

486
00:16:54,339 --> 00:16:55,956
actually, I think press

487
00:16:56,333 --> 00:16:58,261
or your ally as long as you know,

488
00:16:58,420 --> 00:16:59,690
as long as you know who you're talking

489
00:16:59,690 --> 00:16:59,849
to,

490
00:17:00,564 --> 00:17:02,311
when when you have a situation like this.

491
00:17:02,708 --> 00:17:05,012
And so when we began pushing in ways

492
00:17:05,012 --> 00:17:07,174
that went outside the system and that we're

493
00:17:07,174 --> 00:17:08,392
in danger of creating

494
00:17:09,007 --> 00:17:09,963
real change.

495
00:17:10,680 --> 00:17:11,180
Suddenly,

496
00:17:11,875 --> 00:17:14,844
management was very hostile even to small reasonable

497
00:17:15,141 --> 00:17:15,434
suggest.

498
00:17:15,954 --> 00:17:16,194
Right?

499
00:17:16,993 --> 00:17:19,150
Like, again, these are our legal rights. Yeah.

500
00:17:19,390 --> 00:17:21,308
But they know more... And it's my job

501
00:17:21,308 --> 00:17:23,066
as a privacy person or it was my

502
00:17:23,066 --> 00:17:26,039
job as a privacy person. To identify things

503
00:17:26,039 --> 00:17:28,359
about the products that could hurt people and

504
00:17:28,359 --> 00:17:30,599
to make sure that they don't. Right? Right?

505
00:17:31,000 --> 00:17:33,400
Suddenly, the company started its stonewall a some

506
00:17:33,400 --> 00:17:33,640
stuff.

507
00:17:34,293 --> 00:17:36,203
And sub diverting even some of its own

508
00:17:36,203 --> 00:17:36,703
procedures

509
00:17:37,078 --> 00:17:38,432
for certain sorts of decisions.

510
00:17:39,387 --> 00:17:41,615
So it became clear over time that the

511
00:17:41,615 --> 00:17:43,740
company had a very effective internal mess

512
00:17:44,336 --> 00:17:46,585
strategy for def using this kind of criticism

513
00:17:46,642 --> 00:17:48,868
and making us all kind of blame each

514
00:17:48,868 --> 00:17:50,696
other for not raising concerns in the right

515
00:17:50,696 --> 00:17:51,912
way or what have you

516
00:17:52,287 --> 00:17:53,891
and that they had been doing this from

517
00:17:53,891 --> 00:17:55,875
the very start anytime we tried to raise

518
00:17:55,875 --> 00:17:56,375
anything

519
00:17:56,906 --> 00:17:58,969
really big that really challenged their power,

520
00:17:59,763 --> 00:18:02,482
especially, you know, the interest of marginalized communities,

521
00:18:02,641 --> 00:18:04,074
both inside and outside the company,

522
00:18:04,632 --> 00:18:07,021
very frequently do wind up at odds with

523
00:18:07,021 --> 00:18:10,138
the profit interest. Oh, sure. Yeah. Yeah. And,

524
00:18:10,296 --> 00:18:12,356
you know, so having studied our labor rights,

525
00:18:12,594 --> 00:18:14,496
we had... And it was an all consuming

526
00:18:14,496 --> 00:18:16,580
couple of years, which I will spare you

527
00:18:16,968 --> 00:18:20,169
but suffice to say it became clear that

528
00:18:20,463 --> 00:18:22,528
with the skill set I had acquired in

529
00:18:22,528 --> 00:18:24,852
talking about all this, I could advance the

530
00:18:24,852 --> 00:18:26,848
cause better from outside the company where I'm

531
00:18:27,008 --> 00:18:28,844
more liberty able everybody to say critical things,

532
00:18:29,483 --> 00:18:31,638
than I could from inside. And therefore, I

533
00:18:31,638 --> 00:18:33,096
pursued a kind of

534
00:18:34,450 --> 00:18:37,730
I I diverted the company's hostile attention toward

535
00:18:37,730 --> 00:18:39,649
myself so that other people wouldn't take the

536
00:18:39,649 --> 00:18:41,490
brunt of it and so that we could

537
00:18:41,490 --> 00:18:43,838
test the limits and know where they were

538
00:18:44,134 --> 00:18:46,202
so that people could make conscious decisions about

539
00:18:46,202 --> 00:18:47,395
when to go over those lines.

540
00:18:48,111 --> 00:18:51,475
And that, you know, IIII started going down

541
00:18:51,475 --> 00:18:53,551
that path consciously about a year before I

542
00:18:53,551 --> 00:18:54,748
separated from the company.

543
00:18:55,626 --> 00:18:57,143
Oh, that sounds terrible. I'm sorry. You had

544
00:18:57,143 --> 00:18:58,580
to go through all that. Oh, no. No.

545
00:18:58,740 --> 00:19:00,269
I'm very proud of it. This is a

546
00:19:00,269 --> 00:19:01,703
part of my life that I look back

547
00:19:01,703 --> 00:19:01,942
on.

548
00:19:02,580 --> 00:19:03,956
You know, I am horrified

549
00:19:04,413 --> 00:19:04,913
at

550
00:19:05,289 --> 00:19:06,426
everything the company

551
00:19:06,740 --> 00:19:08,980
does and did, but I am very proud

552
00:19:08,980 --> 00:19:10,119
of this, and

553
00:19:11,059 --> 00:19:11,380
you know,

554
00:19:13,140 --> 00:19:13,299
yeah.

555
00:19:15,622 --> 00:19:17,450
Okay. Let me ask you this question. I'm

556
00:19:17,450 --> 00:19:18,801
not sure how much perspective you've got in

557
00:19:18,801 --> 00:19:21,289
this. But what is your sense for

558
00:19:21,679 --> 00:19:24,079
how Google compares to some of the other

559
00:19:24,079 --> 00:19:26,400
ad driven companies like Facebook or Twitter. And

560
00:19:26,400 --> 00:19:27,440
some of these other companies that have these

561
00:19:27,440 --> 00:19:29,059
quote unquote free services that

562
00:19:29,440 --> 00:19:31,053
we are the product. How do you think

563
00:19:31,053 --> 00:19:33,049
they they all compare? Who's who's the worst

564
00:19:33,049 --> 00:19:34,326
day to collection you collect collects the most

565
00:19:34,326 --> 00:19:36,003
stuff? I don't know. What... Do you have

566
00:19:36,003 --> 00:19:37,280
a dev take on that? A hot take?

567
00:19:37,998 --> 00:19:38,238
Oh,

568
00:19:38,956 --> 00:19:39,675
well, yes.

569
00:19:40,645 --> 00:19:41,997
If you would ask me why I was

570
00:19:41,997 --> 00:19:43,350
working there, I would have said that, of

571
00:19:43,350 --> 00:19:45,100
course, Google is the best because they have

572
00:19:45,100 --> 00:19:47,658
all these privacy people. Right? I don't know

573
00:19:47,658 --> 00:19:49,722
that I believe that. On the other hand,

574
00:19:49,960 --> 00:19:52,420
the competition is bragging about their mel fe.

575
00:19:53,055 --> 00:19:55,277
So Google kind of tries to play, like,

576
00:19:56,008 --> 00:19:58,481
we're the good ones to regulators. And in

577
00:19:58,481 --> 00:20:01,033
my role, I was occasionally the 1 responding

578
00:20:01,033 --> 00:20:03,426
to regulatory inquiries by the way, through a

579
00:20:03,426 --> 00:20:04,782
lawyer. Like I didn't get my name on

580
00:20:04,782 --> 00:20:08,311
it, pleasure. Google tries to, like, control the

581
00:20:08,311 --> 00:20:09,989
narrative by being, like, we do all this

582
00:20:09,989 --> 00:20:12,786
privacy research. We have this groundbreaking math.

583
00:20:13,505 --> 00:20:13,825
Right?

584
00:20:14,397 --> 00:20:16,147
We're the good ones, you can trust us

585
00:20:16,147 --> 00:20:18,613
to lead the way, but it's kinda tell

586
00:20:18,693 --> 00:20:20,363
Dungeons and Dragons joke that play with your

587
00:20:20,363 --> 00:20:22,610
audience. It's Are you kidding? It's called firewalls

588
00:20:22,610 --> 00:20:24,524
don't up dragons? I thought it might be.

589
00:20:24,684 --> 00:20:25,721
I thought it might be. So there's an

590
00:20:25,721 --> 00:20:28,512
old joke, right? The party of adventurers is

591
00:20:28,512 --> 00:20:30,346
in the woods, and they're attacked by a

592
00:20:30,346 --> 00:20:30,745
dragon.

593
00:20:31,320 --> 00:20:32,920
And and the party members are running from

594
00:20:32,920 --> 00:20:34,279
the dragon. And somebody says,

595
00:20:34,840 --> 00:20:36,840
what are you doing? You can't get away

596
00:20:36,840 --> 00:20:38,680
from the dragon and and and the other

597
00:20:38,680 --> 00:20:40,119
party members says, I don't have to run

598
00:20:40,119 --> 00:20:41,571
fast than the drink. And I only have

599
00:20:41,571 --> 00:20:43,087
to run faster than you. Right.

600
00:20:43,805 --> 00:20:45,721
Right. It's... And that's what Google is doing

601
00:20:45,721 --> 00:20:47,875
here. They're trying to look better than Facebook.

602
00:20:48,449 --> 00:20:51,082
It's a very mobile. It's a very mobile.

603
00:20:51,401 --> 00:20:51,641
Right.

604
00:20:52,439 --> 00:20:54,673
And the trouble should not accept it. Right.

605
00:20:54,992 --> 00:20:55,471
Right. Right.

606
00:20:56,523 --> 00:20:58,834
Well, okay. So I wanna talk about, I

607
00:20:58,834 --> 00:21:00,986
wanna talk about privacy and enhancing technologies today

608
00:21:00,986 --> 00:21:04,029
and and maybe why how we please. You

609
00:21:04,029 --> 00:21:05,470
could use some this data for the for

610
00:21:05,470 --> 00:21:07,630
good enough for Evil. But and I use

611
00:21:07,630 --> 00:21:10,190
that term specifically in relation to Google since

612
00:21:10,190 --> 00:21:12,199
they used to be the ones who carried

613
00:21:12,199 --> 00:21:14,585
that banner, But... Okay. So I don't know

614
00:21:14,585 --> 00:21:15,938
how far this goes back and if this

615
00:21:15,938 --> 00:21:17,609
is something you actually saw why you were

616
00:21:17,609 --> 00:21:18,347
there, but

617
00:21:18,737 --> 00:21:20,644
they have this thing they're calling privacy sandbox.

618
00:21:20,882 --> 00:21:23,265
And they have... The the death of the

619
00:21:23,265 --> 00:21:25,354
third party cookie, that bell has been

620
00:21:25,903 --> 00:21:27,812
wrong many times or many years. They can't...

621
00:21:27,971 --> 00:21:29,800
They just push it back again. But Google's

622
00:21:29,959 --> 00:21:30,993
Had copy. They make the money off of

623
00:21:30,993 --> 00:21:31,630
this. And so,

624
00:21:32,840 --> 00:21:33,960
I guess, maybe, I don't know if it

625
00:21:33,960 --> 00:21:35,400
was Gdpr the push of this or if

626
00:21:35,400 --> 00:21:36,600
they just the writing on the wall or

627
00:21:36,600 --> 00:21:38,920
maybe because every other browser is automatically blocking

628
00:21:38,920 --> 00:21:40,680
third party cookies or maybe all the above.

629
00:21:41,080 --> 00:21:43,247
But they have come up with. This these

630
00:21:43,247 --> 00:21:44,778
technologies that they claim

631
00:21:45,150 --> 00:21:47,926
is much more privacy preserving. And honestly, on

632
00:21:47,926 --> 00:21:48,323
paper,

633
00:21:48,878 --> 00:21:50,330
it kinda looks better

634
00:21:50,719 --> 00:21:53,192
So I'm I'm curious. What did you see

635
00:21:53,192 --> 00:21:56,222
around privacy sandbox in this topic Api, which

636
00:21:56,222 --> 00:21:58,296
was flock before that or flock before that.

637
00:21:58,615 --> 00:22:00,505
What what's your take on this

638
00:22:01,103 --> 00:22:04,134
technology. And at least in theory, could this

639
00:22:04,134 --> 00:22:06,687
be more privacy preserving than the behavioral tracking

640
00:22:06,687 --> 00:22:07,405
stuff that we do now?

641
00:22:08,460 --> 00:22:11,740
So I was involved in early strategy discussions

642
00:22:11,740 --> 00:22:13,440
around what became privacy sandbox.

643
00:22:14,139 --> 00:22:15,899
At the time I left Google, I had

644
00:22:15,899 --> 00:22:17,594
material criticisms systems of it which I could

645
00:22:17,594 --> 00:22:20,390
not share with the public. Fortunately, everything from

646
00:22:20,390 --> 00:22:22,468
that time is now public information, and I

647
00:22:22,468 --> 00:22:24,785
can talk about it. I haven't spoken out

648
00:22:24,785 --> 00:22:27,902
about because actually Apple's privacy people jumped on

649
00:22:27,902 --> 00:22:29,810
that and said everything I would have said

650
00:22:29,810 --> 00:22:29,969
anyway.

651
00:22:31,161 --> 00:22:32,911
Good on them. Right? They're playing this Vr

652
00:22:32,911 --> 00:22:35,074
game also. They want to be sure. They

653
00:22:35,074 --> 00:22:36,670
want to run faster than Google. Right. Right.

654
00:22:36,830 --> 00:22:39,784
Right. Right. But at least when these, you

655
00:22:39,784 --> 00:22:42,578
know, large scale power structures conflict with each

656
00:22:42,578 --> 00:22:44,175
other, sometimes the public benefits,

657
00:22:45,145 --> 00:22:47,770
Right. So my perspective is,

658
00:22:50,633 --> 00:22:50,872
look.

659
00:22:51,603 --> 00:22:53,352
As a member of a marginalized group.

660
00:22:54,227 --> 00:22:55,817
I am a recent immigrant from the United

661
00:22:55,976 --> 00:22:56,874
States to Canada

662
00:22:57,408 --> 00:23:00,270
because the United States is becoming increasingly hostile

663
00:23:00,270 --> 00:23:01,066
to trans people.

664
00:23:01,637 --> 00:23:03,226
And I don't wish to be on the

665
00:23:03,226 --> 00:23:05,769
receiving end of whatever happens there. Mh. And

666
00:23:05,928 --> 00:23:07,302
I I have the wherewithal

667
00:23:07,835 --> 00:23:10,236
to do that. Very few people do. Right.

668
00:23:10,633 --> 00:23:12,304
Nowhere in the world will be far enough

669
00:23:12,304 --> 00:23:14,213
if America really goes down the dark path,

670
00:23:14,452 --> 00:23:14,952
but

671
00:23:16,217 --> 00:23:17,492
So it's more question of where you stand

672
00:23:17,492 --> 00:23:18,710
in fight, but

673
00:23:19,883 --> 00:23:23,070
we are facing a potential reality where within

674
00:23:23,070 --> 00:23:24,597
5 years, it could be a crime to

675
00:23:24,597 --> 00:23:26,188
be trans in the United States,

676
00:23:26,824 --> 00:23:29,688
which, you know, again, look, I'm old enough

677
00:23:29,688 --> 00:23:32,014
that, like, when I was born, I could

678
00:23:32,014 --> 00:23:35,125
have been institutional for being trans. Right? The

679
00:23:35,125 --> 00:23:36,960
queer community has been here before, and we

680
00:23:36,960 --> 00:23:39,033
will, you know, we will survive it again

681
00:23:39,033 --> 00:23:41,841
this time. We will anything necessary to survive,

682
00:23:42,081 --> 00:23:42,581
but

683
00:23:43,199 --> 00:23:45,194
so I'm not saying like, poor us. But

684
00:23:45,354 --> 00:23:48,547
I am saying, like, it's important that these

685
00:23:48,547 --> 00:23:49,367
online platforms

686
00:23:50,158 --> 00:23:50,658
treat

687
00:23:51,113 --> 00:23:54,716
the demographics of their users with appropriate seriousness

688
00:23:55,011 --> 00:23:57,813
and not leak that information everywhere. And it's

689
00:23:57,813 --> 00:24:00,125
not just queer stuff. You know, They could

690
00:24:00,125 --> 00:24:02,199
they could leak racial information. They could leak

691
00:24:02,199 --> 00:24:04,672
religious information. They they could leak all sorts

692
00:24:04,672 --> 00:24:06,926
of things that people would be subject to

693
00:24:06,985 --> 00:24:09,953
discrimination if 4. Right. Right. And it's really

694
00:24:09,953 --> 00:24:12,667
important that they not do that. And the

695
00:24:12,667 --> 00:24:14,285
problem with the privacy sandbox

696
00:24:14,663 --> 00:24:16,121
thing is right

697
00:24:16,513 --> 00:24:18,421
you know, these technologies have changed over the

698
00:24:18,421 --> 00:24:20,567
years, but I believe it's still fundamentally based

699
00:24:20,567 --> 00:24:23,611
on this idea that it does client side

700
00:24:23,906 --> 00:24:26,867
computation to classify users into about 10000 buckets,

701
00:24:27,505 --> 00:24:30,298
possibly more possibly less, the number may be

702
00:24:30,298 --> 00:24:32,692
debatable. The number I'm sure it'll change. And

703
00:24:32,852 --> 00:24:34,939
I'm sure the company is would be happy

704
00:24:34,939 --> 00:24:36,928
to change the number because they're like, oh,

705
00:24:37,008 --> 00:24:38,360
we did the math, and this is better

706
00:24:38,360 --> 00:24:41,305
now. We addressed your concern. But the point

707
00:24:41,305 --> 00:24:41,805
is

708
00:24:42,498 --> 00:24:44,347
it's enough. Right? Like,

709
00:24:45,704 --> 00:24:49,376
all these marginalized groups are, like, several percentage

710
00:24:49,376 --> 00:24:50,674
points of society

711
00:24:51,052 --> 00:24:52,090
are from this background.

712
00:24:52,582 --> 00:24:55,278
And so any reasonable number that's useful for

713
00:24:55,278 --> 00:24:56,175
advertising purposes

714
00:24:56,706 --> 00:24:58,926
is also going to be useful for identifying

715
00:24:58,926 --> 00:24:59,719
marginalized groups.

716
00:25:00,369 --> 00:25:02,116
And because it's a black box machine learning

717
00:25:02,116 --> 00:25:03,388
system, you can't even know.

718
00:25:03,944 --> 00:25:06,565
Did it do that. It probably did. It

719
00:25:06,565 --> 00:25:07,065
probably

720
00:25:07,931 --> 00:25:08,431
discovered

721
00:25:09,124 --> 00:25:12,407
that people who buy rainbow sneakers are

722
00:25:12,781 --> 00:25:13,599
more likely

723
00:25:14,372 --> 00:25:17,892
to, you know, also by other products. Right?

724
00:25:18,131 --> 00:25:20,943
Right. They probably discovered that proxy for this

725
00:25:21,160 --> 00:25:23,393
because it's told to do that. It's told

726
00:25:23,393 --> 00:25:26,219
maximize profit and it would be illegal to

727
00:25:26,755 --> 00:25:28,132
use some of these categories

728
00:25:28,588 --> 00:25:29,863
for targeted advertising,

729
00:25:30,580 --> 00:25:31,935
especially I believe employment status.

730
00:25:32,668 --> 00:25:35,062
For targeted advertising, if you knew that's that's

731
00:25:35,062 --> 00:25:37,297
what it's doing. But because it's an una

732
00:25:37,297 --> 00:25:39,770
unacceptable machine learning system it wander that bias.

733
00:25:40,343 --> 00:25:41,619
Right. And that's... And I And I think

734
00:25:41,619 --> 00:25:42,575
that's where a lot of this... A lot

735
00:25:42,575 --> 00:25:43,235
of people

736
00:25:43,691 --> 00:25:44,967
missed the boat on this is is that

737
00:25:44,967 --> 00:25:46,083
a lot of these things. And I think

738
00:25:46,083 --> 00:25:47,597
because I think even if when I remember

739
00:25:47,597 --> 00:25:50,315
reading about topic Api. Like, they specifically, I

740
00:25:50,315 --> 00:25:52,482
think said that they were not going to

741
00:25:52,934 --> 00:25:54,046
allow or whatever.

742
00:25:54,522 --> 00:25:57,159
Categories that would be, you know, the sexuality

743
00:25:57,159 --> 00:25:58,829
and and and and religion and then we're

744
00:25:58,829 --> 00:26:00,182
gonna kick those up, But you're right in

745
00:26:00,182 --> 00:26:02,966
that. If you've got enough data, you could

746
00:26:02,966 --> 00:26:03,879
find proxy

747
00:26:04,337 --> 00:26:05,694
information that could get you right back to

748
00:26:05,694 --> 00:26:07,371
a lot of that same stuff. It's it's

749
00:26:07,371 --> 00:26:10,245
like location information. If you have 3 locations

750
00:26:10,245 --> 00:26:11,682
spots for somebody, you could pretty much tell

751
00:26:11,682 --> 00:26:13,689
who they are because you know, where they

752
00:26:13,689 --> 00:26:15,438
are from 9 to 5 Monday to Friday

753
00:26:15,438 --> 00:26:17,187
is probably where they work, where they are

754
00:26:17,187 --> 00:26:18,460
every night is probably where they live and

755
00:26:18,460 --> 00:26:19,493
where they go on Sunday is where they

756
00:26:19,493 --> 00:26:21,900
worship, and if that that'll nail down a

757
00:26:21,900 --> 00:26:23,500
lot of people just with those 3 data

758
00:26:23,500 --> 00:26:23,900
points. Right?

759
00:26:24,700 --> 00:26:24,940
Yeah.

760
00:26:25,660 --> 00:26:26,859
Identify people or even though none of that

761
00:26:26,859 --> 00:26:28,859
it had identifying information, and there's just a

762
00:26:28,859 --> 00:26:30,470
lot correlation that we've done, especially now in

763
00:26:30,470 --> 00:26:31,369
the age of Ai

764
00:26:31,906 --> 00:26:33,662
with, you know, machine learning models. So, you

765
00:26:33,662 --> 00:26:35,656
know, we're getting really good at finding patterns.

766
00:26:36,388 --> 00:26:38,797
And so, yeah. I I I'd I certainly

767
00:26:39,252 --> 00:26:40,922
understand that that is a a limitation and

768
00:26:41,002 --> 00:26:43,904
I... How this data, even though it doesn't

769
00:26:43,962 --> 00:26:44,462
directly

770
00:26:45,079 --> 00:26:47,552
capture some of these more sensitive bits of

771
00:26:47,552 --> 00:26:49,627
information. Can... You could still derive it. Right?

772
00:26:50,757 --> 00:26:52,900
Exactly. And we should be assuming because the

773
00:26:52,900 --> 00:26:56,074
model is given data that definitely contains this,

774
00:26:56,630 --> 00:26:57,662
and it's given a goal.

775
00:26:58,233 --> 00:27:00,800
That depends on this, and it is complicated

776
00:27:00,857 --> 00:27:02,368
enough to find these patterns. We know that

777
00:27:02,368 --> 00:27:04,515
the technology is good at finding patterns.

778
00:27:04,992 --> 00:27:06,185
That is 1 of the few things that

779
00:27:06,185 --> 00:27:08,824
it really does robust at this point. Right.

780
00:27:09,062 --> 00:27:11,050
And that's why it's useful for making money.

781
00:27:11,845 --> 00:27:13,992
And we should be assuming that it is

782
00:27:13,992 --> 00:27:15,287
probably finding

783
00:27:15,741 --> 00:27:17,275
these demographic categories

784
00:27:18,063 --> 00:27:19,977
in the absence of a mechanism to prevent

785
00:27:19,977 --> 00:27:21,652
it from doing that, and I am not

786
00:27:21,652 --> 00:27:22,870
aware of any such mechanism.

787
00:27:24,843 --> 00:27:26,039
Okay. Let let's see if we could do

788
00:27:26,039 --> 00:27:27,576
a complete 1 80

789
00:27:28,367 --> 00:27:30,519
Yeah. On data collection. Let's see if we

790
00:27:30,519 --> 00:27:32,511
can somehow find this silver lining and all

791
00:27:32,511 --> 00:27:33,330
this. So

792
00:27:33,946 --> 00:27:35,482
Alright. So we we are surrounded

793
00:27:35,874 --> 00:27:38,368
by smart devices that collect a crap ton

794
00:27:38,745 --> 00:27:41,696
of highly personal information, either from built in

795
00:27:41,696 --> 00:27:42,972
sensors, which we may or may not be

796
00:27:42,972 --> 00:27:44,567
aware of or just by asking us foot

797
00:27:44,567 --> 00:27:46,338
for us, like the know, signing up for

798
00:27:46,338 --> 00:27:47,857
apps or signing up for accounts. A lot

799
00:27:47,857 --> 00:27:49,215
of times they I ask us for all

800
00:27:49,215 --> 00:27:51,453
sorts of information that we give up because

801
00:27:51,453 --> 00:27:52,651
either we think we need to or because

802
00:27:52,651 --> 00:27:54,584
they they require us to. And, you know,

803
00:27:54,824 --> 00:27:57,301
we in the privacy space talk endlessly about

804
00:27:57,301 --> 00:27:59,778
how revealing this information can be and and

805
00:27:59,778 --> 00:28:01,617
how this could be used against us. So...

806
00:28:01,777 --> 00:28:03,555
But is there some way

807
00:28:04,422 --> 00:28:06,325
that we could process all this data or

808
00:28:06,325 --> 00:28:08,466
some subset of this data privately so that

809
00:28:08,466 --> 00:28:09,577
we could do good things with the 2.

810
00:28:09,735 --> 00:28:11,266
You know, how might our rampant

811
00:28:11,894 --> 00:28:15,091
data collection actually benefits society. And I will

812
00:28:15,091 --> 00:28:16,630
say it as an engineer as a Devops

813
00:28:16,689 --> 00:28:19,327
engineer in particular, at Cisco. We had devices

814
00:28:19,327 --> 00:28:21,345
all over the planet, and we use data

815
00:28:21,819 --> 00:28:23,259
to find out how they're doing it and

816
00:28:23,259 --> 00:28:25,339
actually, we... 1 of my models was is

817
00:28:25,500 --> 00:28:27,259
I wanna find problems before our customers do.

818
00:28:27,659 --> 00:28:29,555
Because I have that capability with all this

819
00:28:29,674 --> 00:28:31,897
you know, tele and telemetry and analytics and

820
00:28:31,897 --> 00:28:33,645
things that are coming in. So I, you

821
00:28:33,645 --> 00:28:34,757
know, for me, I was for a. And

822
00:28:34,757 --> 00:28:36,107
of course, it's not quite the same privacy

823
00:28:36,107 --> 00:28:37,933
things when we're selling things to the companies

824
00:28:37,933 --> 00:28:39,700
it's a different ballgame game. But that is

825
00:28:39,700 --> 00:28:41,140
an instance for me As certainly an engineer.

826
00:28:41,380 --> 00:28:43,539
I understand how I'd want a lot of

827
00:28:43,539 --> 00:28:44,660
data. And I could do some good things

828
00:28:44,660 --> 00:28:45,140
for that data.

829
00:28:45,940 --> 00:28:47,714
So can we... What what are some string

830
00:28:47,714 --> 00:28:49,414
ways where we might be able to

831
00:28:50,115 --> 00:28:52,195
do some good with all this massive amounts

832
00:28:52,195 --> 00:28:53,154
of data that we're collecting.

833
00:28:54,195 --> 00:28:56,683
Yeah. So you raise an interesting example with

834
00:28:56,683 --> 00:28:58,518
a networking data. And I do believe that's

835
00:28:58,518 --> 00:29:00,432
important although though it verge on, like,

836
00:29:01,309 --> 00:29:02,505
the thing where I think we need to

837
00:29:02,505 --> 00:29:04,910
be very careful as data that's about speech.

838
00:29:05,545 --> 00:29:07,632
Right, because we're in an age of increasing

839
00:29:07,925 --> 00:29:10,623
censorship, and I would view, like network traffic

840
00:29:10,623 --> 00:29:13,261
as something to be very cautious about because

841
00:29:13,261 --> 00:29:15,489
even the metadata who's talking to who is

842
00:29:15,489 --> 00:29:16,365
highly sensitive.

843
00:29:16,922 --> 00:29:17,160
Right?

844
00:29:18,513 --> 00:29:20,025
By the way, I have the greatest respect

845
00:29:20,025 --> 00:29:21,550
for the people who do large scale networking

846
00:29:21,550 --> 00:29:24,166
working. That seems like incredibly difficult work due

847
00:29:24,166 --> 00:29:24,983
to the many

848
00:29:25,434 --> 00:29:26,307
organizations involved.

849
00:29:27,258 --> 00:29:29,816
Certainly, you know, it's the same promise that

850
00:29:29,816 --> 00:29:32,207
science has always offered if we understand our

851
00:29:32,207 --> 00:29:34,757
world if we understand ourselves, we have the

852
00:29:34,757 --> 00:29:36,431
power to make change. Right?

853
00:29:37,083 --> 00:29:38,297
It's fundamentally.

854
00:29:38,671 --> 00:29:39,330
I think

855
00:29:40,498 --> 00:29:41,871
there there's a lot of

856
00:29:42,325 --> 00:29:43,754
there's a lot of ways it can do

857
00:29:43,754 --> 00:29:46,715
good. We saw this interest example early in

858
00:29:46,715 --> 00:29:47,932
the Covid pandemic

859
00:29:48,390 --> 00:29:48,890
with

860
00:29:49,506 --> 00:29:51,181
actually a number of my friends were involved

861
00:29:51,181 --> 00:29:52,239
in designing this

862
00:29:52,856 --> 00:29:56,455
specification that was going to collect contact tracing

863
00:29:56,455 --> 00:29:58,680
data. Yeah. Google App actually worked together a

864
00:29:58,680 --> 00:30:00,190
little bit on that. Didn't think. They did.

865
00:30:00,508 --> 00:30:02,097
They did. It was a, you know, a

866
00:30:02,097 --> 00:30:03,460
number of people

867
00:30:03,856 --> 00:30:06,471
who share my background and and skill set

868
00:30:06,471 --> 00:30:06,867
effectively,

869
00:30:07,501 --> 00:30:09,402
really got together and worked hard,

870
00:30:09,893 --> 00:30:11,954
to design this protocol because they saw the

871
00:30:11,954 --> 00:30:14,254
danger that if they don't do this, governments

872
00:30:14,254 --> 00:30:15,840
are going to do this, and the thing

873
00:30:15,840 --> 00:30:17,188
that governments are going to do,

874
00:30:17,839 --> 00:30:19,599
as much as I love that governments can

875
00:30:19,599 --> 00:30:20,960
hold corporations accountable,

876
00:30:21,279 --> 00:30:23,839
they are also large power structures, which are

877
00:30:23,839 --> 00:30:26,567
perhaps a little bit full of themselves sometimes

878
00:30:26,567 --> 00:30:28,421
and assume that they can do no wrong

879
00:30:28,797 --> 00:30:30,969
when they actually can hurt people

880
00:30:31,345 --> 00:30:33,710
even just through collecting data because it's data

881
00:30:33,987 --> 00:30:36,918
about every movement people make. So they designed

882
00:30:36,918 --> 00:30:39,137
a protocol, which is kind of double blind,

883
00:30:39,375 --> 00:30:40,667
which neither you

884
00:30:41,451 --> 00:30:44,575
nor the server nor anybody you meet

885
00:30:44,949 --> 00:30:46,562
knows that you have done that

886
00:30:47,175 --> 00:30:49,821
until this kind of mathematical operation happens,

887
00:30:50,688 --> 00:30:54,053
which combines all this data, and which only

888
00:30:54,269 --> 00:30:54,769
succeeds

889
00:30:55,542 --> 00:30:57,634
when all the parts are there

890
00:30:58,104 --> 00:30:59,463
which means that you were in the same

891
00:30:59,463 --> 00:31:01,540
place at the same time as somebody else.

892
00:31:02,019 --> 00:31:03,697
And then the 2 of you are notified

893
00:31:03,697 --> 00:31:04,736
that you were in the same place at

894
00:31:04,736 --> 00:31:05,455
the same time,

895
00:31:06,029 --> 00:31:07,545
so you can be like, hey, I had

896
00:31:07,704 --> 00:31:09,618
Covid. Sorry. You have Covid.

897
00:31:11,054 --> 00:31:13,128
Nice idea that peep that people care about

898
00:31:13,128 --> 00:31:15,281
each other enough to prevent them from getting

899
00:31:15,281 --> 00:31:15,781
deadly

900
00:31:16,174 --> 00:31:16,674
illnesses.

901
00:31:17,133 --> 00:31:17,373
Right.

902
00:31:18,492 --> 00:31:20,010
Right. I thought I would have liked to

903
00:31:20,010 --> 00:31:20,650
live in that world.

904
00:31:22,807 --> 00:31:24,566
Yeah. That was a fascinating aid case study,

905
00:31:24,725 --> 00:31:27,220
and I the fact that the Google Apple

906
00:31:27,220 --> 00:31:28,819
worked together on that. I thought was actually

907
00:31:28,819 --> 00:31:30,819
kind of, I don't know, heartwarming it away.

908
00:31:31,460 --> 00:31:33,059
It was. It was inspiring and I want

909
00:31:33,059 --> 00:31:35,892
to emphasize it was it was upper management

910
00:31:35,950 --> 00:31:37,543
took P our credit for it, but this

911
00:31:37,543 --> 00:31:39,534
was the hard work of the the individual

912
00:31:39,534 --> 00:31:40,034
contributors

913
00:31:40,570 --> 00:31:42,641
who... The people who are fighting for what

914
00:31:42,641 --> 00:31:43,756
they believe to be right.

915
00:31:44,409 --> 00:31:46,484
And who see moral dangers here that they

916
00:31:46,484 --> 00:31:49,038
wish to steer the ship away from. They

917
00:31:49,038 --> 00:31:49,857
dropped everything

918
00:31:50,235 --> 00:31:52,390
and worked on this. They pulled everybody in.

919
00:31:52,804 --> 00:31:54,554
Was this driven by the engineers and the

920
00:31:54,554 --> 00:31:56,462
and the the employees as opposed to management

921
00:31:56,462 --> 00:31:58,212
where did this come from? Well, I was

922
00:31:58,212 --> 00:31:59,803
outside at that time. I was doing some

923
00:31:59,803 --> 00:32:01,808
self searching in the period but when I

924
00:32:01,808 --> 00:32:03,404
left Google and when I joined the nonprofit

925
00:32:03,404 --> 00:32:05,798
world. So I don't know who initiated the

926
00:32:05,798 --> 00:32:07,314
effort, but I can say there is no

927
00:32:07,314 --> 00:32:09,091
way it could possibly have happened

928
00:32:09,484 --> 00:32:11,159
without the engineers viewing it as a moral

929
00:32:11,159 --> 00:32:11,478
duty.

930
00:32:12,036 --> 00:32:13,392
Well, there's some other interesting things I've heard.

931
00:32:13,551 --> 00:32:14,907
You know, what would people talk about,

932
00:32:15,625 --> 00:32:17,380
all of our cell phones, for example, have

933
00:32:17,380 --> 00:32:20,419
acc in them. And they Somebody was talking

934
00:32:20,419 --> 00:32:22,799
about, what if we used everybody's cell phone

935
00:32:22,799 --> 00:32:24,466
to get us tremors,

936
00:32:24,942 --> 00:32:27,163
you know, small tremors that with... If you

937
00:32:27,163 --> 00:32:29,573
had a million of these devices in an

938
00:32:29,573 --> 00:32:32,529
area, you could maybe locate an impending earthquake,

939
00:32:32,769 --> 00:32:35,007
or they were doing... You mentioned Covid. I

940
00:32:35,007 --> 00:32:37,245
know that they were doing wastewater sampling at

941
00:32:37,245 --> 00:32:39,098
the ways the water treatment plants they could

942
00:32:39,098 --> 00:32:41,116
kinda see at a bulk level and a

943
00:32:41,655 --> 00:32:43,493
what kind of viral load was in the

944
00:32:43,493 --> 00:32:44,852
community and a lot of times predict when

945
00:32:44,852 --> 00:32:45,811
they were gonna be outbreaks.

946
00:32:46,530 --> 00:32:47,030
Drug.

947
00:32:47,663 --> 00:32:50,447
Yeah. Yeah all... Yeah. Drug interactions are a

948
00:32:50,447 --> 00:32:52,617
great 1 because that's a very data centric

949
00:32:52,674 --> 00:32:53,174
because

950
00:32:53,628 --> 00:32:55,776
most patients who have a complex medical situation

951
00:32:55,776 --> 00:32:57,460
are dozens of medications.

952
00:32:58,096 --> 00:32:59,787
And so it's very hard to dis

953
00:33:00,718 --> 00:33:03,101
what caused to this interaction. You know there

954
00:33:03,101 --> 00:33:04,825
was an interaction, but you don't know which

955
00:33:04,944 --> 00:33:07,744
pair of medications caused it. And so, yes.

956
00:33:08,464 --> 00:33:09,764
Aggregated data

957
00:33:10,065 --> 00:33:11,345
could do a great job there, But at

958
00:33:11,345 --> 00:33:12,464
the same time,

959
00:33:13,119 --> 00:33:15,615
the mist of medications that somebody is on

960
00:33:15,833 --> 00:33:18,388
is probably unique to them. Mh. And so

961
00:33:18,388 --> 00:33:19,984
what you do about that. I think it's

962
00:33:19,984 --> 00:33:22,161
a complicated question. I actually

963
00:33:22,714 --> 00:33:24,630
I believe that health care privacy is an

964
00:33:24,630 --> 00:33:26,088
area where the law needs

965
00:33:26,546 --> 00:33:28,063
further advancement. Everybody,

966
00:33:28,941 --> 00:33:30,617
I like to say the p and Hipaa

967
00:33:30,617 --> 00:33:33,426
stands for portability not privacy. Which it does

968
00:33:33,426 --> 00:33:35,422
by the way. Yes. It sure does.

969
00:33:36,060 --> 00:33:38,934
People incorrectly assume that because there's a law

970
00:33:38,934 --> 00:33:41,224
about this and because medical providers do take

971
00:33:41,423 --> 00:33:42,161
law seriously

972
00:33:42,614 --> 00:33:44,202
that the law must say the things they

973
00:33:44,202 --> 00:33:47,061
wanted to say that our modern expectations and

974
00:33:47,061 --> 00:33:48,990
fears about how this could be misused

975
00:33:49,616 --> 00:33:52,233
are actually legally binding. They're not. I I

976
00:33:52,233 --> 00:33:53,264
would like to break with the fact that

977
00:33:53,264 --> 00:33:55,088
the, we... You know, we say that there's

978
00:33:55,088 --> 00:33:56,778
a right to privacy and shrine to the

979
00:33:56,833 --> 00:33:59,888
constitution the word privacy does not appear once

980
00:34:00,023 --> 00:34:01,691
in the constitution or other the bill rights.

981
00:34:02,882 --> 00:34:04,708
That is correct. Now I am not a

982
00:34:04,708 --> 00:34:05,208
lawyer

983
00:34:05,518 --> 00:34:06,950
like to bring up the rhetoric point that

984
00:34:07,109 --> 00:34:09,735
Ro Wade was decided as a privacy argument.

985
00:34:10,054 --> 00:34:12,521
Yeah. It is nobody's business what is happening

986
00:34:12,521 --> 00:34:14,565
to your body. And I think that's

987
00:34:15,001 --> 00:34:17,166
still morally the correct argument.

988
00:34:18,015 --> 00:34:20,871
The word privacy often gets used as, like,

989
00:34:21,839 --> 00:34:24,393
a catch all term for. I have some

990
00:34:24,393 --> 00:34:26,549
reason not to tell you this, and it's

991
00:34:26,549 --> 00:34:27,986
none of your business what that reason is.

992
00:34:28,146 --> 00:34:29,662
I don't have to justify to it to

993
00:34:29,662 --> 00:34:29,822
you.

994
00:34:31,114 --> 00:34:32,709
Right Yeah. And I I love the students

995
00:34:32,709 --> 00:34:34,304
quota. This. And I I bring in a

996
00:34:34,304 --> 00:34:36,138
public too often. But, you, you know, you

997
00:34:36,138 --> 00:34:37,175
quoted say I think this is not kind

998
00:34:37,175 --> 00:34:39,328
a Reddit Ama where you said, arguing that

999
00:34:39,328 --> 00:34:41,253
you don't care about right to privacy because

1000
00:34:41,253 --> 00:34:43,421
you have nothing to hide, is no different

1001
00:34:43,477 --> 00:34:44,827
than saying you don't care about free speech

1002
00:34:44,827 --> 00:34:46,337
because you have nothing to say. It's a

1003
00:34:46,337 --> 00:34:47,687
right. The whole point with rights is you

1004
00:34:47,687 --> 00:34:48,775
don't have to fight for them. They just

1005
00:34:48,894 --> 00:34:50,723
are. You you don't have to justify them.

1006
00:34:50,882 --> 00:34:51,360
You have them.

1007
00:34:52,076 --> 00:34:53,905
And and I... And that's different than a

1008
00:34:53,905 --> 00:34:55,019
lot of other area of law.

1009
00:34:55,909 --> 00:34:57,577
And I will say as somebody in my

1010
00:34:57,577 --> 00:34:59,563
professional opinion of somebody who spent every day

1011
00:34:59,563 --> 00:35:00,278
for 5 years,

1012
00:35:01,390 --> 00:35:03,297
identifying the ways that that that that software

1013
00:35:03,297 --> 00:35:04,194
can hurt people

1014
00:35:04,744 --> 00:35:06,905
everybody has something worth hiding. It's just most

1015
00:35:06,905 --> 00:35:08,664
don't know what it is until too late.

1016
00:35:09,704 --> 00:35:09,944
Yes.

1017
00:35:11,065 --> 00:35:13,079
Alright. So given that we care about privacy,

1018
00:35:13,400 --> 00:35:14,940
and let's talk about some of these privacy

1019
00:35:15,000 --> 00:35:15,719
preserving technologies.

1020
00:35:16,679 --> 00:35:19,333
There are actually several efforts underway. To do

1021
00:35:19,333 --> 00:35:20,843
this, which is really kinda really cool. And

1022
00:35:20,843 --> 00:35:21,797
I and I wanna make sure that people

1023
00:35:21,797 --> 00:35:22,751
are aware of these, because I don't think

1024
00:35:22,751 --> 00:35:23,546
it gets enough attention.

1025
00:35:24,420 --> 00:35:25,851
1 of them is call... And their call,

1026
00:35:26,010 --> 00:35:28,157
again, we're recalling them privacy enhanced technologies. It's

1027
00:35:28,157 --> 00:35:29,607
actually the term of art. It it's a

1028
00:35:29,687 --> 00:35:31,762
Pet or a pet. The pet sounds weird.

1029
00:35:32,001 --> 00:35:32,639
It's a Pet.

1030
00:35:33,358 --> 00:35:35,752
And 1 of the ones is called differential

1031
00:35:35,752 --> 00:35:36,231
privacy.

1032
00:35:36,804 --> 00:35:38,474
Let's start with that 1 Can you maybe

1033
00:35:38,474 --> 00:35:40,384
explain what that is in in simple terms?

1034
00:35:40,623 --> 00:35:43,169
And then, know, how does it work to

1035
00:35:43,169 --> 00:35:43,669
balance

1036
00:35:44,536 --> 00:35:46,786
the utility of all this data while still

1037
00:35:46,843 --> 00:35:47,639
preserving our privacy.

1038
00:35:48,354 --> 00:35:50,025
So that's a really interesting 1.

1039
00:35:51,471 --> 00:35:54,511
Differential privacy is for situations where you are

1040
00:35:54,805 --> 00:35:57,901
publishing some data, think like the Census records

1041
00:35:57,901 --> 00:35:58,457
or something,

1042
00:35:58,869 --> 00:36:00,630
but you want to protect the people who

1043
00:36:00,630 --> 00:36:02,869
are in the dataset set. So you want

1044
00:36:02,869 --> 00:36:03,530
to say

1045
00:36:03,989 --> 00:36:07,030
there are, like, 10000 people living in this

1046
00:36:07,429 --> 00:36:08,329
in this suburb

1047
00:36:08,644 --> 00:36:10,157
but you don't want anybody to be able

1048
00:36:10,157 --> 00:36:12,069
to identify who a specific person is.

1049
00:36:13,025 --> 00:36:14,162
And that is challenging

1050
00:36:14,777 --> 00:36:17,587
because there are sophisticated attacks these days

1051
00:36:18,058 --> 00:36:20,052
which can take. You'd think that, like, rounding

1052
00:36:20,052 --> 00:36:22,045
it off the nearest thousand is safe, but

1053
00:36:22,045 --> 00:36:22,524
it's not.

1054
00:36:23,242 --> 00:36:25,155
You can compare different versions of the same

1055
00:36:25,155 --> 00:36:26,772
dataset set, and you can

1056
00:36:27,721 --> 00:36:30,668
identify individuals track their movements across time Or

1057
00:36:30,668 --> 00:36:32,340
cross correlate with a similar set of data?

1058
00:36:32,897 --> 00:36:34,411
Or Cross Car weight with a similar set

1059
00:36:34,411 --> 00:36:35,366
of data exactly.

1060
00:36:36,019 --> 00:36:38,816
And so differential privacy gives you a mathematical

1061
00:36:38,816 --> 00:36:39,956
guarantee about

1062
00:36:40,414 --> 00:36:42,972
the extent of such attacks. You can never

1063
00:36:42,972 --> 00:36:44,810
prevent all such attacks. The only way to

1064
00:36:44,810 --> 00:36:46,903
do that is did not release the data,

1065
00:36:47,222 --> 00:36:49,619
but differential privacy kind of blur the data.

1066
00:36:49,859 --> 00:36:52,257
It adds noise to it in a very

1067
00:36:52,257 --> 00:36:52,757
calculated

1068
00:36:53,136 --> 00:36:54,275
way. It's a very

1069
00:36:54,908 --> 00:36:55,408
specific

1070
00:36:55,786 --> 00:36:58,261
type of noise that follows, I believe a

1071
00:36:58,261 --> 00:36:58,979
logistic pattern,

1072
00:36:59,697 --> 00:37:00,597
such that

1073
00:37:00,974 --> 00:37:02,991
these attackers can never gain

1074
00:37:03,384 --> 00:37:05,541
more than a certain amount of information that

1075
00:37:05,541 --> 00:37:07,619
you choose in advance and that you call

1076
00:37:07,619 --> 00:37:10,675
by the mathematical constant Epsilon. You choose your

1077
00:37:11,067 --> 00:37:11,567
privacy,

1078
00:37:12,101 --> 00:37:14,430
Epsilon that you're willing to lose this much

1079
00:37:14,646 --> 00:37:15,146
privacy,

1080
00:37:15,759 --> 00:37:17,532
and then you red your data

1081
00:37:18,065 --> 00:37:18,860
before publishing it.

1082
00:37:19,514 --> 00:37:21,592
They actually did this with a... Sorry. Go

1083
00:37:21,592 --> 00:37:21,752
ahead.

1084
00:37:22,631 --> 00:37:24,309
I was what are the examples I've often

1085
00:37:24,309 --> 00:37:25,827
heard quoted it explains this is like, okay.

1086
00:37:25,987 --> 00:37:27,426
Let's got to these 10 pounds of records

1087
00:37:27,426 --> 00:37:29,118
with 10000 people, and you've got let's say

1088
00:37:29,118 --> 00:37:29,994
you've got all their ages.

1089
00:37:30,791 --> 00:37:32,384
And so age is something that kinda narrows

1090
00:37:32,384 --> 00:37:33,898
it down pretty quickly. Right? We live between,

1091
00:37:33,977 --> 00:37:36,925
you know, 0 and 90 years old ish,

1092
00:37:37,403 --> 00:37:37,801
statistically,

1093
00:37:38,294 --> 00:37:39,731
you know, and if I knew what specific

1094
00:37:39,731 --> 00:37:41,886
age you were on this specific year that

1095
00:37:41,886 --> 00:37:43,563
might, you know, dial me in. But let's

1096
00:37:43,563 --> 00:37:46,037
say, I go through all 10000 records and

1097
00:37:46,197 --> 00:37:46,856
I fuzz

1098
00:37:47,327 --> 00:37:49,232
just a little bit. Everybody's age. Like, you

1099
00:37:49,232 --> 00:37:50,818
know, my age is reported as this is

1100
00:37:50,818 --> 00:37:52,723
not my age. My age report is 35.

1101
00:37:53,119 --> 00:37:54,883
But I'm gonna... I'm gonna say Carrie was

1102
00:37:54,883 --> 00:37:56,080
37 and Irene,

1103
00:37:56,878 --> 00:37:58,154
and I know this is is probably not

1104
00:37:58,154 --> 00:38:00,149
outrage. They are 25, and,

1105
00:38:00,708 --> 00:38:02,316
no. We're gonna say they're 20 they're they're

1106
00:38:02,316 --> 00:38:05,097
23. So they're both off by 2. But

1107
00:38:05,097 --> 00:38:06,925
if you took the average of everybody's ages,

1108
00:38:07,084 --> 00:38:08,276
you could still get the average age of

1109
00:38:08,276 --> 00:38:10,025
everybody in that 10000 thing and use that

1110
00:38:10,025 --> 00:38:12,505
average ages you know, for your data, but

1111
00:38:12,505 --> 00:38:15,686
any individual record was wrong. And therefore been

1112
00:38:15,686 --> 00:38:17,516
been holidays. I've quite flattered by that.

1113
00:38:20,010 --> 00:38:21,050
IIII

1114
00:38:21,050 --> 00:38:22,650
have not been in my twenties for long.

1115
00:38:23,050 --> 00:38:23,769
Long time.

1116
00:38:24,809 --> 00:38:26,890
No. No. Yeah. That's exactly right. It's a

1117
00:38:26,890 --> 00:38:28,920
very well shows an example that you just

1118
00:38:29,137 --> 00:38:30,730
adjusted a couple of years,

1119
00:38:31,128 --> 00:38:33,120
and that makes it harder to do that,

1120
00:38:33,359 --> 00:38:33,598
like,

1121
00:38:34,729 --> 00:38:35,229
cross

1122
00:38:35,606 --> 00:38:38,000
correlation. You can't eliminate this record when you're

1123
00:38:38,000 --> 00:38:38,739
cross car

1124
00:38:39,196 --> 00:38:40,552
because it could be this person or it

1125
00:38:40,552 --> 00:38:41,350
could be that person.

1126
00:38:42,004 --> 00:38:43,440
And therefore, you can't narrow down on other

1127
00:38:43,440 --> 00:38:46,233
people because you couldn't that 1. Right? And

1128
00:38:46,233 --> 00:38:47,509
so you only have to add a small

1129
00:38:47,509 --> 00:38:48,626
amount of noise usually.

1130
00:38:49,599 --> 00:38:50,954
Is there a particular place we've talked about

1131
00:38:50,954 --> 00:38:52,310
a census? And that is actually where I've

1132
00:38:52,310 --> 00:38:53,687
heard this particular technology

1133
00:38:54,463 --> 00:38:56,457
referred to a lot. So to not being

1134
00:38:56,457 --> 00:38:58,530
able to identify individual records within sets data,

1135
00:38:58,690 --> 00:39:00,851
But where else might this be handy? Where

1136
00:39:00,851 --> 00:39:03,315
else where else might this particular... It we're

1137
00:39:03,315 --> 00:39:05,858
where differential privacy be a good technology to

1138
00:39:05,858 --> 00:39:07,858
use where we to allow us to process

1139
00:39:07,858 --> 00:39:09,841
data and get some good, you know, statistical

1140
00:39:09,841 --> 00:39:11,348
data... Information on data without,

1141
00:39:12,379 --> 00:39:13,252
identifying individual people.

1142
00:39:13,984 --> 00:39:16,061
You raised the earthquakes you Scenario and I

1143
00:39:16,061 --> 00:39:17,820
do think that's a very useful 1.

1144
00:39:18,459 --> 00:39:20,137
Now location is a

1145
00:39:20,776 --> 00:39:22,135
very sensitive thing.

1146
00:39:22,868 --> 00:39:25,099
In a sense your location almost is your

1147
00:39:25,099 --> 00:39:25,817
identity. Right?

1148
00:39:26,534 --> 00:39:28,446
There was only 1 person who was where

1149
00:39:28,446 --> 00:39:30,453
you are at this time of day. There's

1150
00:39:30,453 --> 00:39:32,045
only 1 person who goes back and forth

1151
00:39:32,045 --> 00:39:34,273
from your home to your workplace every day.

1152
00:39:34,751 --> 00:39:37,378
Right. So it's difficult to make that private,

1153
00:39:37,537 --> 00:39:40,422
but know, but that's a mathematical parameter that

1154
00:39:40,422 --> 00:39:42,416
you can decide how much you're willing to

1155
00:39:42,416 --> 00:39:43,395
lose and

1156
00:39:44,091 --> 00:39:46,005
put that into the system, and it's very

1157
00:39:46,005 --> 00:39:48,513
likely that you could still get something

1158
00:39:49,123 --> 00:39:51,661
useful for, like, early warnings of earthquakes out

1159
00:39:51,661 --> 00:39:53,405
of it. Now I say they're very likely,

1160
00:39:53,564 --> 00:39:55,760
but I have not done the earthquake math

1161
00:39:55,880 --> 00:39:58,197
here. Right? I'm not a geo. Right. Right

1162
00:39:58,276 --> 00:40:00,753
Right. But from my background, it seems like

1163
00:40:00,753 --> 00:40:01,552
it should be possible.

1164
00:40:02,192 --> 00:40:03,790
You know, health care, like,

1165
00:40:04,684 --> 00:40:08,464
again, tracking, not just not epidemics, but tracking

1166
00:40:08,605 --> 00:40:09,105
emerging

1167
00:40:10,284 --> 00:40:10,784
outbreaks

1168
00:40:11,085 --> 00:40:11,585
right

1169
00:40:11,899 --> 00:40:13,733
is a scenario where could be very useful

1170
00:40:13,733 --> 00:40:15,727
because it can let you know, hey, there's

1171
00:40:15,727 --> 00:40:18,301
a cluster in this region. Maybe,

1172
00:40:18,679 --> 00:40:21,884
you know, look for any big events that

1173
00:40:21,884 --> 00:40:24,036
possibly didn't have adequate hygiene precautions,

1174
00:40:24,514 --> 00:40:27,064
something like that. And it's just a starting

1175
00:40:27,064 --> 00:40:29,376
point for the epidemiologist to do the work.

1176
00:40:29,867 --> 00:40:32,641
That until recently I would have talked up

1177
00:40:32,641 --> 00:40:34,304
at great length, how much work they do

1178
00:40:34,304 --> 00:40:34,938
to protect us.

1179
00:40:36,047 --> 00:40:36,206
Yeah.

1180
00:40:37,251 --> 00:40:39,234
So... Yeah. So it seems like what... Where

1181
00:40:39,234 --> 00:40:41,297
this particular technology really comes into play as

1182
00:40:41,297 --> 00:40:43,518
a situations where we can glean a lot

1183
00:40:43,518 --> 00:40:46,958
of important information from aggregate data and statistics

1184
00:40:46,958 --> 00:40:49,994
about the aggregate data where we don't... Where

1185
00:40:49,994 --> 00:40:52,392
we wanna protect the individual privacy of of

1186
00:40:52,392 --> 00:40:55,043
the all each record of data, but we

1187
00:40:55,043 --> 00:40:56,878
still might need access to those records. I

1188
00:40:56,878 --> 00:40:58,315
don't know. Move do follow up or something.

1189
00:40:58,794 --> 00:41:01,028
So different privacy works is, like, a a

1190
00:41:01,028 --> 00:41:03,115
set of data where you wanna do statistics

1191
00:41:03,115 --> 00:41:05,419
around that set of data. The differential privacy

1192
00:41:05,419 --> 00:41:06,929
is a prop... Is is a transformation you

1193
00:41:06,929 --> 00:41:10,282
make to an entire dataset set. Alright. It's

1194
00:41:10,282 --> 00:41:12,668
it's not suitable for situations where you might

1195
00:41:12,668 --> 00:41:14,440
look at individual records later

1196
00:41:14,894 --> 00:41:17,837
unless you applied in much more sophisticated ways.

1197
00:41:18,489 --> 00:41:20,487
That plan from the start for the fact

1198
00:41:20,487 --> 00:41:22,085
that you're going to be digging into details

1199
00:41:22,085 --> 00:41:24,482
later, and how many details you might dig

1200
00:41:24,482 --> 00:41:26,240
into. You have to set that down on

1201
00:41:26,240 --> 00:41:26,695
yourself

1202
00:41:27,054 --> 00:41:28,652
before you start, and you have to take

1203
00:41:28,652 --> 00:41:29,872
into account every

1204
00:41:30,491 --> 00:41:31,849
inquiry that will be made over the life

1205
00:41:31,849 --> 00:41:34,726
of the system. So it's much more fraught

1206
00:41:34,726 --> 00:41:36,978
and difficult to use it in that kind

1207
00:41:36,978 --> 00:41:37,137
of,

1208
00:41:38,011 --> 00:41:40,556
you're going to dig into this data case

1209
00:41:40,556 --> 00:41:41,295
by case

1210
00:41:41,908 --> 00:41:42,408
scenario.

1211
00:41:42,942 --> 00:41:44,692
I would think that researchers is doing that.

1212
00:41:45,343 --> 00:41:47,335
Or are not well served by differential privacy

1213
00:41:47,335 --> 00:41:50,044
and should be instead doing more to protect

1214
00:41:50,044 --> 00:41:51,579
the way the data is used

1215
00:41:52,036 --> 00:41:53,629
and who has access to it and things

1216
00:41:53,629 --> 00:41:55,873
of that nature. Which you should always do

1217
00:41:55,873 --> 00:41:58,419
anyway, but they should be more restrictive about

1218
00:41:58,419 --> 00:42:00,806
it. And so the next technology I wanted

1219
00:42:00,806 --> 00:42:03,114
to bring up, maybe addresses that issue, and

1220
00:42:03,114 --> 00:42:05,355
and it's 1 that's really bizarre. Almost sounds

1221
00:42:05,355 --> 00:42:06,150
too good to be true.

1222
00:42:07,421 --> 00:42:10,442
It's called homo warfare encryption. My understanding is

1223
00:42:10,442 --> 00:42:12,849
this basically allows you to do

1224
00:42:13,878 --> 00:42:15,734
computations on encrypted data

1225
00:42:16,272 --> 00:42:19,066
without first decrypt that data. This sounds like

1226
00:42:19,066 --> 00:42:20,663
something where I might be able to do

1227
00:42:20,663 --> 00:42:22,992
a process on, let's say, data from someone's

1228
00:42:22,992 --> 00:42:24,826
medical record without actually being able to see

1229
00:42:24,826 --> 00:42:26,681
what's in that record. But

1230
00:42:27,298 --> 00:42:29,052
this... I admit. This is something nice lot

1231
00:42:29,052 --> 00:42:30,807
less sure about. So what... Explain those what

1232
00:42:30,807 --> 00:42:31,046
this is.

1233
00:42:31,940 --> 00:42:32,980
It's a nice thought.

1234
00:42:33,860 --> 00:42:35,940
Now I want to preface this by saying

1235
00:42:36,019 --> 00:42:37,960
I am not a crypto

1236
00:42:38,500 --> 00:42:40,830
expert. I know more even than I think

1237
00:42:40,830 --> 00:42:43,143
most programmers about crypto photography, but I am

1238
00:42:43,143 --> 00:42:43,882
not a crypto

1239
00:42:44,259 --> 00:42:46,652
expert. However, I have worked with people who

1240
00:42:46,652 --> 00:42:46,891
are.

1241
00:42:47,609 --> 00:42:48,747
My impression

1242
00:42:49,459 --> 00:42:51,775
is that it is a long way before

1243
00:42:51,775 --> 00:42:54,172
homo morph encryption can deliver on those promises.

1244
00:42:54,891 --> 00:42:57,140
So it's a nice thought. Right? Like, we

1245
00:42:57,140 --> 00:42:59,133
don't have to trust the device. We we

1246
00:42:59,133 --> 00:43:01,205
can run them somebody else's computer. Like, we

1247
00:43:01,205 --> 00:43:02,959
could run then Google's computer and Google won't

1248
00:43:02,959 --> 00:43:03,596
see what we do.

1249
00:43:04,249 --> 00:43:06,103
That is not the reality of that technology

1250
00:43:06,639 --> 00:43:08,255
today. What happens

1251
00:43:08,631 --> 00:43:11,394
with most of these approaches is that

1252
00:43:11,911 --> 00:43:13,343
it kind of depends on what you mean

1253
00:43:13,343 --> 00:43:13,843
by

1254
00:43:14,217 --> 00:43:16,524
computation. First off. Right? Like, you can never

1255
00:43:16,524 --> 00:43:19,864
do, like, word processing this way. Right. Because

1256
00:43:19,864 --> 00:43:21,079
that has too many

1257
00:43:21,945 --> 00:43:23,873
control points in it, It's making too many

1258
00:43:24,087 --> 00:43:26,492
decisions. If you're putting it through a simple

1259
00:43:26,626 --> 00:43:27,126
mathematical

1260
00:43:27,578 --> 00:43:29,923
transformation, if you are a credit card processor.

1261
00:43:30,620 --> 00:43:31,998
And you want to

1262
00:43:32,376 --> 00:43:34,213
compare your data with a different credit card

1263
00:43:34,213 --> 00:43:34,713
processor,

1264
00:43:35,090 --> 00:43:36,847
but neither of you trust each other for

1265
00:43:36,847 --> 00:43:37,565
business reasons.

1266
00:43:38,137 --> 00:43:39,887
Because these companies are all terrified of each

1267
00:43:39,887 --> 00:43:40,047
other.

1268
00:43:41,240 --> 00:43:41,797
They they are.

1269
00:43:45,153 --> 00:43:47,644
Then then that's a scenario where some fancy

1270
00:43:47,782 --> 00:43:48,759
encryption technologies,

1271
00:43:49,693 --> 00:43:52,481
which might be described as partial morph, so

1272
00:43:52,481 --> 00:43:54,407
that's not how I would put it. Because

1273
00:43:54,407 --> 00:43:57,135
the term homo morph has been sensational,

1274
00:43:57,750 --> 00:43:59,660
and I prefer to avoid it. I would

1275
00:43:59,660 --> 00:44:02,604
have said something like comm crypto systems or

1276
00:44:02,604 --> 00:44:03,104
something

1277
00:44:03,734 --> 00:44:06,445
comp or private set intersection to refer to

1278
00:44:06,445 --> 00:44:08,041
the end goal that you're trying to achieve

1279
00:44:08,041 --> 00:44:10,992
there as the names of those technologies, even

1280
00:44:10,992 --> 00:44:12,529
though it's technically a homo,

1281
00:44:13,066 --> 00:44:15,253
when I say that people start

1282
00:44:16,343 --> 00:44:17,480
dabbling about blockchain.

1283
00:44:20,717 --> 00:44:21,592
Oh, blockchain.

1284
00:44:22,644 --> 00:44:24,482
It's been supplied by Ai in the news

1285
00:44:24,482 --> 00:44:25,600
at least for the most part. We don't

1286
00:44:25,600 --> 00:44:27,278
hear about it quite so much, but about

1287
00:44:27,278 --> 00:44:28,178
1 of my favorite

1288
00:44:28,557 --> 00:44:31,129
engineering flow chart diagrams was You know, you

1289
00:44:31,129 --> 00:44:32,487
you you ask a question. You go to

1290
00:44:32,487 --> 00:44:34,085
the choice point and you got 2 exits

1291
00:44:34,085 --> 00:44:35,603
at yes or no? And the question is,

1292
00:44:35,683 --> 00:44:37,681
do you need home... Do you you need

1293
00:44:37,681 --> 00:44:39,855
blockchain and and both both of them go

1294
00:44:39,855 --> 00:44:40,335
to know.

1295
00:44:41,695 --> 00:44:43,215
We I always I always love that chart

1296
00:44:43,215 --> 00:44:45,695
anyway, but you. III agree with that. At

1297
00:44:45,695 --> 00:44:46,014
any anyway.

1298
00:44:47,787 --> 00:44:49,620
Alright. So what about the other 1? There's

1299
00:44:49,620 --> 00:44:52,091
some other very interesting privacy enhancing technologies out

1300
00:44:52,091 --> 00:44:54,084
there. Let's let's rattle off a few of

1301
00:44:54,084 --> 00:44:55,611
of some of the other more promising What

1302
00:44:55,611 --> 00:44:56,722
what are some other things that are being

1303
00:44:56,722 --> 00:44:58,253
done out there to try to

1304
00:44:58,625 --> 00:45:00,845
allow us to process data without necessarily given

1305
00:45:00,845 --> 00:45:03,957
away people's information, private information? Sure. So I

1306
00:45:03,957 --> 00:45:05,792
touched on the moment ago. There are these

1307
00:45:05,792 --> 00:45:08,982
interesting things secure multi part computation. Mh.

1308
00:45:09,540 --> 00:45:11,774
Is this neat technique where you combine data

1309
00:45:11,774 --> 00:45:12,992
from 2 separate sources

1310
00:45:13,465 --> 00:45:15,625
it's some fancy encryption that I don't think

1311
00:45:15,625 --> 00:45:16,905
it's a good idea to go into.

1312
00:45:17,945 --> 00:45:19,945
I understand I understand this. This is 1

1313
00:45:19,945 --> 00:45:22,184
of those few topics where I can hold

1314
00:45:22,184 --> 00:45:24,355
my own. But I don't think that your

1315
00:45:24,355 --> 00:45:26,670
readers want to hear this math. Yeah. Just

1316
00:45:26,670 --> 00:45:28,026
give just give us the D.

1317
00:45:28,505 --> 00:45:30,601
Yeah. The short version of it is

1318
00:45:30,914 --> 00:45:32,909
I have some information about myself.

1319
00:45:33,547 --> 00:45:36,421
You have some information about yourself. We want

1320
00:45:36,421 --> 00:45:38,416
to find out what what we have in

1321
00:45:38,416 --> 00:45:38,735
common.

1322
00:45:39,469 --> 00:45:39,710
Right?

1323
00:45:40,349 --> 00:45:43,230
We've agreed on some criteria where, like, if

1324
00:45:43,230 --> 00:45:45,869
you and I both have the same favorite

1325
00:45:45,869 --> 00:45:47,802
flavor of ice cream, then we want to

1326
00:45:47,802 --> 00:45:50,035
know about that. Otherwise, I'd prefer not to

1327
00:45:50,035 --> 00:45:51,709
know about your terrible taste and ice cream.

1328
00:45:52,187 --> 00:45:53,702
Right. Sounds a lot like a 0 knowledge

1329
00:45:53,702 --> 00:45:54,818
proof. Is it is it based on the

1330
00:45:54,818 --> 00:45:55,456
similar concept?

1331
00:45:56,584 --> 00:45:58,409
It it is an analogous concept and I

1332
00:45:58,409 --> 00:46:00,814
believe some mechanisms for 0 knowledge proofs

1333
00:46:01,425 --> 00:46:03,409
are along these lines. We can get into

1334
00:46:03,409 --> 00:46:04,202
that in a moment.

1335
00:46:04,773 --> 00:46:06,517
But secure multi party computation,

1336
00:46:07,468 --> 00:46:10,243
we both encrypt our stuff in this fancy

1337
00:46:10,243 --> 00:46:10,560
way,

1338
00:46:11,115 --> 00:46:12,011
and then we

1339
00:46:12,558 --> 00:46:14,308
have a neutral third party,

1340
00:46:15,342 --> 00:46:17,808
combine it in a way where, like, only

1341
00:46:17,808 --> 00:46:19,103
only the third party

1342
00:46:19,494 --> 00:46:21,412
can see whether anything was the same,

1343
00:46:22,051 --> 00:46:24,548
Something to that effect, or there are other

1344
00:46:24,688 --> 00:46:25,188
protocols

1345
00:46:25,567 --> 00:46:27,345
where we don't have a third party

1346
00:46:27,659 --> 00:46:29,248
and 1 of us is doing the combining,

1347
00:46:29,407 --> 00:46:31,495
but we still only see the thing

1348
00:46:32,108 --> 00:46:34,650
that agrees. Okay. Don't. And any of, you,

1349
00:46:34,984 --> 00:46:36,341
It doesn't require us to have a trusted

1350
00:46:36,341 --> 00:46:37,699
at third party. You can work that way,

1351
00:46:37,778 --> 00:46:38,896
but it doesn't have to work that way.

1352
00:46:39,694 --> 00:46:41,211
That's an option that can make the for

1353
00:46:41,211 --> 00:46:43,207
these protocols easier to design. There's a number

1354
00:46:43,207 --> 00:46:44,539
of ways to implement this

1355
00:46:45,619 --> 00:46:47,860
which all have different constraints, different types of

1356
00:46:47,860 --> 00:46:50,500
properties we can set. Okay. Anything else like,

1357
00:46:50,660 --> 00:46:52,260
if we'd we'd briefly touched on 0 dollars

1358
00:46:52,260 --> 00:46:53,380
proofs. So let's go back to that rick.

1359
00:46:54,032 --> 00:46:55,541
I've actually tried to describe this myself before,

1360
00:46:55,621 --> 00:46:56,994
and it's it's tricky, but

1361
00:46:57,607 --> 00:46:58,878
how does that play into this? What how

1362
00:46:58,878 --> 00:46:59,514
is that a Pet?

1363
00:47:00,706 --> 00:47:03,422
0 knowledge proofs are very interested technology. I

1364
00:47:03,422 --> 00:47:06,284
have yet to see an application of it

1365
00:47:06,284 --> 00:47:08,668
that really fires me up for, like, this

1366
00:47:08,668 --> 00:47:10,735
is using its potential well, but it's a

1367
00:47:10,735 --> 00:47:13,138
cool thought. Mean, like 2 billionaires figuring out

1368
00:47:13,138 --> 00:47:14,732
who has more money than the other. That's

1369
00:47:14,732 --> 00:47:17,223
that's the classic example. Right? Right. That is

1370
00:47:17,281 --> 00:47:18,396
terrible useful for, yeah.

1371
00:47:19,210 --> 00:47:21,369
Yeah. Exactly. But it's like, I want to

1372
00:47:21,369 --> 00:47:21,869
know,

1373
00:47:22,170 --> 00:47:23,389
like, if you can

1374
00:47:23,769 --> 00:47:25,849
can afford to show up to this fancy

1375
00:47:25,849 --> 00:47:27,769
party or whatever, but I don't want to

1376
00:47:27,769 --> 00:47:28,510
know anything

1377
00:47:28,824 --> 00:47:30,976
other than did somebody pay for this party.

1378
00:47:31,295 --> 00:47:32,570
I don't even wanna to know who you

1379
00:47:32,570 --> 00:47:35,518
are. Right? A 0 knowledge proof is an

1380
00:47:35,518 --> 00:47:36,530
argument that let's

1381
00:47:37,046 --> 00:47:40,385
you convince me that you belong here. 0II

1382
00:47:40,385 --> 00:47:42,848
do know of a good. Actually, cloudflare. They...

1383
00:47:43,007 --> 00:47:44,756
They're the ones who applied this very well

1384
00:47:44,756 --> 00:47:46,844
and while back. They have this thing where...

1385
00:47:47,085 --> 00:47:48,764
And I don't know whether Cloud fire deserve

1386
00:47:48,764 --> 00:47:50,864
the public trust or whether this particular

1387
00:47:51,244 --> 00:47:53,565
technology is even sufficiently rigorous.

1388
00:47:53,979 --> 00:47:56,133
But it's a cool idea. We have to

1389
00:47:56,133 --> 00:47:58,127
not let the fact that it's cool blind

1390
00:47:58,127 --> 00:48:00,440
us to as that's hard for me to

1391
00:48:00,440 --> 00:48:01,716
do, but I I feel you. Go ahead.

1392
00:48:01,955 --> 00:48:03,710
I I am very subject to that bias.

1393
00:48:04,204 --> 00:48:06,838
As well, but it's important. Like, but it

1394
00:48:06,838 --> 00:48:09,152
is really cool. Like, Cloudflare has this thing

1395
00:48:09,152 --> 00:48:09,951
where it's like,

1396
00:48:10,510 --> 00:48:12,446
you have to, like, do a cap

1397
00:48:12,918 --> 00:48:15,223
to, like, view a website. Right?

1398
00:48:15,780 --> 00:48:18,084
And you get a cookie, but they claim

1399
00:48:18,084 --> 00:48:20,253
that they can't identify who you are

1400
00:48:20,564 --> 00:48:22,162
even though they set a cookie with a

1401
00:48:22,162 --> 00:48:25,598
unique identifier. Right. It's weird. Right. And what

1402
00:48:25,598 --> 00:48:28,326
they've done is they've constructed this proof that,

1403
00:48:28,484 --> 00:48:31,975
like somebody solved this cap, but they've swapped

1404
00:48:31,975 --> 00:48:33,880
your data with, like, 30 other people and

1405
00:48:33,880 --> 00:48:35,149
they don't know which 1 is you.

1406
00:48:35,784 --> 00:48:35,942
Right?

1407
00:48:36,594 --> 00:48:37,711
Somebody did a approve.

1408
00:48:38,349 --> 00:48:39,488
1 of those people

1409
00:48:40,105 --> 00:48:43,057
is viewing this site now. Right? But I

1410
00:48:43,057 --> 00:48:43,856
don't know which 1.

1411
00:48:44,826 --> 00:48:46,576
We're cloud for just know which 1. Sounds

1412
00:48:46,576 --> 00:48:48,008
like a crypto coin mixer.

1413
00:48:49,042 --> 00:48:50,258
It... It's analogous.

1414
00:48:50,633 --> 00:48:51,133
And

1415
00:48:51,444 --> 00:48:53,598
it's difficult to talk about your knowledge pros

1416
00:48:53,598 --> 00:48:55,693
because they are a genuine useful technology,

1417
00:48:56,548 --> 00:48:57,048
which

1418
00:48:57,426 --> 00:49:00,159
as of so many other things, the the

1419
00:49:00,217 --> 00:49:03,494
cryptocurrency people took this existing terminology and decided

1420
00:49:03,494 --> 00:49:06,594
this belongs to us now. Right? Yeah. It's

1421
00:49:06,594 --> 00:49:09,637
very frustrating because these technologies do have real

1422
00:49:09,773 --> 00:49:12,012
potential which it's hard to get to,

1423
00:49:12,568 --> 00:49:14,021
while while there's this

1424
00:49:14,474 --> 00:49:16,459
multi level marketing scheme in the way.

1425
00:49:17,809 --> 00:49:18,206
Yeah. Right.

1426
00:49:19,500 --> 00:49:21,820
Okay. Alright. So as we wrap up, looking

1427
00:49:21,820 --> 00:49:23,420
at some of these Pet, which which of

1428
00:49:23,420 --> 00:49:25,659
these actually do you feel have the most

1429
00:49:25,659 --> 00:49:26,480
chance of

1430
00:49:27,114 --> 00:49:28,954
of succeeding, You know, which 1 has the

1431
00:49:28,954 --> 00:49:31,195
best chance at widespread adoption and really helping

1432
00:49:31,195 --> 00:49:32,554
us. And then and then what's gonna drive

1433
00:49:32,554 --> 00:49:34,210
that adoption is what's gonna

1434
00:49:34,568 --> 00:49:36,324
bringing these things actually happen because the these

1435
00:49:36,324 --> 00:49:37,840
things slow things down and make things more

1436
00:49:37,840 --> 00:49:39,755
less convenient. These are all ivory tower things

1437
00:49:39,835 --> 00:49:41,510
Is anybody actually going to adopt these?

1438
00:49:42,481 --> 00:49:44,470
Well, many of them are adopted, and it's

1439
00:49:44,470 --> 00:49:46,402
because these companies are afraid of regulators,

1440
00:49:46,856 --> 00:49:48,527
and so they try to do proactive things.

1441
00:49:49,018 --> 00:49:50,685
The problem is the proactive things they do

1442
00:49:50,685 --> 00:49:53,145
were constantly broken and don't work, or they

1443
00:49:53,145 --> 00:49:54,494
have a carve out where we didn't cover

1444
00:49:54,494 --> 00:49:56,795
this corner case. You know, as programs we

1445
00:49:56,795 --> 00:49:58,878
have these things corner cases where it's like,

1446
00:49:59,355 --> 00:50:01,186
oh, you... You're this type of person, but

1447
00:50:01,186 --> 00:50:02,618
you're also that type of person at the

1448
00:50:02,618 --> 00:50:04,210
same time. I didn't know that was possible.

1449
00:50:04,767 --> 00:50:06,438
But it turns out there's a million people

1450
00:50:06,438 --> 00:50:08,628
a day. Who fall into that category,

1451
00:50:09,167 --> 00:50:11,724
and we didn't protect those people's privacy. Right?

1452
00:50:12,283 --> 00:50:12,603
Oops.

1453
00:50:14,521 --> 00:50:16,925
And so what to make it real. It

1454
00:50:16,925 --> 00:50:19,233
needs not just be something that companies talk

1455
00:50:19,233 --> 00:50:20,983
about, it needs to be something that they're

1456
00:50:20,983 --> 00:50:22,016
held accountable for.

1457
00:50:22,588 --> 00:50:24,650
And for that, we need regulatory scrutiny.

1458
00:50:25,047 --> 00:50:27,292
I firmly believe that we need regulatory scrutiny

1459
00:50:27,347 --> 00:50:28,085
because regulators

1460
00:50:29,330 --> 00:50:30,385
are independent

1461
00:50:30,773 --> 00:50:32,865
and they can. It's very similar to the

1462
00:50:32,922 --> 00:50:35,413
aviation model where we have this worldwide system

1463
00:50:35,788 --> 00:50:36,425
of, you know,

1464
00:50:37,460 --> 00:50:39,701
the Faa had and the Nt b have

1465
00:50:39,701 --> 00:50:40,439
in house

1466
00:50:40,891 --> 00:50:43,750
technical expertise and investigative authority. Perhaps not as

1467
00:50:43,750 --> 00:50:45,973
much as they should have, I've heard from

1468
00:50:45,973 --> 00:50:47,084
aviation people that they do like it to

1469
00:50:47,084 --> 00:50:47,402
be better.

1470
00:50:48,054 --> 00:50:49,968
But in in computers, we have none of

1471
00:50:49,968 --> 00:50:51,722
that, and we should be aspiring to have

1472
00:50:51,722 --> 00:50:52,680
even that. Right?

1473
00:50:53,318 --> 00:50:55,630
The companies should have to explain this to

1474
00:50:55,630 --> 00:50:57,067
experts, they should have to show how it

1475
00:50:57,067 --> 00:50:59,787
works. They should have to go into detail

1476
00:50:59,787 --> 00:51:01,856
on it. They should have to earn that

1477
00:51:01,856 --> 00:51:02,651
public trust.

1478
00:51:03,526 --> 00:51:06,968
Without outside truly independent scrutiny by people whose

1479
00:51:06,968 --> 00:51:09,702
paychecks don't come from the same organization

1480
00:51:10,159 --> 00:51:11,276
that cannot happen.

1481
00:51:12,009 --> 00:51:14,324
And it's not for lack of moral will

1482
00:51:14,324 --> 00:51:16,479
on the part of the privacy workers. It's

1483
00:51:16,479 --> 00:51:19,112
for, you know, it's your job. They'll fire

1484
00:51:19,112 --> 00:51:20,310
you if you push too hard.

1485
00:51:20,883 --> 00:51:23,274
Yeah. And and, you know, and III

1486
00:51:23,353 --> 00:51:24,469
I'd I make this argument been a lot

1487
00:51:24,469 --> 00:51:25,585
about regulation. A lot of people as soon

1488
00:51:25,585 --> 00:51:26,940
as you say the word regulation, you lose

1489
00:51:26,940 --> 00:51:28,946
half the audience, but I I'd there are

1490
00:51:28,946 --> 00:51:31,331
reasons for regulations. And and as you point

1491
00:51:31,331 --> 00:51:32,524
out, there are reasons why I don't have

1492
00:51:32,524 --> 00:51:34,671
to personally inspect my plane before I get

1493
00:51:34,671 --> 00:51:36,443
out and fly. There are reasons why

1494
00:51:36,754 --> 00:51:38,429
I don't have to poison test my food

1495
00:51:38,429 --> 00:51:39,545
before I eat it. There are reasons why

1496
00:51:39,705 --> 00:51:41,699
I don't have to do drug tests before

1497
00:51:41,858 --> 00:51:43,294
I take a drug. Someone's done that for

1498
00:51:43,294 --> 00:51:44,809
me, and that... And they did that because

1499
00:51:44,809 --> 00:51:46,166
they were required to do so by some

1500
00:51:46,166 --> 00:51:46,817
loss somewhere.

1501
00:51:47,375 --> 00:51:50,159
And that's that's regulation. I don't think everybody

1502
00:51:50,159 --> 00:51:52,648
in the world appreciates how often,

1503
00:51:53,103 --> 00:51:53,342
you know,

1504
00:51:54,233 --> 00:51:57,025
restaurants actually do cause food poisoning, but it

1505
00:51:57,025 --> 00:51:59,577
gets caught very early because of these regulations,

1506
00:51:59,975 --> 00:52:02,448
and so hardly anybody is affected. The problem

1507
00:52:02,448 --> 00:52:03,086
gets fixed.

1508
00:52:03,819 --> 00:52:05,976
And so it's very... You can't see something

1509
00:52:05,976 --> 00:52:08,373
that doesn't happen, so you don't see a

1510
00:52:08,373 --> 00:52:11,010
million cases of food poisoning because we prevented

1511
00:52:11,010 --> 00:52:11,170
them.

1512
00:52:11,822 --> 00:52:14,442
Right? But if you look at, like,

1513
00:52:15,235 --> 00:52:17,538
local news reports, or you look at any,

1514
00:52:17,697 --> 00:52:20,176
like, public records on this topic. It's very

1515
00:52:20,176 --> 00:52:22,088
clear that these are widespread problems.

1516
00:52:22,885 --> 00:52:25,514
The Fda inspect food before it goes on

1517
00:52:25,514 --> 00:52:26,311
sale in the stores.

1518
00:52:27,121 --> 00:52:28,951
They catch a lot of problems and a

1519
00:52:28,951 --> 00:52:30,724
lot of products get taken off the shelves

1520
00:52:31,258 --> 00:52:32,213
before people get sick.

1521
00:52:32,929 --> 00:52:33,167
Right?

1522
00:52:33,818 --> 00:52:34,318
And

1523
00:52:34,773 --> 00:52:36,442
the company made the case to us as

1524
00:52:36,442 --> 00:52:38,192
workers and at 1 time 9 life I

1525
00:52:38,192 --> 00:52:40,099
believed it, that they should be subject only

1526
00:52:40,099 --> 00:52:42,326
to what they call self regulation, they should

1527
00:52:42,326 --> 00:52:45,284
be judged during execution are. Mh. And they

1528
00:52:45,284 --> 00:52:47,694
made the case because they believe that rapid

1529
00:52:47,990 --> 00:52:51,492
innovation is only possible in that light. But,

1530
00:52:51,651 --> 00:52:52,708
no. This is

1531
00:52:53,496 --> 00:52:55,427
this is a bluff game. The company

1532
00:52:56,438 --> 00:52:57,813
intentionally does not

1533
00:52:58,426 --> 00:53:01,630
educate the public or regulatory authorities on these

1534
00:53:01,686 --> 00:53:02,482
engineering details.

1535
00:53:02,974 --> 00:53:04,246
So that they can then go to the

1536
00:53:04,246 --> 00:53:05,382
press and claim

1537
00:53:05,996 --> 00:53:08,461
that any criticism is not informed by the

1538
00:53:08,461 --> 00:53:10,130
technical realities of the situation.

1539
00:53:10,926 --> 00:53:12,516
So that they can then say, like,

1540
00:53:13,168 --> 00:53:15,315
Well, these politicians made this fall without our

1541
00:53:15,315 --> 00:53:17,143
input. Yeah. Because you refused to give your

1542
00:53:17,143 --> 00:53:17,303
input.

1543
00:53:18,018 --> 00:53:18,814
Right? Like,

1544
00:53:19,450 --> 00:53:21,610
somebody has to they're playing chicken. You know,

1545
00:53:22,086 --> 00:53:23,275
I want to be in a world where

1546
00:53:23,275 --> 00:53:25,335
there is engagement where there is back and

1547
00:53:25,335 --> 00:53:27,341
forth where it's a public discussion

1548
00:53:27,951 --> 00:53:28,451
because

1549
00:53:28,759 --> 00:53:30,827
First off, these things are very detailed centric.

1550
00:53:30,986 --> 00:53:33,235
We can't get to good outcomes by accident

1551
00:53:33,531 --> 00:53:35,463
or by hoping. Hope is not a strategy

1552
00:53:35,679 --> 00:53:37,682
as my former and player love to say,

1553
00:53:38,318 --> 00:53:39,830
you know, the public is a stakeholder in

1554
00:53:39,830 --> 00:53:41,898
these discussions and we need to understand and

1555
00:53:41,898 --> 00:53:42,398
care

1556
00:53:42,853 --> 00:53:45,651
about the details to of all this, and

1557
00:53:45,651 --> 00:53:48,774
we need to, you know, make sure that

1558
00:53:49,069 --> 00:53:50,579
the rules that we put in place are

1559
00:53:50,579 --> 00:53:53,362
rules that serve the public interest that actually

1560
00:53:53,362 --> 00:53:55,762
address the concerns we care about and don't

1561
00:53:55,762 --> 00:53:57,454
just en ent bad practices.

1562
00:53:58,465 --> 00:54:00,134
I'm not optimistic for the state of it

1563
00:54:00,134 --> 00:54:01,644
in the United States. I don't see the

1564
00:54:01,644 --> 00:54:03,098
political will here, but

1565
00:54:03,409 --> 00:54:05,567
Europe is moving in a helpful direction. I

1566
00:54:05,567 --> 00:54:06,546
was part of Gdpr

1567
00:54:06,925 --> 00:54:09,024
implementation. It was the only time

1568
00:54:09,802 --> 00:54:11,880
that I really saw what the company could

1569
00:54:11,880 --> 00:54:13,835
do when it was highly motivated.

1570
00:54:14,375 --> 00:54:15,894
Right? Why aren't it why aren't they doing

1571
00:54:15,894 --> 00:54:16,535
that all the time?

1572
00:54:17,175 --> 00:54:18,855
Right. Well, you know, I I don't wanna

1573
00:54:18,855 --> 00:54:20,789
be an ap a capital deposit. But I

1574
00:54:20,949 --> 00:54:21,910
I think a lot of this is these

1575
00:54:21,910 --> 00:54:22,550
companies are,

1576
00:54:23,269 --> 00:54:25,590
their shareholders demand return. And when they see

1577
00:54:25,590 --> 00:54:26,869
money on the table, they said, you gotta

1578
00:54:26,869 --> 00:54:28,324
pick that up. Mean, and that money in

1579
00:54:28,324 --> 00:54:30,242
this case is monetizing your data because there's

1580
00:54:30,242 --> 00:54:32,579
nothing saying they camps. So it's perfectly legal

1581
00:54:32,639 --> 00:54:34,157
for them to do so and the shareholders

1582
00:54:34,157 --> 00:54:35,836
like, well, you... Would you make more money

1583
00:54:35,836 --> 00:54:37,690
if you did that, They're lot Yeah. Well

1584
00:54:37,690 --> 00:54:38,250
then do that.

1585
00:54:39,369 --> 00:54:41,690
So, you know, I think I I I'd

1586
00:54:41,690 --> 00:54:43,950
be very interested to know what level of

1587
00:54:44,423 --> 00:54:45,855
some of the executive of these companies would

1588
00:54:45,855 --> 00:54:47,763
be happy to have regulations, so they wouldn't

1589
00:54:47,763 --> 00:54:49,512
have to bow to shareholders demands to do

1590
00:54:49,512 --> 00:54:50,784
this. I think they're at least some out

1591
00:54:50,784 --> 00:54:52,079
there. Maybe I'm a wide eyed.

1592
00:54:52,548 --> 00:54:53,583
I think there are at least some people

1593
00:54:53,583 --> 00:54:56,210
out there, like, please regulators this. Please because

1594
00:54:56,289 --> 00:54:57,483
I don't wanna do this anymore, but my

1595
00:54:57,483 --> 00:54:59,075
shareholders demand, and there's nothing saying I can't

1596
00:54:59,075 --> 00:54:59,473
so I will.

1597
00:55:00,269 --> 00:55:02,672
I don't know. I think that if that

1598
00:55:02,672 --> 00:55:04,422
is true, it's true in the company's whose

1599
00:55:04,422 --> 00:55:06,809
primary income is from something other than advertising.

1600
00:55:07,923 --> 00:55:10,571
Again, I had pretty good visibility upper management

1601
00:55:10,707 --> 00:55:12,233
because I was... Got very good at reading

1602
00:55:12,233 --> 00:55:13,822
the Tea leaves and knowing who was saying

1603
00:55:13,822 --> 00:55:15,807
what and so on. And because I talked

1604
00:55:15,807 --> 00:55:17,181
to people throughout the company

1605
00:55:17,635 --> 00:55:19,342
and got I

1606
00:55:19,957 --> 00:55:22,265
believe that executives want to believe that they

1607
00:55:22,265 --> 00:55:23,061
are good people,

1608
00:55:23,857 --> 00:55:25,290
and I also believe that when you have

1609
00:55:25,290 --> 00:55:26,166
billions of dollars,

1610
00:55:26,738 --> 00:55:28,411
you believe that being a good person is

1611
00:55:28,411 --> 00:55:29,286
something you can buy.

1612
00:55:30,880 --> 00:55:32,074
It's it's just distort.

1613
00:55:32,951 --> 00:55:35,442
Anyone who has that level of power to

1614
00:55:35,674 --> 00:55:38,623
you know, frankly spin up entire little organizations

1615
00:55:38,623 --> 00:55:40,695
to sell a narrative on their behalf anytime

1616
00:55:40,695 --> 00:55:41,434
they choose

1617
00:55:41,811 --> 00:55:43,166
just on the drop of a hat,

1618
00:55:43,978 --> 00:55:46,046
gets used to using it and starts to

1619
00:55:46,046 --> 00:55:48,216
believe that they deserve it. And,

1620
00:55:48,989 --> 00:55:50,660
you know, I don't want to, like,

1621
00:55:51,455 --> 00:55:51,955
dismantle

1622
00:55:52,983 --> 00:55:56,112
I I don't want to take an unnecessarily

1623
00:55:56,250 --> 00:55:57,923
ideological position here. I'm trying to make a

1624
00:55:57,923 --> 00:55:58,902
practical argument

1625
00:55:59,358 --> 00:55:59,916
that, like,

1626
00:56:00,568 --> 00:56:02,793
it almost doesn't matter what's in executive hearts

1627
00:56:02,793 --> 00:56:04,224
at the end of the day, although,

1628
00:56:05,416 --> 00:56:07,562
it... Because it matters what they do.

1629
00:56:08,135 --> 00:56:09,574
And they will do the right thing when

1630
00:56:09,574 --> 00:56:11,094
they are incentivized to do the right thing.

1631
00:56:11,255 --> 00:56:13,255
When we have some actual ground rules. Right?

1632
00:56:13,894 --> 00:56:15,175
And then perhaps none of us will have

1633
00:56:15,175 --> 00:56:17,186
to move too closely at what... What works

1634
00:56:17,186 --> 00:56:17,981
in the hearts of man?

1635
00:56:20,128 --> 00:56:22,377
Alright. So before we go, why don't you

1636
00:56:22,513 --> 00:56:23,706
tell us... What it was next for you

1637
00:56:23,706 --> 00:56:25,456
and what and in internet safety labs? What

1638
00:56:25,456 --> 00:56:26,012
are guys up to?

1639
00:56:26,824 --> 00:56:29,216
Yeah. You know, we are a tiny nonprofit.

1640
00:56:29,615 --> 00:56:32,088
We can always use funding as I'm sure

1641
00:56:32,088 --> 00:56:34,640
every non nonprofit Mh. Says constantly. It's my

1642
00:56:34,640 --> 00:56:36,336
job to say that. I'm trying to

1643
00:56:36,888 --> 00:56:37,844
Noticed I noticed.

1644
00:56:38,482 --> 00:56:41,590
But, we are a tiny nonprofit with I

1645
00:56:41,590 --> 00:56:43,901
think a really solid team. We're building out,

1646
00:56:43,981 --> 00:56:45,213
like, testing

1647
00:56:45,591 --> 00:56:47,506
methodologies so we can look at this software

1648
00:56:47,506 --> 00:56:50,219
and say, does it do these things to

1649
00:56:50,219 --> 00:56:52,466
what extent does it do these things? Who's

1650
00:56:52,466 --> 00:56:54,316
better and who's worse? And

1651
00:56:54,769 --> 00:56:56,437
in what ways are they better or worse?

1652
00:56:56,755 --> 00:56:57,811
We want to

1653
00:56:58,343 --> 00:56:59,931
inform, like, maybe if you're...

1654
00:57:00,424 --> 00:57:02,099
We have a focus right now in educational

1655
00:57:02,099 --> 00:57:04,890
technology because that's been an area with a

1656
00:57:04,890 --> 00:57:06,963
lot of very obvious problems at a very

1657
00:57:06,963 --> 00:57:07,463
vulnerable

1658
00:57:07,840 --> 00:57:08,000
population.

1659
00:57:08,797 --> 00:57:11,208
If you're, like, the person who decides for,

1660
00:57:11,288 --> 00:57:13,685
like, a school district, what what apps to

1661
00:57:13,685 --> 00:57:14,244
we use,

1662
00:57:14,804 --> 00:57:16,562
you can look at our research and we

1663
00:57:16,562 --> 00:57:18,000
will tell you which apps are better or

1664
00:57:18,000 --> 00:57:19,875
worse. That is all public

1665
00:57:20,175 --> 00:57:22,015
information free on our website for anybody.

1666
00:57:22,575 --> 00:57:26,015
We also hope that over time regulators will

1667
00:57:26,015 --> 00:57:27,375
pay attention because they're the ones who have

1668
00:57:27,375 --> 00:57:30,520
the power to take in here. Right? Technology

1669
00:57:30,577 --> 00:57:32,248
companies have the power to take action, but

1670
00:57:32,248 --> 00:57:33,760
in our experience, they don't want to hear

1671
00:57:33,760 --> 00:57:36,067
it, which makes sense. We're telling them about

1672
00:57:36,067 --> 00:57:37,992
some additional money that they're gonna spend to

1673
00:57:37,992 --> 00:57:39,979
fix a problem. They'd rather not spend it.

1674
00:57:40,138 --> 00:57:41,330
Right? It's cheaper not to.

1675
00:57:42,045 --> 00:57:44,112
So regulators are the ones who have the

1676
00:57:44,112 --> 00:57:45,860
ability to take action. And I think there's

1677
00:57:45,860 --> 00:57:47,789
a lot of, like, work to be done

1678
00:57:47,789 --> 00:57:49,550
over the next few years to, like,

1679
00:57:50,910 --> 00:57:53,410
bridge the knowledge from industry

1680
00:57:53,949 --> 00:57:55,170
to everywhere

1681
00:57:56,042 --> 00:57:59,468
where actual policy decisions are made and where

1682
00:57:59,468 --> 00:58:00,924
we can do informed

1683
00:58:02,097 --> 00:58:02,597
Nuance

1684
00:58:03,866 --> 00:58:04,924
protections that,

1685
00:58:05,381 --> 00:58:07,556
you know, protect innovation while

1686
00:58:08,013 --> 00:58:11,203
moving forward with protecting society as well. Right?

1687
00:58:12,734 --> 00:58:14,969
Alright. Well, good luck with that. We need

1688
00:58:14,969 --> 00:58:15,129
it.

1689
00:58:16,007 --> 00:58:16,507
So

1690
00:58:16,965 --> 00:58:18,722
I... I very much hope you succeed for

1691
00:58:18,722 --> 00:58:20,557
all for all of our sakes. Thanks so

1692
00:58:20,557 --> 00:58:21,675
much for coming out of the, Irene.

1693
00:58:22,487 --> 00:58:24,397
My pleasure. Thank you for covering these topics.

1694
00:58:24,556 --> 00:58:25,273
They're so important.

1695
00:58:31,735 --> 00:58:33,255
It was really great talking with Irene.

1696
00:58:33,655 --> 00:58:35,574
They had some very interesting perspectives,

1697
00:58:36,135 --> 00:58:38,215
and it was just so interesting to talk

1698
00:58:38,215 --> 00:58:40,304
to somebody who's actually been in the belly

1699
00:58:40,304 --> 00:58:42,053
of the beast who has has worked for

1700
00:58:42,212 --> 00:58:43,746
Google, a company that I

1701
00:58:44,358 --> 00:58:46,505
bash on a lot at least in terms

1702
00:58:46,505 --> 00:58:47,061
of privacy.

1703
00:58:47,553 --> 00:58:48,584
I mean, I I will say it and

1704
00:58:48,663 --> 00:58:50,249
I've said multiple times. They make some really

1705
00:58:50,249 --> 00:58:53,183
great products. Some of their, you know, tools

1706
00:58:53,183 --> 00:58:54,873
and apps and whatever are

1707
00:58:55,660 --> 00:58:56,380
best in class.

1708
00:58:57,420 --> 00:59:00,160
Google Docs is wonderful, but the privacy aspects

1709
00:59:00,219 --> 00:59:01,980
just ruin everything as far as I'm concerned.

1710
00:59:02,794 --> 00:59:04,145
So I was really glad to have the

1711
00:59:04,145 --> 00:59:05,360
opportunity to talk to somebody

1712
00:59:05,973 --> 00:59:07,801
who has fence a dime on the inside

1713
00:59:07,801 --> 00:59:09,152
about what things are like at Google and

1714
00:59:09,152 --> 00:59:11,079
how they work there. And doing a little

1715
00:59:11,079 --> 00:59:12,360
bit of research trying to find some links

1716
00:59:12,360 --> 00:59:14,440
for the show. I did run across an

1717
00:59:14,440 --> 00:59:14,920
article that,

1718
00:59:15,960 --> 00:59:17,320
Irene posted on medium.

1719
00:59:18,212 --> 00:59:19,406
When they left Google,

1720
00:59:20,043 --> 00:59:21,498
kind of as a plea to

1721
00:59:21,873 --> 00:59:24,500
their former employees about not giving up on

1722
00:59:24,500 --> 00:59:26,569
privacy and how Irene was not giving up

1723
00:59:26,569 --> 00:59:27,206
on privacy.

1724
00:59:27,619 --> 00:59:28,358
And how

1725
00:59:29,131 --> 00:59:29,529
frankly,

1726
00:59:30,086 --> 00:59:33,031
employees, tech employees working for these big tech

1727
00:59:33,031 --> 00:59:33,531
companies

1728
00:59:34,146 --> 00:59:37,113
have power. Have say, can

1729
00:59:38,057 --> 00:59:41,174
band together to push back from the inside

1730
00:59:41,863 --> 00:59:44,820
on some of these privacy invasive projects

1731
00:59:45,527 --> 00:59:47,198
and they have done so. Some of these

1732
00:59:47,198 --> 00:59:49,744
were government projects with, like, facial recognition,

1733
00:59:50,620 --> 00:59:52,212
and things like that, where they have been

1734
00:59:52,212 --> 00:59:53,883
successful. I'm sure there's cases where they haven't

1735
00:59:53,883 --> 00:59:55,573
as well. But if you are,

1736
00:59:56,452 --> 00:59:57,411
a tech employee,

1737
00:59:58,050 --> 01:00:00,367
and you are seeing your company, do some

1738
01:00:00,367 --> 01:00:02,379
things that you think are not right, You

1739
01:00:02,379 --> 01:00:04,217
do have some power to make change within

1740
01:00:04,217 --> 01:00:05,815
your company. I know it could be tough.

1741
01:00:06,215 --> 01:00:07,653
You might need to band together with some

1742
01:00:07,653 --> 01:00:08,612
of your other employees,

1743
01:00:09,572 --> 01:00:10,690
but you can make a difference.

1744
01:00:11,742 --> 01:00:13,970
Advocate for privacy where you can, push back

1745
01:00:13,970 --> 01:00:17,572
on, you know, privacy invasive technologies and over

1746
01:00:17,630 --> 01:00:18,266
collection of data,

1747
01:00:18,997 --> 01:00:21,462
and maybe take some of these privacy enhancing

1748
01:00:21,462 --> 01:00:23,767
technologies that we've been talking about and suggest

1749
01:00:23,767 --> 01:00:25,835
that to your employers that you try to

1750
01:00:25,835 --> 01:00:27,743
incorporate those into your products where you can,

1751
01:00:28,394 --> 01:00:30,312
for the benefit of all of us. So

1752
01:00:30,312 --> 01:00:31,590
as usual, there would be some bonus Q

1753
01:00:31,590 --> 01:00:32,469
and a with my guest,

1754
01:00:33,428 --> 01:00:35,666
for the patrons on the bonus podcast on

1755
01:00:35,826 --> 01:00:36,326
Thursday,

1756
01:00:36,880 --> 01:00:38,079
We talk a little bit more about what

1757
01:00:38,079 --> 01:00:39,840
it was like to work inside the privacy

1758
01:00:39,840 --> 01:00:41,760
department of Google. We talk about data mini

1759
01:00:41,760 --> 01:00:44,000
randomization. We talk about how to properly a,

1760
01:00:44,159 --> 01:00:45,335
modern hard drive, including

1761
01:00:46,014 --> 01:00:47,692
Ssds, which are kinda tricky you were erase

1762
01:00:47,692 --> 01:00:49,690
properly. So Patrons you will get that on

1763
01:00:49,770 --> 01:00:50,969
Thursday as usual.

1764
01:00:52,407 --> 01:00:54,405
Okay. Now it's time to talk about the

1765
01:00:54,405 --> 01:00:56,497
book surge. So again, you can get all

1766
01:00:56,497 --> 01:00:58,413
the details just by going to fd DSD

1767
01:00:58,413 --> 01:01:00,488
dot me slash book search or find the

1768
01:01:00,488 --> 01:01:02,005
link in the show notes, but here's kinda

1769
01:01:02,005 --> 01:01:02,643
of what's going on.

1770
01:01:05,052 --> 01:01:06,486
If you have a notice, I I really

1771
01:01:06,486 --> 01:01:08,956
don't like marketing. I'm not I'm not good

1772
01:01:08,956 --> 01:01:10,868
at it, and that probably says a lot

1773
01:01:10,868 --> 01:01:12,637
about, you know, where my book in this

1774
01:01:12,637 --> 01:01:14,232
podcast are at this point. I, they're doing

1775
01:01:14,232 --> 01:01:16,486
fine. But I mean, I always feel like

1776
01:01:16,943 --> 01:01:18,698
if I was better at self promotion,

1777
01:01:20,373 --> 01:01:21,465
that I would be

1778
01:01:21,825 --> 01:01:24,305
reaching more people. And I don't wanna be

1779
01:01:24,305 --> 01:01:26,385
famous. I don't wanna be rich. But I

1780
01:01:26,385 --> 01:01:27,985
do wanna help people. I'd wanna reach as

1781
01:01:27,985 --> 01:01:29,605
many people as I possibly can

1782
01:01:30,144 --> 01:01:30,644
because

1783
01:01:31,517 --> 01:01:33,190
Again, we can kinda build this herd immunity

1784
01:01:33,190 --> 01:01:35,500
of the more people that take these basic

1785
01:01:35,500 --> 01:01:37,969
privacy and security measures, being pick the low

1786
01:01:37,969 --> 01:01:39,562
hanging fruit. The better off we will all

1787
01:01:39,562 --> 01:01:39,722
be.

1788
01:01:40,535 --> 01:01:42,614
Because your privacy and security overlap mine.

1789
01:01:43,335 --> 01:01:44,614
You know, it's a we thing not a

1790
01:01:44,614 --> 01:01:46,054
methane thing as I often like to say.

1791
01:01:46,775 --> 01:01:47,255
And so,

1792
01:01:48,547 --> 01:01:50,782
If I believe that, and I do, wholeheartedly.

1793
01:01:51,501 --> 01:01:53,336
I want people to read my book, and

1794
01:01:53,496 --> 01:01:54,933
I want people to send the podcast and

1795
01:01:54,933 --> 01:01:56,210
the newsletter and all that sort of stuff.

1796
01:01:56,625 --> 01:01:58,625
So I... I kinda got inspired and I'm

1797
01:01:58,625 --> 01:02:00,224
trying to force myself out of my comfort

1798
01:02:00,224 --> 01:02:01,285
zone a little bit more

1799
01:02:01,664 --> 01:02:04,498
and and try to do some promotion. And

1800
01:02:04,635 --> 01:02:06,885
honestly, what I really need, which I also

1801
01:02:07,420 --> 01:02:09,886
don't often do is ask for help, specifically,

1802
01:02:10,682 --> 01:02:12,512
your help, my audience.

1803
01:02:13,163 --> 01:02:15,072
I want to crowds some of this stuff

1804
01:02:15,072 --> 01:02:17,299
where I can. And so, I thought of

1805
01:02:17,299 --> 01:02:18,890
a few different things, and the first 1

1806
01:02:18,970 --> 01:02:19,686
I'm going to try.

1807
01:02:20,736 --> 01:02:22,329
Is what I'm calling a book surge.

1808
01:02:22,887 --> 01:02:25,994
So on Amazon, there's, you know, millions probably

1809
01:02:25,994 --> 01:02:28,145
literally millions of books on Amazon and to

1810
01:02:28,145 --> 01:02:29,992
try to get notice that Amazon is really

1811
01:02:29,992 --> 01:02:30,627
hard to do.

1812
01:02:31,342 --> 01:02:34,123
So unless you come up in Amazon search

1813
01:02:34,123 --> 01:02:34,623
results

1814
01:02:35,077 --> 01:02:37,142
or unless you somehow pop up somewhere else

1815
01:02:37,142 --> 01:02:40,190
on Amazon, it's part of some top secret

1816
01:02:40,190 --> 01:02:42,829
algorithm that places your book on a web

1817
01:02:42,829 --> 01:02:43,329
page

1818
01:02:43,710 --> 01:02:45,880
within Amazon where they would see it. And

1819
01:02:45,880 --> 01:02:49,291
so while their algorithms are completely top secret

1820
01:02:49,291 --> 01:02:50,084
and proprietary,

1821
01:02:51,194 --> 01:02:54,630
you know, most people agree that to rise

1822
01:02:54,630 --> 01:02:56,650
to the top of search results

1823
01:02:56,949 --> 01:02:59,829
or to get featured somewhere on the website,

1824
01:03:00,469 --> 01:03:02,800
it would depend on things like how many

1825
01:03:02,800 --> 01:03:05,023
books you've sold? How many people are have

1826
01:03:05,023 --> 01:03:07,484
looked at your book, what the ratings are

1827
01:03:07,484 --> 01:03:09,310
for your book, How many ratings you have

1828
01:03:09,310 --> 01:03:11,077
for your book, And in particular,

1829
01:03:11,715 --> 01:03:15,008
every book on Amazon gets assigned to 3

1830
01:03:15,146 --> 01:03:15,646
categories

1831
01:03:16,264 --> 01:03:17,541
where if you look at the webpage page

1832
01:03:17,541 --> 01:03:18,909
may have just scrolled past this. But if

1833
01:03:18,909 --> 01:03:20,260
you look at any book on Amazon and

1834
01:03:20,260 --> 01:03:21,930
scroll down a little bit, it'll have the

1835
01:03:21,930 --> 01:03:24,713
overall ranking of how a book is selling

1836
01:03:24,713 --> 01:03:27,336
within all possible books. But right below that,

1837
01:03:27,908 --> 01:03:29,024
there's 3 categories.

1838
01:03:29,581 --> 01:03:32,131
And these are sometimes negotiated with the author

1839
01:03:32,131 --> 01:03:32,848
or the publisher.

1840
01:03:33,884 --> 01:03:36,275
It's it's actually quite the opaque process. Let

1841
01:03:36,275 --> 01:03:36,832
me tell you.

1842
01:03:37,723 --> 01:03:40,431
But there's these 3 categories. And they give

1843
01:03:40,431 --> 01:03:42,660
you a ranking for where you currently are

1844
01:03:42,660 --> 01:03:44,412
within that hour, I think. I think it's

1845
01:03:44,412 --> 01:03:45,686
an hourly updated thing.

1846
01:03:46,418 --> 01:03:48,412
Where you are selling within that category. And

1847
01:03:48,412 --> 01:03:50,486
then varies quite a bit. I mean, if

1848
01:03:50,486 --> 01:03:52,161
there's a rush on book sales, you know,

1849
01:03:52,320 --> 01:03:53,597
then not, you know, in your book doesn't

1850
01:03:53,597 --> 01:03:55,032
it sell anymore than normal, then it's gonna

1851
01:03:55,032 --> 01:03:56,880
fall down because A whole bunch of other

1852
01:03:56,880 --> 01:03:58,474
books are selling more. Or if it's a

1853
01:03:58,474 --> 01:04:00,146
really slow day. And for some reason,

1854
01:04:00,863 --> 01:04:02,615
your book has a few more people buying

1855
01:04:02,615 --> 01:04:04,527
it than normal, that might pop you up

1856
01:04:04,527 --> 01:04:06,291
a little bit. So it really varies quite

1857
01:04:06,291 --> 01:04:08,672
a bit. But that is 1 way that

1858
01:04:08,672 --> 01:04:09,943
people would see the book because if they're

1859
01:04:09,943 --> 01:04:11,292
clicking on a book category if they're just

1860
01:04:11,292 --> 01:04:12,880
kinda, I'll show me show me books in

1861
01:04:12,880 --> 01:04:14,071
the computer privacy area.

1862
01:04:14,723 --> 01:04:17,117
Then there's a list, and there's a ranked

1863
01:04:17,117 --> 01:04:18,952
list. And so what you really wanna be

1864
01:04:18,952 --> 01:04:20,547
is you wanna be on that first page

1865
01:04:20,547 --> 01:04:22,630
of results. And by the way, this is

1866
01:04:22,630 --> 01:04:24,058
just drives me. That's 1 of 1 of

1867
01:04:24,058 --> 01:04:25,010
the categories I'm in,

1868
01:04:25,804 --> 01:04:27,629
I was looking at. And, like, of the

1869
01:04:27,629 --> 01:04:28,264
top 10 books.

1870
01:04:29,393 --> 01:04:30,611
3 of the books

1871
01:04:31,468 --> 01:04:33,623
were all basically titled the same thing and

1872
01:04:33,623 --> 01:04:34,900
we're the same kind of book and there

1873
01:04:34,900 --> 01:04:35,799
it was wtf

1874
01:04:36,177 --> 01:04:37,135
is my password.

1875
01:04:37,947 --> 01:04:39,860
And so their their password journals. There places

1876
01:04:39,860 --> 01:04:41,773
for people to write passwords down. 3 of

1877
01:04:41,773 --> 01:04:42,570
the top 10 books,

1878
01:04:43,447 --> 01:04:45,121
anyway, it it is what it is.

1879
01:04:46,092 --> 01:04:48,245
If you wanna get seen, another way to

1880
01:04:48,245 --> 01:04:50,238
get noticed is to be in those, you

1881
01:04:50,238 --> 01:04:52,390
know, on the top search results in the

1882
01:04:52,390 --> 01:04:53,927
first page, usually of

1883
01:04:54,239 --> 01:04:56,239
any of those kind of book category lists.

1884
01:04:56,880 --> 01:04:58,559
So what can you help me? Okay. So

1885
01:04:58,559 --> 01:05:00,800
for the next week, here are some things

1886
01:05:00,800 --> 01:05:03,059
that you can do to help me

1887
01:05:03,534 --> 01:05:05,855
get more notice on Amazon to to raise

1888
01:05:05,855 --> 01:05:08,434
up in these rankings to to to hopefully

1889
01:05:08,894 --> 01:05:10,755
prime their algorithms to

1890
01:05:11,295 --> 01:05:13,306
get seen more more people. So first of

1891
01:05:13,306 --> 01:05:14,659
all, I would like to try to get

1892
01:05:14,659 --> 01:05:16,092
in the top 10 of what of the

1893
01:05:16,092 --> 01:05:18,081
3 categories that my book has been assigned

1894
01:05:18,081 --> 01:05:19,991
to. If you go to the article, you

1895
01:05:19,991 --> 01:05:20,150
can...

1896
01:05:20,804 --> 01:05:22,324
It'll take you to those 3 categories and

1897
01:05:22,324 --> 01:05:23,065
show you

1898
01:05:23,605 --> 01:05:25,764
what the current top listings are in those

1899
01:05:25,764 --> 01:05:26,244
categories.

1900
01:05:26,885 --> 01:05:28,324
And I think the main way to get

1901
01:05:28,324 --> 01:05:30,410
there honestly is this straight up book sales.

1902
01:05:30,727 --> 01:05:31,759
Some of these other things I'm gonna talk

1903
01:05:31,759 --> 01:05:33,267
about. I'm sure feed into this as well,

1904
01:05:33,664 --> 01:05:35,490
but this is where I've been saying. If

1905
01:05:35,490 --> 01:05:37,712
you were thinking about buying my book for

1906
01:05:37,712 --> 01:05:39,080
yourself, for someone else,

1907
01:05:40,036 --> 01:05:40,536
maybe

1908
01:05:41,311 --> 01:05:42,825
for a Christmas present, you can start your

1909
01:05:42,905 --> 01:05:45,375
Christmas shopping early. Maybe you wanna donate it

1910
01:05:45,375 --> 01:05:47,287
to your library or something like that. If

1911
01:05:47,287 --> 01:05:49,536
you if you were thinking about buying a

1912
01:05:49,536 --> 01:05:50,733
copy of my book for any reason,

1913
01:05:51,371 --> 01:05:53,925
do it this week. So between Monday, July

1914
01:05:53,925 --> 01:05:55,600
eighth and Sunday, July fourteenth.

1915
01:05:56,175 --> 01:05:57,534
You know, I'm trying to get this kinda

1916
01:05:57,534 --> 01:05:58,034
concentrated

1917
01:05:58,414 --> 01:06:00,255
so that this is all happening at 1

1918
01:06:00,255 --> 01:06:02,275
time, and hopefully focuses

1919
01:06:02,974 --> 01:06:06,098
the effect. Also, I really really need book

1920
01:06:06,098 --> 01:06:06,257
reviews.

1921
01:06:07,213 --> 01:06:09,124
Amazon really likes those book reviews to come

1922
01:06:09,124 --> 01:06:11,751
from people who are verified purchasers of that

1923
01:06:11,751 --> 01:06:13,999
book. So if you have purchased the book

1924
01:06:13,999 --> 01:06:16,631
on Amazon, I would very much appreciate if

1925
01:06:16,631 --> 01:06:19,343
you could leave a 5 star review. You

1926
01:06:19,343 --> 01:06:21,257
don't have to put any text with that

1927
01:06:21,257 --> 01:06:23,507
review, just putting the star review will be

1928
01:06:23,507 --> 01:06:25,982
very, very helpful, but a little text goes

1929
01:06:25,982 --> 01:06:28,137
a long way. Just a sentence or 2

1930
01:06:28,137 --> 01:06:29,974
would be great. This is the fifth edition

1931
01:06:29,974 --> 01:06:31,890
of my book and unfortunately, Amazon doesn't let

1932
01:06:31,890 --> 01:06:34,216
me carry forward the reviews from previous additions.

1933
01:06:34,614 --> 01:06:36,129
So all those great reviews I've gotten in

1934
01:06:36,129 --> 01:06:37,722
previous additions in my book are all gone.

1935
01:06:37,882 --> 01:06:39,157
That's a clean slate when I do a

1936
01:06:39,157 --> 01:06:39,635
new edition.

1937
01:06:40,367 --> 01:06:42,037
Despite the fact that I've sold hundreds of

1938
01:06:42,037 --> 01:06:43,310
these books on Amazon,

1939
01:06:44,105 --> 01:06:46,651
I've only got 27 reviews. Now I would

1940
01:06:46,651 --> 01:06:48,659
have like to get feedback. If there's something

1941
01:06:48,659 --> 01:06:50,015
you don't like about the book, There will

1942
01:06:50,015 --> 01:06:50,813
be a next edition.

1943
01:06:51,530 --> 01:06:53,046
But if you if you don't feel comfortable

1944
01:06:53,046 --> 01:06:54,482
giving it a 5 star review, then just

1945
01:06:54,482 --> 01:06:55,917
send me an email. I would I would

1946
01:06:55,917 --> 01:06:57,925
very much like to get feedback, set it

1947
01:06:57,925 --> 01:07:00,473
to feedback at firewalls don't dragons dot com.

1948
01:07:00,712 --> 01:07:02,225
But if you like the book, I would

1949
01:07:02,225 --> 01:07:04,136
love for you to put a 5 star

1950
01:07:04,136 --> 01:07:05,092
review on Amazon.

1951
01:07:05,823 --> 01:07:06,699
The other thing I wanna do is I

1952
01:07:06,699 --> 01:07:08,450
wanna get my book ranking up. Right now,

1953
01:07:08,609 --> 01:07:10,918
it's 4.4 stars out of 5. I would

1954
01:07:10,918 --> 01:07:12,032
really love to get that up to about

1955
01:07:12,032 --> 01:07:13,146
the 4.8 range.

1956
01:07:13,719 --> 01:07:16,106
And since there are so few reviews, wouldn't

1957
01:07:16,106 --> 01:07:17,460
take too many 5 star reviews to get

1958
01:07:17,460 --> 01:07:19,052
it up there. So if that would... That

1959
01:07:19,052 --> 01:07:20,246
that shouldn't be too hard to do if

1960
01:07:20,246 --> 01:07:22,090
we can get some more reviews. And there's

1961
01:07:22,090 --> 01:07:23,438
2 more things you could do that you

1962
01:07:23,438 --> 01:07:25,183
might not be thinking about. First of all,

1963
01:07:25,262 --> 01:07:26,690
if you've spent some time on Amazon, and

1964
01:07:26,690 --> 01:07:28,277
you've looked at a book. You've probably noticed

1965
01:07:28,277 --> 01:07:30,197
or actually probably any product. But if you

1966
01:07:30,197 --> 01:07:32,108
look at the book, right underneath that book,

1967
01:07:32,506 --> 01:07:34,656
they will often recommend 2 or 3 other

1968
01:07:34,656 --> 01:07:36,965
books under a heading and says, people also

1969
01:07:36,965 --> 01:07:38,135
bought this

1970
01:07:38,494 --> 01:07:39,394
And so

1971
01:07:39,855 --> 01:07:42,255
if you already have my book, 1 of

1972
01:07:42,255 --> 01:07:44,094
the best ways to get other people to

1973
01:07:44,094 --> 01:07:44,815
see in my book.

1974
01:07:45,550 --> 01:07:47,150
Is to go into 1 of the 3

1975
01:07:47,150 --> 01:07:48,989
categories. And again, you could find them on

1976
01:07:48,989 --> 01:07:50,429
my website or just go to Amazon. You'll

1977
01:07:50,429 --> 01:07:52,190
see in there in that little area I

1978
01:07:52,190 --> 01:07:52,829
just told you about.

1979
01:07:53,883 --> 01:07:56,436
Go buy 1 of the other top 10

1980
01:07:56,436 --> 01:07:56,936
books

1981
01:07:57,474 --> 01:07:59,070
in 1 of those other 3 categories.

1982
01:07:59,628 --> 01:08:01,039
When you buy both books

1983
01:08:01,398 --> 01:08:03,472
that increases the odds that someone else looking

1984
01:08:03,472 --> 01:08:05,387
at 1 of those other books. We'll see

1985
01:08:05,387 --> 01:08:07,380
my book. In my book honestly is really

1986
01:08:07,380 --> 01:08:08,976
not like other books, so I'm not really

1987
01:08:08,976 --> 01:08:10,890
in competition with the with these other books.

1988
01:08:11,381 --> 01:08:12,730
So really it helps both of us at

1989
01:08:12,730 --> 01:08:14,078
that point. If you see 1, you see

1990
01:08:14,078 --> 01:08:15,585
the other. So that's great. And the other

1991
01:08:15,585 --> 01:08:16,617
thing you can do, and I'm not sure

1992
01:08:16,617 --> 01:08:18,996
how much this factors into Amazon's algorithms.

1993
01:08:19,569 --> 01:08:21,483
But I have seen places on Amazon where

1994
01:08:21,483 --> 01:08:23,976
it lists, you know, like most wished for

1995
01:08:24,514 --> 01:08:24,753
books.

1996
01:08:25,631 --> 01:08:27,624
So a very simple thing you can do

1997
01:08:27,624 --> 01:08:29,473
is just add my book whether you've that

1998
01:08:29,473 --> 01:08:31,703
are not. Add my book to your Amazon

1999
01:08:31,703 --> 01:08:33,217
wish list. And you can do it on

2000
01:08:33,217 --> 01:08:34,572
private list wish list, It doesn't have to

2001
01:08:34,572 --> 01:08:36,244
be public. But just adding my book to

2002
01:08:36,244 --> 01:08:38,099
a wish list, I think will also

2003
01:08:39,204 --> 01:08:41,768
Feed into some of Amazon's cryptic algorithms

2004
01:08:42,140 --> 01:08:44,203
to hopefully get it some more attention. So

2005
01:08:44,203 --> 01:08:46,203
a couple more thoughts before we go. First

2006
01:08:46,203 --> 01:08:47,100
of all, I

2007
01:08:47,475 --> 01:08:49,328
I really want this to be honest

2008
01:08:49,782 --> 01:08:52,487
and sincere and as organic as it can

2009
01:08:52,487 --> 01:08:54,097
be. With me asking you to do this.

2010
01:08:54,337 --> 01:08:56,095
Please, you know, don't spam a bunch of

2011
01:08:56,095 --> 01:08:58,731
people about this, You know, don't just write

2012
01:08:58,731 --> 01:09:00,584
review if you've never read the book, be

2013
01:09:00,584 --> 01:09:02,822
honest and, you know, keep the right spirit

2014
01:09:02,822 --> 01:09:04,420
in mind. I I do not wanna try

2015
01:09:04,420 --> 01:09:06,118
to hack the algorithm

2016
01:09:06,578 --> 01:09:09,215
or do any weird search engine optimization.

2017
01:09:10,029 --> 01:09:12,660
Stuff that's not for real. I want this

2018
01:09:12,660 --> 01:09:14,973
to be real. Also as maybe a little

2019
01:09:14,973 --> 01:09:18,163
incentive, to say thank you to recognize people

2020
01:09:18,163 --> 01:09:20,011
that help me with this, I've got a

2021
01:09:20,011 --> 01:09:21,844
couple more things I'm gonna try, but over

2022
01:09:21,844 --> 01:09:23,357
the next, you know, couple weeks or months.

2023
01:09:23,516 --> 01:09:25,985
But basically, when this kind of campaign those

2024
01:09:25,985 --> 01:09:26,224
over.

2025
01:09:26,941 --> 01:09:27,896
I'm gonna look back and say,

2026
01:09:28,548 --> 01:09:30,216
who really contributed? Who helped me do this?

2027
01:09:30,374 --> 01:09:31,327
Who made it impact,

2028
01:09:32,042 --> 01:09:34,185
and that'll be a completely subjective thing on

2029
01:09:34,185 --> 01:09:36,328
my part, but I'm am going to gather

2030
01:09:36,328 --> 01:09:39,125
together those people, and put together a Zoom

2031
01:09:39,125 --> 01:09:40,795
call where we could talk to me live.

2032
01:09:41,033 --> 01:09:42,465
I'll do... I'll talk to you about the

2033
01:09:42,465 --> 01:09:43,578
book a little bit. We'll do a little,

2034
01:09:43,817 --> 01:09:45,089
maybe a Q and a a little, maybe

2035
01:09:45,089 --> 01:09:47,077
an Ama session and ask me anything session.

2036
01:09:47,569 --> 01:09:48,760
And just gotta hang off a little bit

2037
01:09:48,760 --> 01:09:50,268
and, you know, give me a chance to

2038
01:09:50,268 --> 01:09:52,173
say live and in person. Thank you. I

2039
01:09:52,173 --> 01:09:53,841
may also give away some swag or something,

2040
01:09:54,000 --> 01:09:55,349
you know, we'll see. But I do wanna

2041
01:09:55,349 --> 01:09:57,350
make sure that I, you know, say thank

2042
01:09:57,350 --> 01:09:59,896
you for helping me out with this. I'm

2043
01:09:59,896 --> 01:10:02,124
still working on some details, but I'm open

2044
01:10:02,124 --> 01:10:04,671
to ideas. And so all I guess I

2045
01:10:04,671 --> 01:10:06,756
ask is is you know, as you're doing

2046
01:10:06,756 --> 01:10:08,828
whatever it is you're doing, you know, keep

2047
01:10:08,828 --> 01:10:10,822
a record of it somehow, maybe save your

2048
01:10:10,822 --> 01:10:12,895
email, take a screenshot, something like that, and

2049
01:10:12,895 --> 01:10:13,930
just hold onto it for now.

2050
01:10:14,662 --> 01:10:16,172
When the time comes, I'll just ask for

2051
01:10:16,172 --> 01:10:17,682
people who want to be involved in this

2052
01:10:17,682 --> 01:10:18,875
to, you know, send me a note that,

2053
01:10:19,034 --> 01:10:21,101
here's here's what I did to help, and

2054
01:10:21,260 --> 01:10:22,850
I will invite as many of these views

2055
01:10:22,850 --> 01:10:25,091
as I can. To this Zoom party.

2056
01:10:26,043 --> 01:10:28,109
And I'll say thank you there. So that's

2057
01:10:28,109 --> 01:10:30,573
all I'm say about that. Thank you for

2058
01:10:30,573 --> 01:10:31,708
listening to Rush

2059
01:10:32,811 --> 01:10:35,666
go to fd DSD dot me slash book

2060
01:10:35,666 --> 01:10:38,440
surge for the full details. And again, you've

2061
01:10:38,440 --> 01:10:40,597
got 1 week to do this. But if

2062
01:10:40,597 --> 01:10:41,235
you miss that,

2063
01:10:41,952 --> 01:10:43,704
stay tuned. I've got a couple other things

2064
01:10:43,784 --> 01:10:45,139
I might be doing in the next weeks

2065
01:10:45,139 --> 01:10:45,616
and months.

2066
01:10:46,254 --> 01:10:47,448
So it's not your last chance. And by

2067
01:10:47,448 --> 01:10:49,462
the way, everything that I've mentioned today

2068
01:10:49,774 --> 01:10:50,970
help me no matter when you do it.

2069
01:10:51,288 --> 01:10:52,484
I'm just trying to focus it all this

2070
01:10:52,484 --> 01:10:53,998
week, in particular to try to bounce up

2071
01:10:53,998 --> 01:10:54,716
in those rankings.

2072
01:10:55,274 --> 01:10:56,787
So there you go. Thank you so much

2073
01:10:56,787 --> 01:10:58,949
for tuning in. Will Take care of everybody

2074
01:10:58,949 --> 01:11:00,934
stay safe out there. We'll have a new

2075
01:11:00,934 --> 01:11:03,950
show next week. And until then, don't get

2076
01:11:03,950 --> 01:11:05,220
caught with your driver's step.