1
00:00:00,000 --> 00:00:00,958
>> DAVID MALAN: All right.

2
00:00:00,958 --> 00:00:01,470
Welcome.

3
00:00:01,470 --> 00:00:02,140
Hello, everyone.

4
00:00:02,140 --> 00:00:03,550
My name is David Malan.

5
00:00:03,550 --> 00:00:05,720
I'm on the computer science
faculty here at Harvard,

6
00:00:05,720 --> 00:00:08,130
and teach a few courses--
most of them related

7
00:00:08,130 --> 00:00:10,720
to introductory computer
science and higher level

8
00:00:10,720 --> 00:00:12,370
concepts that derive from that.

9
00:00:12,370 --> 00:00:16,660
The next couple of days are not so much
though about building from the ground

10
00:00:16,660 --> 00:00:18,920
up, as we might in a typical
undergraduate course,

11
00:00:18,920 --> 00:00:22,300
but looking at computer science as it
relates to business and to decision

12
00:00:22,300 --> 00:00:24,360
making-- really from
the top down so that we

13
00:00:24,360 --> 00:00:27,170
can accommodate a range of
backgrounds, as you'll soon see,

14
00:00:27,170 --> 00:00:30,994
both less technical and more
technical, and also a number of goals

15
00:00:30,994 --> 00:00:31,660
that folks have.

16
00:00:31,660 --> 00:00:34,680
In fact, we thought we'd start
off by taking a look at some

17
00:00:34,680 --> 00:00:37,210
of the demographics we have here.

18
00:00:37,210 --> 00:00:39,390
But first, let's take a
look at where we're headed.

19
00:00:39,390 --> 00:00:44,280
>> So today, we have four
blocks focused for the day.

20
00:00:44,280 --> 00:00:48,560
First up, we will focus on
privacy, security, and society.

21
00:00:48,560 --> 00:00:51,415
And we'll do this by way of a
couple of case studies of sorts.

22
00:00:51,415 --> 00:00:54,040
Very much in the news of late
has been a certain company called

23
00:00:54,040 --> 00:00:57,580
Apple and a certain agency known
as the FBI, as you might have read.

24
00:00:57,580 --> 00:00:59,890
And we'll use this as an
opportunity to discuss

25
00:00:59,890 --> 00:01:02,470
exactly what some of
the underlying issues

26
00:01:02,470 --> 00:01:06,072
there are, why it's interesting,
what it means technologically,

27
00:01:06,072 --> 00:01:07,780
and use that to
transition more generally

28
00:01:07,780 --> 00:01:11,290
to a discussion about security
and making decisions there on.

29
00:01:11,290 --> 00:01:13,180
>> Two, looking at encryption specifically.

30
00:01:13,180 --> 00:01:16,388
So we'll look a little more technically
at what it means to actually scramble

31
00:01:16,388 --> 00:01:17,580
or encrypt information.

32
00:01:17,580 --> 00:01:19,413
And then we'll take a
look at Dropbox, which

33
00:01:19,413 --> 00:01:22,060
is one of these very popular
file sharing tools these days.

34
00:01:22,060 --> 00:01:26,450
You might use it, or Box, or SkyDrive,
or the more recent incarnation

35
00:01:26,450 --> 00:01:27,680
thereof and so forth.

36
00:01:27,680 --> 00:01:29,950
And we'll take a look at some of
the underlying security and privacy

37
00:01:29,950 --> 00:01:30,742
implications there.

38
00:01:30,742 --> 00:01:33,408
We'll have a break, and then we'll
look at internet technologies

39
00:01:33,408 --> 00:01:35,240
in the latter half of
this morning-- trying

40
00:01:35,240 --> 00:01:39,000
to give you a better sense of how
that thing works that many of you

41
00:01:39,000 --> 00:01:41,860
are connected to at the moment--
certainly use most every day--

42
00:01:41,860 --> 00:01:45,560
and what the implications are there
for performance, for hardware,

43
00:01:45,560 --> 00:01:49,000
for software, and any number of
other attributes, specifically

44
00:01:49,000 --> 00:01:52,000
trying to tease apart a whole bunch
of acronyms that you might have seen

45
00:01:52,000 --> 00:01:53,990
or might even use, but
don't necessarily know

46
00:01:53,990 --> 00:01:56,040
what's going on underneath the hood.

47
00:01:56,040 --> 00:01:59,500
And we'll take a look to at the
process of actually getting a company

48
00:01:59,500 --> 00:02:03,164
or getting a entity online on the
web, and what that actually means.

49
00:02:03,164 --> 00:02:04,580
Then we'll have a break for lunch.

50
00:02:04,580 --> 00:02:07,790
We'll come back and take a look at cloud
computing, and also designing server

51
00:02:07,790 --> 00:02:09,889
architectures more
generally so that you'll

52
00:02:09,889 --> 00:02:13,160
walk out with a better understanding,
hopefully, of this buzzword "cloud

53
00:02:13,160 --> 00:02:14,867
computing," but what it actually means.

54
00:02:14,867 --> 00:02:16,950
And if you're trying to
build a business or you're

55
00:02:16,950 --> 00:02:19,484
trying to expand a business,
exactly what you need to know

56
00:02:19,484 --> 00:02:22,400
and what you need to do in order to
handle increasing numbers of users

57
00:02:22,400 --> 00:02:25,190
online, and what kind of decisions
you have to make around that.

58
00:02:25,190 --> 00:02:28,380
>> And then in the last part of today,
we'll take a look at web development

59
00:02:28,380 --> 00:02:29,330
specifically.

60
00:02:29,330 --> 00:02:31,690
We won't get our hands
too dirty, but I thought

61
00:02:31,690 --> 00:02:34,689
it might be enlightening if we actually
do get our hands a little dirty,

62
00:02:34,689 --> 00:02:40,052
and take a look at something called
HTML, CSS, and an actual server set

63
00:02:40,052 --> 00:02:43,010
up so that you'll create a little
web page for yourself, even if you've

64
00:02:43,010 --> 00:02:43,718
done this before.

65
00:02:43,718 --> 00:02:47,010
But we'll talk about what the
interesting ideas are underlying that

66
00:02:47,010 --> 00:02:48,920
and what actually is
happening every time you

67
00:02:48,920 --> 00:02:54,470
go to Google Apps, or Facebook, or
any number of other web-based tools.

68
00:02:54,470 --> 00:02:57,670
>> Tomorrow, meanwhile,
we'll transition to a look

69
00:02:57,670 --> 00:02:59,890
in the morning at
computational thinking--

70
00:02:59,890 --> 00:03:03,090
a fancy way of describing
how a computer might think

71
00:03:03,090 --> 00:03:05,710
or a human versed in
computing might think--

72
00:03:05,710 --> 00:03:09,330
a little more methodical, a little
more algorithmic, as we might say.

73
00:03:09,330 --> 00:03:13,630
And we won't go too deeply
into programming per se,

74
00:03:13,630 --> 00:03:15,590
but we'll focus on some
of the tenets that you

75
00:03:15,590 --> 00:03:18,800
see in programming and computer
science-- abstraction, algorithms,

76
00:03:18,800 --> 00:03:21,751
and how you represent data, and
why that's actually interesting.

77
00:03:21,751 --> 00:03:23,750
We will take somewhat of
a look at reprogramming

78
00:03:23,750 --> 00:03:25,350
in the latter half of tomorrow morning.

79
00:03:25,350 --> 00:03:27,266
We'll get your hands a
little dirty with that,

80
00:03:27,266 --> 00:03:30,150
but only so that we have some
context for talking about some

81
00:03:30,150 --> 00:03:34,080
of the terms of art that an engineer or
a programmer might actually use, things

82
00:03:34,080 --> 00:03:37,959
you might hear or see on a whiteboard
when engineers are designing something.

83
00:03:37,959 --> 00:03:40,000
In the latter half of
tomorrow, we'll take a look

84
00:03:40,000 --> 00:03:42,220
at what might be called
technology stacks.

85
00:03:42,220 --> 00:03:44,170
In other words, most
people today don't really

86
00:03:44,170 --> 00:03:46,230
sit down with an empty
screen in front of them

87
00:03:46,230 --> 00:03:48,900
and start building some application
or building some website.

88
00:03:48,900 --> 00:03:50,690
You stand on the
shoulders of others using

89
00:03:50,690 --> 00:03:54,257
things called frameworks and libraries,
many of them open source these days.

90
00:03:54,257 --> 00:03:56,340
So we'll give you a sense
of what all that's about

91
00:03:56,340 --> 00:04:01,790
and how you go about designing software
and choosing those ingredients.

92
00:04:01,790 --> 00:04:05,189
>> And then we'll conclude with a look
at web programming specifically

93
00:04:05,189 --> 00:04:06,980
and some of the
technologies related there,

94
00:04:06,980 --> 00:04:11,200
too-- things like databases,
open source, or commercial APIs,

95
00:04:11,200 --> 00:04:14,200
or application programming
interfaces, and then one such language

96
00:04:14,200 --> 00:04:15,570
that you might use with that.

97
00:04:15,570 --> 00:04:18,800
So it'll be a mix of conceptual
introductions, a mix of hands on,

98
00:04:18,800 --> 00:04:21,019
and a mix for discussion throughout.

99
00:04:21,019 --> 00:04:25,500
>> But before we do that, let me give you
an answer to a couple of the questions

100
00:04:25,500 --> 00:04:26,834
that everyone here was asked.

101
00:04:26,834 --> 00:04:29,000
How would you describe your
comfort with technology?

102
00:04:29,000 --> 00:04:30,250
We have a bit of a range here.

103
00:04:30,250 --> 00:04:34,342
So six people said somewhat comfortable,
five said very, and two said not very.

104
00:04:34,342 --> 00:04:36,800
So that should lend itself to
some interesting discussions.

105
00:04:36,800 --> 00:04:41,630
>> And please, at any point, whether you
are in the not very or very categories,

106
00:04:41,630 --> 00:04:44,430
do push back if either
I'm assuming too much

107
00:04:44,430 --> 00:04:46,700
or speaking at too high of a level.

108
00:04:46,700 --> 00:04:47,680
Do bring me back down.

109
00:04:47,680 --> 00:04:50,260
And conversely, if you'd
like to get a little more

110
00:04:50,260 --> 00:04:53,580
into the weeds with some topic
technically, by all means push on that.

111
00:04:53,580 --> 00:04:57,280
I'm to happy to answer down
to 0s and 1s if need be.

112
00:04:57,280 --> 00:04:59,620
>> Do you have any programming
experience in any language?

113
00:04:59,620 --> 00:05:04,020
Just to calibrate, almost everyone
has no prior programming experience,

114
00:05:04,020 --> 00:05:04,730
which is great.

115
00:05:04,730 --> 00:05:08,672
And even for those that do, we won't
spend too much time actually teaching

116
00:05:08,672 --> 00:05:10,880
how to program, but rather
just giving you a taste so

117
00:05:10,880 --> 00:05:13,890
that we can then move from there
and talk at a higher level about why

118
00:05:13,890 --> 00:05:16,360
some of those concepts are interesting.

119
00:05:16,360 --> 00:05:18,884
>> This and more will all
be available online.

120
00:05:18,884 --> 00:05:21,300
In fact, if there's one URL
you want to keep open in a tab

121
00:05:21,300 --> 00:05:23,770
throughout today and tomorrow, you
might want to go to this one here.

122
00:05:23,770 --> 00:05:25,150
And that's a copy of the slides.

123
00:05:25,150 --> 00:05:27,774
And any changes we make over the
course of today or discussions

124
00:05:27,774 --> 00:05:30,200
that we annotate on the slides,
they'll be there instantly

125
00:05:30,200 --> 00:05:33,330
if you just reload your browser.

126
00:05:33,330 --> 00:05:36,160
So I'll give you a
moment to jot that down,

127
00:05:36,160 --> 00:05:40,150
and you'll then be able
to see exactly what I see.

128
00:05:40,150 --> 00:05:44,230
>> But before we forge ahead, I
thought it might be helpful,

129
00:05:44,230 --> 00:05:46,390
especially since we're
an intimate group, just

130
00:05:46,390 --> 00:05:49,180
to get to know each other a
little bit and perhaps say

131
00:05:49,180 --> 00:05:52,437
where you're from, or what you do, and
what you're hoping to get out of today

132
00:05:52,437 --> 00:05:54,770
and tomorrow, ideally so that
you might find one or more

133
00:05:54,770 --> 00:05:57,840
like minded spirits or folks to
talk to during break or lunch.

134
00:05:57,840 --> 00:06:00,060
And I'll jump us around
somewhat randomly.

135
00:06:00,060 --> 00:06:01,740
Arwa, you'd like to say hello, first?

136
00:06:01,740 --> 00:06:02,406
>> AUDIENCE: Hello.

137
00:06:02,406 --> 00:06:03,740
Good morning, everyone.

138
00:06:03,740 --> 00:06:07,216
My name is Arwa. [INAUDIBLE].

139
00:06:07,216 --> 00:06:10,632
I work at at my sector like
banking, business [INAUDIBLE].

140
00:06:10,632 --> 00:06:22,464

141
00:06:22,464 --> 00:06:23,130
DAVID MALAN: OK.

142
00:06:23,130 --> 00:06:23,510
Wonderful.

143
00:06:23,510 --> 00:06:24,009
Welcome.

144
00:06:24,009 --> 00:06:24,715
Andrew.

145
00:06:24,715 --> 00:06:25,340
AUDIENCE: Yeah.

146
00:06:25,340 --> 00:06:26,040
Hi, everyone.

147
00:06:26,040 --> 00:06:27,492
I'm Andrew [INAUDIBLE].

148
00:06:27,492 --> 00:06:29,670
So I work for a technology
company, Red Hat,

149
00:06:29,670 --> 00:06:32,505
which is a big open source company.

150
00:06:32,505 --> 00:06:34,404
I have a business
background so [INAUDIBLE]

151
00:06:34,404 --> 00:06:39,008
get more versed into making
solution oriented investments,

152
00:06:39,008 --> 00:06:41,640
I just need to know what
people are talking about.

153
00:06:41,640 --> 00:06:44,305
So I lead our global partner operations.

154
00:06:44,305 --> 00:06:46,740
I've been doing that
for about five years.

155
00:06:46,740 --> 00:06:48,201
Your overview is fantastic.

156
00:06:48,201 --> 00:06:50,368
I'm really looking to pick
up all those [INAUDIBLE].

157
00:06:50,368 --> 00:06:51,325
DAVID MALAN: Wonderful.

158
00:06:51,325 --> 00:06:52,200
Glad to have you.

159
00:06:52,200 --> 00:06:53,197
Chris.

160
00:06:53,197 --> 00:06:54,191
>> AUDIENCE: Good morning.

161
00:06:54,191 --> 00:06:55,185
My name is Chris Pratt.

162
00:06:55,185 --> 00:06:57,173
I work for a company called [INAUDIBLE].

163
00:06:57,173 --> 00:07:00,155

164
00:07:00,155 --> 00:07:03,137
It's a family business, so I
do a lot of different projects.

165
00:07:03,137 --> 00:07:11,089
And right now, I'm focused on technology
initiatives and managing our IT staff.

166
00:07:11,089 --> 00:07:15,296
So I'm here to get a more high
level and broad understanding

167
00:07:15,296 --> 00:07:20,525
of the types of things
that [INAUDIBLE] is

168
00:07:20,525 --> 00:07:25,762
doing and familiar with so I can help
them make the decisions [INAUDIBLE].

169
00:07:25,762 --> 00:07:26,720
DAVID MALAN: Wonderful.

170
00:07:26,720 --> 00:07:27,466
Welcome aboard.

171
00:07:27,466 --> 00:07:29,330
Olivier, is it?

172
00:07:29,330 --> 00:07:29,953
>> AUDIENCE: Yes.

173
00:07:29,953 --> 00:07:34,507
So I'm French living in Switzerland
working for [INAUDIBLE].

174
00:07:34,507 --> 00:07:41,270
It's a [INAUDIBLE] corporations.

175
00:07:41,270 --> 00:07:45,770
So we're collecting money when
there's a disaster and everything.

176
00:07:45,770 --> 00:07:47,770
And I'm teaching some strategies there.

177
00:07:47,770 --> 00:07:55,004
>> So I have to work on [INAUDIBLE]
digital projects, but also

178
00:07:55,004 --> 00:07:57,439
quite technological projects.

179
00:07:57,439 --> 00:08:01,822
So the idea for me is really to
be able to make better decisions

180
00:08:01,822 --> 00:08:04,762
and being better informed of
what I'm really [INAUDIBLE].

181
00:08:04,762 --> 00:08:05,720
DAVID MALAN: Wonderful.

182
00:08:05,720 --> 00:08:06,219
Welcome.

183
00:08:06,219 --> 00:08:09,146
And Roman or Roman, is it?

184
00:08:09,146 --> 00:08:10,437
AUDIENCE: I'm from [INAUDIBLE].

185
00:08:10,437 --> 00:08:13,888

186
00:08:13,888 --> 00:08:16,353
And I'm responsible for the [INAUDIBLE].

187
00:08:16,353 --> 00:08:19,311
And in the team we-- we're
a cross functional team

188
00:08:19,311 --> 00:08:21,776
so we work with engineers.

189
00:08:21,776 --> 00:08:25,530
And what I'm looking forward to is
being able to communicate better

190
00:08:25,530 --> 00:08:27,505
with engineers.

191
00:08:27,505 --> 00:08:28,005
[INAUDIBLE]

192
00:08:28,005 --> 00:08:31,780

193
00:08:31,780 --> 00:08:32,870
>> DAVID MALAN: Wonderful.

194
00:08:32,870 --> 00:08:34,080
And Karina.

195
00:08:34,080 --> 00:08:36,580
>> AUDIENCE: I'm Karina from Montreal.

196
00:08:36,580 --> 00:08:47,630
I'm on [INAUDIBLE] of
province of Quebec.

197
00:08:47,630 --> 00:08:49,726
Sorry, for my English.

198
00:08:49,726 --> 00:08:56,520
And I'm here to better understand what
my programmer or supplier explained

199
00:08:56,520 --> 00:08:59,246
to me.

200
00:08:59,246 --> 00:08:59,745
[INAUDIBLE]

201
00:08:59,745 --> 00:09:04,334

202
00:09:04,334 --> 00:09:05,000
DAVID MALAN: Oh.

203
00:09:05,000 --> 00:09:05,506
Wonderful.

204
00:09:05,506 --> 00:09:07,630
Well, if I ever speak too
quickly, do slow me down.

205
00:09:07,630 --> 00:09:08,720
And I'm happy to repeat.

206
00:09:08,720 --> 00:09:09,595
>> AUDIENCE: [INAUDIBLE]

207
00:09:09,595 --> 00:09:12,130

208
00:09:12,130 --> 00:09:12,880
DAVID MALAN: Sure.

209
00:09:12,880 --> 00:09:13,680
No worries.

210
00:09:13,680 --> 00:09:14,707
And Nikisa, is it?

211
00:09:14,707 --> 00:09:15,290
AUDIENCE: Yes.

212
00:09:15,290 --> 00:09:15,790
Thank you.

213
00:09:15,790 --> 00:09:18,830
My name is Nikisa, and I'm [INAUDIBLE].

214
00:09:18,830 --> 00:09:19,330
[INAUDIBLE]

215
00:09:19,330 --> 00:09:24,574

216
00:09:24,574 --> 00:09:25,846
I am myself [INAUDIBLE].

217
00:09:25,846 --> 00:09:30,560

218
00:09:30,560 --> 00:09:42,825
So I'm always confused with [INAUDIBLE]
whatever you are really [INAUDIBLE].

219
00:09:42,825 --> 00:10:01,682

220
00:10:01,682 --> 00:10:02,430
>> DAVID MALAN: OK.

221
00:10:02,430 --> 00:10:03,030
Wonderful.

222
00:10:03,030 --> 00:10:03,842
Welcome.

223
00:10:03,842 --> 00:10:05,060
Victoria.

224
00:10:05,060 --> 00:10:06,030
>> AUDIENCE: I'm Victoria.

225
00:10:06,030 --> 00:10:07,485
I live in Czech Republic.

226
00:10:07,485 --> 00:10:09,414
I work for [INAUDIBLE] Enterprise.

227
00:10:09,414 --> 00:10:11,122
And even though it is
an IT company, it's

228
00:10:11,122 --> 00:10:14,074
possible that in an IT
company [INAUDIBLE].

229
00:10:14,074 --> 00:10:16,240
So I'm focused on business
development, and whenever

230
00:10:16,240 --> 00:10:18,781
I go to a customer meeting, I
have to take a technical person

231
00:10:18,781 --> 00:10:21,017
with me because my
customer asks questions

232
00:10:21,017 --> 00:10:24,736
about technical side of
the story. [INAUDIBLE].

233
00:10:24,736 --> 00:10:27,110
They talk to each other, but
then I have no understanding

234
00:10:27,110 --> 00:10:28,352
of what they're discussing.

235
00:10:28,352 --> 00:10:30,060
So I'd like to get a
better understanding

236
00:10:30,060 --> 00:10:35,800
because I think it would
help myself [INAUDIBLE]

237
00:10:35,800 --> 00:10:37,800
with my relationship with
the customers as well.

238
00:10:37,800 --> 00:10:39,320
>> DAVID MALAN: And it's a good
point for me to chime in.

239
00:10:39,320 --> 00:10:41,350
There's only so much we'll be
able to do in just two days.

240
00:10:41,350 --> 00:10:43,450
But among the goals,
I would hope, is that,

241
00:10:43,450 --> 00:10:45,800
after the next couple of
days, at least more words

242
00:10:45,800 --> 00:10:47,970
will look familiar as
you're poking around online.

243
00:10:47,970 --> 00:10:50,980
And you'll have a better
sense of what to Google,

244
00:10:50,980 --> 00:10:52,910
or what words actually
mean something and what

245
00:10:52,910 --> 00:10:55,620
might be fluffy marketing
speak so that, over time, you

246
00:10:55,620 --> 00:10:58,670
can build up that comfort and
hopefully displace the person that

247
00:10:58,670 --> 00:11:00,410
has to tag along each time.

248
00:11:00,410 --> 00:11:01,726
Ben.

249
00:11:01,726 --> 00:11:03,225
AUDIENCE: My name's Ben [INAUDIBLE].

250
00:11:03,225 --> 00:11:06,207
I'm a technology transaction
attorney. [INAUDIBLE].

251
00:11:06,207 --> 00:11:10,310
And I'm really here to just get
a better understanding of what

252
00:11:10,310 --> 00:11:18,231
CTOs and Engineers at [INAUDIBLE] legal
side of structuring things [INAUDIBLE].

253
00:11:18,231 --> 00:11:24,072

254
00:11:24,072 --> 00:11:25,030
DAVID MALAN: Wonderful.

255
00:11:25,030 --> 00:11:26,600
AUDIENCE: And Dan.

256
00:11:26,600 --> 00:11:28,220
Hi, everybody.

257
00:11:28,220 --> 00:11:29,220
My name's Dan.

258
00:11:29,220 --> 00:11:30,230
I live local here.

259
00:11:30,230 --> 00:11:32,100
I'm from Andover.

260
00:11:32,100 --> 00:11:36,200
I work locally at a software
company, Kronos Incorporated.

261
00:11:36,200 --> 00:11:39,090
Been in software over
20 years and [INAUDIBLE]

262
00:11:39,090 --> 00:11:42,470
marketing and development type jobs.

263
00:11:42,470 --> 00:11:49,176
For the last five years, I've managed
a team of technical cloud consultants

264
00:11:49,176 --> 00:11:50,880
in presales fashion.

265
00:11:50,880 --> 00:11:53,880
>> So I picked up a lot
of concepts on the way.

266
00:11:53,880 --> 00:11:56,620
And so I do a lot of
technical discussions.

267
00:11:56,620 --> 00:11:57,870
But I can only take it so far.

268
00:11:57,870 --> 00:12:01,420
Similar to Victoria,
lots of times I get lost

269
00:12:01,420 --> 00:12:03,348
and need to call in a technical person.

270
00:12:03,348 --> 00:12:06,720
So I'm just looking to string
a lot of technology jargon

271
00:12:06,720 --> 00:12:09,738
together so I get a
better understanding so I

272
00:12:09,738 --> 00:12:11,462
can have more informed conversations.

273
00:12:11,462 --> 00:12:12,420
DAVID MALAN: Excellent.

274
00:12:12,420 --> 00:12:14,410
Well, ultimately, we can
steer the next couple of days

275
00:12:14,410 --> 00:12:15,890
in any direction folks would like.

276
00:12:15,890 --> 00:12:18,040
We have a straw man for
both today and tomorrow.

277
00:12:18,040 --> 00:12:21,020
But by all means, feel free to
steer us either during the session

278
00:12:21,020 --> 00:12:22,965
or during breaks or lunch
if there's something

279
00:12:22,965 --> 00:12:24,340
you'd like to get off your chest.

280
00:12:24,340 --> 00:12:26,759
And let me emphasize, there
really is no dumb question.

281
00:12:26,759 --> 00:12:28,550
And if you feel like
your question is dumb,

282
00:12:28,550 --> 00:12:32,319
by all means just ask me more quietly
during breaks, or lunch, or the like.

283
00:12:32,319 --> 00:12:35,360
But rest assured, we seem to be in
very good company-- very mixed company

284
00:12:35,360 --> 00:12:37,630
here, both internationally
and technically.

285
00:12:37,630 --> 00:12:41,500
So feel free to share as
comfortably as you'd like.

286
00:12:41,500 --> 00:12:44,050
>> So why don't we take a
look, again, in this context

287
00:12:44,050 --> 00:12:48,520
of privacy, security, and society
at this particular case involving

288
00:12:48,520 --> 00:12:49,990
Apple and the FBI.

289
00:12:49,990 --> 00:12:52,640
And you might be generally
familiar with this case.

290
00:12:52,640 --> 00:12:54,550
It's hard to escape
mention of it these days.

291
00:12:54,550 --> 00:12:59,150
>> Out of curiosity, how
many of you have iPhones?

292
00:12:59,150 --> 00:13:00,760
Almost everyone.

293
00:13:00,760 --> 00:13:02,230
And you have an Android phone?

294
00:13:02,230 --> 00:13:06,460
So fortunately, even though
this is a little biased

295
00:13:06,460 --> 00:13:10,280
toward iPhone specifically, the
reality is the Android operating system

296
00:13:10,280 --> 00:13:13,610
by Google has so many of the similar
features to what Apple is doing.

297
00:13:13,610 --> 00:13:16,290
>> They simply happen to be
in the spotlight right now,

298
00:13:16,290 --> 00:13:19,042
and they've been particularly
on the cutting edge when

299
00:13:19,042 --> 00:13:21,000
it comes to actually
locking down these devices

300
00:13:21,000 --> 00:13:23,250
more and more with
each iteration of iOS,

301
00:13:23,250 --> 00:13:26,660
the operating system that
actually runs on Apple's devices.

302
00:13:26,660 --> 00:13:29,370
So why don't we take a
look here just to set

303
00:13:29,370 --> 00:13:34,430
the stage at what the actual issue is.

304
00:13:34,430 --> 00:13:37,310
So what's going on with Apple
and the FBI to the extent

305
00:13:37,310 --> 00:13:40,403
that you're familiar with the issue?

306
00:13:40,403 --> 00:13:43,970
>> AUDIENCE: The FBI wants to get access to
the data, which is encrypted by Apple.

307
00:13:43,970 --> 00:13:48,220
>> DAVID MALAN: Exactly, so the FBI wants
to get access to data that's encrypted.

308
00:13:48,220 --> 00:13:49,220
So first, step back.

309
00:13:49,220 --> 00:13:52,914
What does it mean for data to be
encrypted, just as a quick definition?

310
00:13:52,914 --> 00:13:56,706
>> AUDIENCE: Somehow secure that
people won't have such easy access

311
00:13:56,706 --> 00:13:59,831
to it [INAUDIBLE].

312
00:13:59,831 --> 00:14:00,580
DAVID MALAN: Yeah.

313
00:14:00,580 --> 00:14:04,130
Exactly So it's some way of obscuring
information so that no one else can,

314
00:14:04,130 --> 00:14:06,350
in theory, access that information.

315
00:14:06,350 --> 00:14:09,480
And so you can just casually
think of it as scrambling.

316
00:14:09,480 --> 00:14:12,549
So if it's an English word
or an English paragraph,

317
00:14:12,549 --> 00:14:15,340
you might just jumble the words up
so that someone might look at it

318
00:14:15,340 --> 00:14:16,357
and it sees nonsense.

319
00:14:16,357 --> 00:14:18,690
But hopefully, there's a way
to rearrange those letters.

320
00:14:18,690 --> 00:14:20,230
>> Now, in reality, it's
much more secure than that

321
00:14:20,230 --> 00:14:22,021
because someone who's
simply diligent could

322
00:14:22,021 --> 00:14:23,930
unscramble the words
with high probability

323
00:14:23,930 --> 00:14:25,490
and figure out what a sentence says.

324
00:14:25,490 --> 00:14:27,573
And in reality, at the end
of the day, all of this

325
00:14:27,573 --> 00:14:29,749
is happening at a very
low level-- 0s and 1s.

326
00:14:29,749 --> 00:14:32,290
And tomorrow morning, we'll talk
about computational thinking

327
00:14:32,290 --> 00:14:35,580
and what it means for data to
be implemented or represented

328
00:14:35,580 --> 00:14:37,109
with just 0s and 1s.

329
00:14:37,109 --> 00:14:38,900
But for today's purposes,
let's just assume

330
00:14:38,900 --> 00:14:41,275
that you have things like
emails, and photos, and videos,

331
00:14:41,275 --> 00:14:43,470
and all of that on a iPhone
or an Android device.

332
00:14:43,470 --> 00:14:46,840
And somehow, that data
is ideally scrambled.

333
00:14:46,840 --> 00:14:51,620
And so there's a suspect in this
particular case, San Bernardino,

334
00:14:51,620 --> 00:14:55,800
where they have the suspect's phone,
and they want to get data off of it.

335
00:14:55,800 --> 00:15:00,780
But in this case, Apple has essentially
said no to some things and yes

336
00:15:00,780 --> 00:15:01,550
to other things.

337
00:15:01,550 --> 00:15:04,130
So they've said yes to a few things
in a manner consistent with what

338
00:15:04,130 --> 00:15:06,588
a lot of US companies would do
when subpoenaed or the like.

339
00:15:06,588 --> 00:15:10,290
They've provided, for instance, the
authorities with the iCloud backup.

340
00:15:10,290 --> 00:15:12,940
So if I'm familiar, iCloud
is this cloud base--

341
00:15:12,940 --> 00:15:17,260
and we'll come back to cloud computing--
this nebulously defined cloud

342
00:15:17,260 --> 00:15:19,270
based service where it
just backs up your data.

343
00:15:19,270 --> 00:15:22,620
And it turns out that you can
access data there unencrypted.

344
00:15:22,620 --> 00:15:25,350
So it's unscrambled when it's
actually being backed up there.

345
00:15:25,350 --> 00:15:26,770
And so Apple's turned that over.

346
00:15:26,770 --> 00:15:28,520
But unfortunately, the
suspect in question

347
00:15:28,520 --> 00:15:32,050
seems to have disabled automatic
iCloud backup some weeks

348
00:15:32,050 --> 00:15:35,080
prior to the FBI obtaining
this particular iPhone.

349
00:15:35,080 --> 00:15:38,570
So there's a few weeks of potential
data that lives on the phone,

350
00:15:38,570 --> 00:15:39,920
but not in iCloud.

351
00:15:39,920 --> 00:15:43,359
And so the FBI wants to actually look
at what's on that particular phone.

352
00:15:43,359 --> 00:15:45,400
Unfortunately, the phone,
like many of ours here,

353
00:15:45,400 --> 00:15:47,050
is protected with the passcode.

354
00:15:47,050 --> 00:15:50,640
And how long are these passcodes
typically-- whether on your phone

355
00:15:50,640 --> 00:15:51,470
or in general?

356
00:15:51,470 --> 00:15:52,420
>> AUDIENCE: Four.

357
00:15:52,420 --> 00:15:52,666
>> DAVID MALAN: Yeah.

358
00:15:52,666 --> 00:15:53,600
So often four digits.

359
00:15:53,600 --> 00:15:55,580
They've started with
newer versions of iOS

360
00:15:55,580 --> 00:15:57,292
to make these passcodes a little longer.

361
00:15:57,292 --> 00:15:59,000
And let's just put
that into perspective.

362
00:15:59,000 --> 00:16:02,010
So if it's a four digit
passcode, that's pretty good.

363
00:16:02,010 --> 00:16:03,760
That's comparable to
what many people have

364
00:16:03,760 --> 00:16:06,510
on their ATMs or their debit cards.

365
00:16:06,510 --> 00:16:08,980
What's the implication for security?

366
00:16:08,980 --> 00:16:10,272
>> Well, let's take a step back.

367
00:16:10,272 --> 00:16:11,980
If you have a four
digit code-- and let's

368
00:16:11,980 --> 00:16:14,230
let's start to ourselves
even before tomorrow morning.

369
00:16:14,230 --> 00:16:15,430
Think computationally.

370
00:16:15,430 --> 00:16:17,060
It's a four digit code.

371
00:16:17,060 --> 00:16:21,390
How would you, as a human off the
street, not necessarily a technophile,

372
00:16:21,390 --> 00:16:26,240
characterize just how secure an
iPhone is if it's using a four digit

373
00:16:26,240 --> 00:16:28,490
passcode-- 0s through 9s.

374
00:16:28,490 --> 00:16:34,006
How do you begin to quantify
the security of an iPhone then?

375
00:16:34,006 --> 00:16:34,630
AUDIENCE: Five?

376
00:16:34,630 --> 00:16:35,379
DAVID MALAN: Five?

377
00:16:35,379 --> 00:16:37,112
And what you mean by five?

378
00:16:37,112 --> 00:16:43,041
>> AUDIENCE: [INAUDIBLE] this technology--
it's easy to access trying from 1001

379
00:16:43,041 --> 00:16:44,790
[INAUDIBLE].

380
00:16:44,790 --> 00:16:46,238
>> DAVID MALAN: OK.

381
00:16:46,238 --> 00:16:49,112
>> AUDIENCE: Try 111, 000, [INAUDIBLE].

382
00:16:49,112 --> 00:16:53,244
And if I [INAUDIBLE] my computer
so many times [INAUDIBLE].

383
00:16:53,244 --> 00:16:54,160
DAVID MALAN: Ah, good.

384
00:16:54,160 --> 00:16:58,830
So already, if we've defined the problem
scenario as this device is secure

385
00:16:58,830 --> 00:17:02,260
because it has a four digit
passcode, an attack on that phone

386
00:17:02,260 --> 00:17:04,980
would simply be to try
all possible numbers.

387
00:17:04,980 --> 00:17:07,210
You might just start 0 0 0 0.

388
00:17:07,210 --> 00:17:10,510
And frighteningly, that is the
default passcode on a lot of devices

389
00:17:10,510 --> 00:17:11,069
these days.

390
00:17:11,069 --> 00:17:13,109
In fact, as an aside,
if you have any device

391
00:17:13,109 --> 00:17:15,579
that supports a wireless
technology called Bluetooth,

392
00:17:15,579 --> 00:17:18,869
the default passcode
very often is 0 0 0 0.

393
00:17:18,869 --> 00:17:23,899
Or maybe, if it's a more secure
device, 0 0 0 0 0-- one additional 0.

394
00:17:23,899 --> 00:17:26,690
So when in doubt, if you need to
get into some device, start there.

395
00:17:26,690 --> 00:17:30,350
>> But of course, if the iPhone shakes or
whatnot, and says, nope, that's not it,

396
00:17:30,350 --> 00:17:34,030
what number might you try after 0 0 0 0?

397
00:17:34,030 --> 00:17:35,390
1 1 1 1.

398
00:17:35,390 --> 00:17:37,010
2 2 2 2.

399
00:17:37,010 --> 00:17:38,410
7 7 7 7-- that's yours?

400
00:17:38,410 --> 00:17:39,800
OK.

401
00:17:39,800 --> 00:17:42,970
You might just brute force,
as a computer scientist

402
00:17:42,970 --> 00:17:44,750
says-- try all possible values.

403
00:17:44,750 --> 00:17:46,740
>> So let's steer back to
the original question.

404
00:17:46,740 --> 00:17:48,020
How secure is an iPhone?

405
00:17:48,020 --> 00:17:52,115
Someone off the street might say very
secure, or not very secure, or medium

406
00:17:52,115 --> 00:17:53,740
secure, but that's kind of meaningless.

407
00:17:53,740 --> 00:17:56,780
It would be nice if we could
ascribe something more quantitative,

408
00:17:56,780 --> 00:17:57,620
even if its numbers.

409
00:17:57,620 --> 00:18:02,184
We don't need fancy math, but just some
numerical estimate or qualification

410
00:18:02,184 --> 00:18:02,850
of the security.

411
00:18:02,850 --> 00:18:04,940
>> So if you've got a four
digit passcode, can we

412
00:18:04,940 --> 00:18:08,440
begin to ascribe some kind
of numeric rating to it?

413
00:18:08,440 --> 00:18:09,630
How secure is it?

414
00:18:09,630 --> 00:18:10,714
>> AUDIENCE: 1 out of 10,000.

415
00:18:10,714 --> 00:18:11,463
DAVID MALAN: Yeah.

416
00:18:11,463 --> 00:18:12,480
So 1 out of 10,000.

417
00:18:12,480 --> 00:18:15,608
Where do you get the 10,000 from?

418
00:18:15,608 --> 00:18:18,036
>> AUDIENCE: All possibilities [INAUDIBLE].

419
00:18:18,036 --> 00:18:19,160
DAVID MALAN: Yeah, exactly.

420
00:18:19,160 --> 00:18:24,260
If you've got a 4 digit code, you can
have 0 0 0 0, or you can have 9 9 9 9,

421
00:18:24,260 --> 00:18:24,760
maximally.

422
00:18:24,760 --> 00:18:27,370
And so that's 10,000 possibilities.

423
00:18:27,370 --> 00:18:28,875
So that seems pretty big.

424
00:18:28,875 --> 00:18:31,000
And it would certainly take
a human quite some time

425
00:18:31,000 --> 00:18:32,370
to try all of those codes.

426
00:18:32,370 --> 00:18:34,920
>> And so suppose, I, during lunch
swiped one of your iPhones

427
00:18:34,920 --> 00:18:36,370
and you have a four digit code.

428
00:18:36,370 --> 00:18:39,317
If I had enough time, maybe
I could type in 0 0 0 0.

429
00:18:39,317 --> 00:18:40,650
And then it shakes and says, no.

430
00:18:40,650 --> 00:18:45,790
0 0 0 1, 0 0 0 2, 0 0 3, and
maybe I can do 1 per second.

431
00:18:45,790 --> 00:18:47,640
So that's 10,000 seconds.

432
00:18:47,640 --> 00:18:50,470
So how long would it take me
in the end to actually get

433
00:18:50,470 --> 00:18:56,070
to decrypting or hacking into
someone's iPhone, given these numbers?

434
00:18:56,070 --> 00:18:57,960
And we'll play with a few perhaps here.

435
00:18:57,960 --> 00:19:00,770
>> Let me go ahead and pull up
overkill of a calculator.

436
00:19:00,770 --> 00:19:05,530
So if it's 10,000 seconds, there
are 60 seconds in a minute,

437
00:19:05,530 --> 00:19:07,240
and there are 60 minutes in an hour.

438
00:19:07,240 --> 00:19:08,870
So it's like 2.7 hours.

439
00:19:08,870 --> 00:19:12,690
So I have to miss the afternoon
sessions, if I started during lunch.

440
00:19:12,690 --> 00:19:16,280
But it would only take me 2.7 hours
to try getting into your iPhone.

441
00:19:16,280 --> 00:19:18,470
>> Now, you might be
familiar with mechanisms

442
00:19:18,470 --> 00:19:22,470
that Apple and soon probably other
companies use to defend against this.

443
00:19:22,470 --> 00:19:25,267
This does not seem or
feel very secure anymore.

444
00:19:25,267 --> 00:19:27,850
And we'll come back in just a
bit to do one more introduction,

445
00:19:27,850 --> 00:19:31,330
unless we feel omitted.

446
00:19:31,330 --> 00:19:34,120
What can we do to make this more secure?

447
00:19:34,120 --> 00:19:35,510
10,000 feels like a lot.

448
00:19:35,510 --> 00:19:39,484
But 2.7 hours does not
really feel like that long.

449
00:19:39,484 --> 00:19:42,650
AUDIENCE: Doesn't it get locked after
three attempts or something like that?

450
00:19:42,650 --> 00:19:44,090
DAVID MALAN: Ah, maybe it does.

451
00:19:44,090 --> 00:19:45,840
In fact, hopefully not
three, because even

452
00:19:45,840 --> 00:19:48,179
I goof on my passcode
three or more times.

453
00:19:48,179 --> 00:19:49,720
So there is typically some threshold.

454
00:19:49,720 --> 00:19:52,320
And I believe in iOS's case,
the default is actually 10.

455
00:19:52,320 --> 00:19:53,538
But similarly--

456
00:19:53,538 --> 00:19:55,130
>> AUDIENCE: [INAUDIBLE]

457
00:19:55,130 --> 00:19:56,630
DAVID MALAN: --similarly reasonable.

458
00:19:56,630 --> 00:19:59,614
So what does that mean-- so
what happens after 10 tries

459
00:19:59,614 --> 00:20:00,780
or whatever number of tries?

460
00:20:00,780 --> 00:20:01,660
>> AUDIENCE: It gets locked.

461
00:20:01,660 --> 00:20:01,990
>> DAVID MALAN: Yeah.

462
00:20:01,990 --> 00:20:03,890
So the phone maybe locks itself down.

463
00:20:03,890 --> 00:20:05,040
>> AUDIENCE: Time delay.

464
00:20:05,040 --> 00:20:05,480
>> DAVID MALAN: Time delay.

465
00:20:05,480 --> 00:20:06,600
Would do you mean by time delay?

466
00:20:06,600 --> 00:20:09,600
>> AUDIENCE: It'll lock the phone for
five minutes, and after five minutes,

467
00:20:09,600 --> 00:20:10,580
you can try again.

468
00:20:10,580 --> 00:20:10,850
>> DAVID MALAN: All right.

469
00:20:10,850 --> 00:20:13,630
But that doesn't feel like it's
solving the problem, right?

470
00:20:13,630 --> 00:20:16,493
Can't I just come back 5 minutes
later and continue hacking on it?

471
00:20:16,493 --> 00:20:17,460
>> AUDIENCE: Yes.

472
00:20:17,460 --> 00:20:18,154
>> DAVID MALAN: OK.

473
00:20:18,154 --> 00:20:20,542
>> AUDIENCE: But after you try
again, it goes to 10 minutes.

474
00:20:20,542 --> 00:20:21,208
DAVID MALAN: Ah.

475
00:20:21,208 --> 00:20:22,660
AUDIENCE: --keeps expanding.

476
00:20:22,660 --> 00:20:24,325
AUDIENCE: So the thing increases but--

477
00:20:24,325 --> 00:20:25,450
DAVID MALAN: Yeah, exactly.

478
00:20:25,450 --> 00:20:28,140
So let's suppose it's
not one per second,

479
00:20:28,140 --> 00:20:33,260
but it takes me for 10,000 codes,
instead of times 1 second for each,

480
00:20:33,260 --> 00:20:35,740
it's actually not even 60 seconds.

481
00:20:35,740 --> 00:20:37,070
It's five minutes.

482
00:20:37,070 --> 00:20:40,950
So now, this is the total number--
this is the total amount of time

483
00:20:40,950 --> 00:20:43,810
I need in order to hack into a phone.

484
00:20:43,810 --> 00:20:46,870
And again, there's 60 seconds in a
minute, and 60 minutes in an hour.

485
00:20:46,870 --> 00:20:50,850
>> So now, we're up to 833 hours.

486
00:20:50,850 --> 00:20:54,787
And if we want to see this precisely,
now we're talking about 34 days.

487
00:20:54,787 --> 00:20:56,870
So it's going to take an
adversary, without sleep,

488
00:20:56,870 --> 00:21:01,250
34 days now to hack into your iPhone,
if there is this five minute delay.

489
00:21:01,250 --> 00:21:03,340
But it's not even just five minutes.

490
00:21:03,340 --> 00:21:05,560
As Kareem said, what
happens after the next--

491
00:21:05,560 --> 00:21:06,210
>> AUDIENCE: After you've tried--

492
00:21:06,210 --> 00:21:06,400
>> DAVID MALAN: --misattempt?

493
00:21:06,400 --> 00:21:09,108
>> AUDIENCE: --five more times, then
it gives you a 10-minute delay.

494
00:21:09,108 --> 00:21:10,450
DAVID MALAN: A 10 minute delay.

495
00:21:10,450 --> 00:21:13,200
And I'm not sure what it is after
that, but maybe it's 20 minutes.

496
00:21:13,200 --> 00:21:14,230
Maybe it's 40 minutes.

497
00:21:14,230 --> 00:21:17,105
And if it is, that's actually an
example of a fairly common technique

498
00:21:17,105 --> 00:21:19,900
in computing known as
exponential backoff, where

499
00:21:19,900 --> 00:21:22,980
this exponentiation usually means
you double something again and again.

500
00:21:22,980 --> 00:21:25,510
>> So that starts out being
not so significant.

501
00:21:25,510 --> 00:21:29,570
But once you start doubling from
2 to 4 to 8 to 16 to 32 to 64,

502
00:21:29,570 --> 00:21:31,650
the gaps really start to widen.

503
00:21:31,650 --> 00:21:35,780
And so it might take a month,
or a year, or a lifetime

504
00:21:35,780 --> 00:21:37,800
to actually get into that device.

505
00:21:37,800 --> 00:21:39,300
Now, there's other mechanisms still.

506
00:21:39,300 --> 00:21:41,350
Time is a good thing
because, in general, this

507
00:21:41,350 --> 00:21:42,940
is a common security technique.

508
00:21:42,940 --> 00:21:46,650
You can't necessarily stop the bad
guys, but you can slow them down.

509
00:21:46,650 --> 00:21:50,480
And because there are finite
resources in life, like living,

510
00:21:50,480 --> 00:21:53,750
you can eventually push
out the threat so far

511
00:21:53,750 --> 00:21:57,510
that even though, sure, the adversary
might get really lucky and try 7 7 7 7

512
00:21:57,510 --> 00:22:01,080
on your phone and get the answer
right, the probability of that

513
00:22:01,080 --> 00:22:02,890
is incredibly low.

514
00:22:02,890 --> 00:22:06,800
And so, generally security is a
function, not of absolute protection,

515
00:22:06,800 --> 00:22:08,390
but of probabilistic protection.

516
00:22:08,390 --> 00:22:13,940
You're just pretty sure that you're
safe from some kind of attack.

517
00:22:13,940 --> 00:22:16,360
>> But that might not be
really good enough.

518
00:22:16,360 --> 00:22:17,780
So what more could you do?

519
00:22:17,780 --> 00:22:21,010
And what more does Apple do,
if people have enabled this,

520
00:22:21,010 --> 00:22:24,700
if an adversary or bad guy tries
to get in more than 10 times,

521
00:22:24,700 --> 00:22:26,070
besides inserting a delay.

522
00:22:26,070 --> 00:22:28,300
What would be a stronger
measure of defense that

523
00:22:28,300 --> 00:22:30,185
might make you sleep better at night?

524
00:22:30,185 --> 00:22:31,430
>> AUDIENCE: Erasing the data.

525
00:22:31,430 --> 00:22:32,250
>> DAVID MALAN: Erase the data.

526
00:22:32,250 --> 00:22:32,750
Yeah.

527
00:22:32,750 --> 00:22:34,750
So in fact, that's a
very common technique

528
00:22:34,750 --> 00:22:36,860
where, much like the
old movies, this message

529
00:22:36,860 --> 00:22:39,360
will self-destruct in 10 seconds.

530
00:22:39,360 --> 00:22:42,050
Very commonly, will
devices, iPhones among them,

531
00:22:42,050 --> 00:22:47,500
just wipe themselves, delete
themselves after 10 incorrect attempts.

532
00:22:47,500 --> 00:22:49,140
So is this a good thing or a bad thing?

533
00:22:49,140 --> 00:22:52,870
Now, let's put on more of
the product manager's hat.

534
00:22:52,870 --> 00:22:55,860
What's good about this?

535
00:22:55,860 --> 00:22:58,540
Why is this a positive feature?

536
00:22:58,540 --> 00:22:59,950
>> [INTERPOSING VOICES]

537
00:22:59,950 --> 00:23:01,350
>> No access to your information.

538
00:23:01,350 --> 00:23:04,710
So now, not only have you
slowed the adversary down,

539
00:23:04,710 --> 00:23:07,320
if you do have those
artificial time delays,

540
00:23:07,320 --> 00:23:10,480
but you've also ensured that if
he or she screws up 10 times,

541
00:23:10,480 --> 00:23:12,730
now the window of
opportunity is just gone.

542
00:23:12,730 --> 00:23:14,090
They've only had 10 attempts.

543
00:23:14,090 --> 00:23:18,050
And the probability of getting the
answer correct out of 10 attempts

544
00:23:18,050 --> 00:23:21,690
when there's 10,000
possibilities is 1 out of 1,000.

545
00:23:21,690 --> 00:23:24,384
So 10 divided by 10,000-- 1 over 1,000.

546
00:23:24,384 --> 00:23:25,800
But even that's not all that good.

547
00:23:25,800 --> 00:23:28,520
So we'll come back to making me
feel better about that probability

548
00:23:28,520 --> 00:23:30,186
because it actually feels somewhat high.

549
00:23:30,186 --> 00:23:31,640
It's 1/10 of a percent.

550
00:23:31,640 --> 00:23:34,734
What's bad about this feature though?

551
00:23:34,734 --> 00:23:36,700
>> AUDIENCE: It's bad because--

552
00:23:36,700 --> 00:23:37,450
DAVID MALAN: Yeah.

553
00:23:37,450 --> 00:23:38,591
What do you mean in my hands?

554
00:23:38,591 --> 00:23:40,591
>> AUDIENCE: If you didn't
lose it, and you're just

555
00:23:40,591 --> 00:23:41,850
trying to get into your phone.

556
00:23:41,850 --> 00:23:42,600
>> DAVID MALAN: Yeah.

557
00:23:42,600 --> 00:23:45,600
So what if there has been no compromise,
your just kind of distracted,

558
00:23:45,600 --> 00:23:47,440
you're an idiot, you
forget your password.

559
00:23:47,440 --> 00:23:50,087
And so it's not that
unreasonable, especially

560
00:23:50,087 --> 00:23:51,920
if you don't log in to
your phone that often

561
00:23:51,920 --> 00:23:55,220
or you're distracted while
doing it, maybe you yourself

562
00:23:55,220 --> 00:23:57,870
mistype your code 11 times.

563
00:23:57,870 --> 00:24:01,110
And now, dammit, you've
just wiped your own device.

564
00:24:01,110 --> 00:24:03,950
So this too is kind of a theme
in computing and computer

565
00:24:03,950 --> 00:24:04,990
science of trade-offs.

566
00:24:04,990 --> 00:24:07,380
There really is rarely a right answer.

567
00:24:07,380 --> 00:24:10,947
There's simply a more preferable
or a less costly answer.

568
00:24:10,947 --> 00:24:12,530
And in this case, there's a trade-off.

569
00:24:12,530 --> 00:24:14,390
One, our data is a
little more secure, if it

570
00:24:14,390 --> 00:24:16,100
gets into the hands of some adversary.

571
00:24:16,100 --> 00:24:18,890
But I can shoot myself in the
foot by wiping, accidentally,

572
00:24:18,890 --> 00:24:22,940
my own data if I don't actually get
that passcode right within the first 10

573
00:24:22,940 --> 00:24:23,699
times.

574
00:24:23,699 --> 00:24:24,490
So what's the push?

575
00:24:24,490 --> 00:24:25,810
How do we fix that?

576
00:24:25,810 --> 00:24:28,500
Do we throw the feature out
altogether, if we're Apple, and say,

577
00:24:28,500 --> 00:24:32,520
this feels bad because we're going to
have-- if we have one irate customer,

578
00:24:32,520 --> 00:24:34,649
this is not a situation
we want to invite.

579
00:24:34,649 --> 00:24:37,190
AUDIENCE: We encrypted and then
we recovered the code somehow

580
00:24:37,190 --> 00:24:41,435
by Apple or whatever [INAUDIBLE].

581
00:24:41,435 --> 00:24:42,726
DAVID MALAN: Can you elaborate?

582
00:24:42,726 --> 00:24:43,601
AUDIENCE: [INAUDIBLE]

583
00:24:43,601 --> 00:24:46,564

584
00:24:46,564 --> 00:24:47,230
DAVID MALAN: OK.

585
00:24:47,230 --> 00:24:52,190
So maybe we don't do this wiping thing,
which feels a little overly dramatic.

586
00:24:52,190 --> 00:24:54,620
Why don't we just keep
the data encrypted?

587
00:24:54,620 --> 00:24:58,120
Well, so in this case, Apple already
does keep the data encrypted.

588
00:24:58,120 --> 00:25:00,920
And what's keeping the adversary
from seeing your encrypted data

589
00:25:00,920 --> 00:25:02,990
is unfortunately that passcode.

590
00:25:02,990 --> 00:25:06,445
>> So the passcode effectively unlocks
the data so that while it's scrambled,

591
00:25:06,445 --> 00:25:09,570
if you're just holding the phone, as
soon as you log in with that passcode,

592
00:25:09,570 --> 00:25:11,650
it's unscrambled and
the user can see it.

593
00:25:11,650 --> 00:25:13,450
So it is already encrypted.

594
00:25:13,450 --> 00:25:17,330
But if we want to avoid wiping
the data, but we somehow

595
00:25:17,330 --> 00:25:20,290
want to have a good answer
on the customer support line

596
00:25:20,290 --> 00:25:25,390
if the absent minded or
forgetful user has accidentally

597
00:25:25,390 --> 00:25:28,390
wiped his or her phone because they
mistyped the password 11 times, what

598
00:25:28,390 --> 00:25:30,050
solution could we offer?

599
00:25:30,050 --> 00:25:33,950
How else could we
solve that problem now?

600
00:25:33,950 --> 00:25:34,450
Yeah.

601
00:25:34,450 --> 00:25:36,074
>> AUDIENCE: Customer service [INAUDIBLE].

602
00:25:36,074 --> 00:25:43,244

603
00:25:43,244 --> 00:25:43,910
DAVID MALAN: OK.

604
00:25:43,910 --> 00:25:45,680
So that's good.

605
00:25:45,680 --> 00:25:49,360
So maybe without using
resorting to wiping,

606
00:25:49,360 --> 00:25:52,850
we could have some out-of-band
mechanism for solving this problem.

607
00:25:52,850 --> 00:25:55,600
And by out-of-band, I mean you
don't interact just with the phone,

608
00:25:55,600 --> 00:25:59,960
maybe you grab someone else's phone or
email and you talk to customer service.

609
00:25:59,960 --> 00:26:02,410
And maybe they ask you the
usual questions of, well,

610
00:26:02,410 --> 00:26:04,290
what's your name, what's
your birthdate, what

611
00:26:04,290 --> 00:26:09,870
are the last four digits of your
social security number or country ID.

612
00:26:09,870 --> 00:26:11,320
>> And what's good about that?

613
00:26:11,320 --> 00:26:13,430
Well, of course, with
high probability, it

614
00:26:13,430 --> 00:26:15,920
lets you and only you into
your phone because maybe they

615
00:26:15,920 --> 00:26:17,099
send a temporary passcode.

616
00:26:17,099 --> 00:26:19,390
And this does not exist in
Apple's case, but maybe they

617
00:26:19,390 --> 00:26:20,820
do send you a temporary passcode.

618
00:26:20,820 --> 00:26:23,040
You get in, and you're back on your way.

619
00:26:23,040 --> 00:26:25,245
But what's the downside
of this solution?

620
00:26:25,245 --> 00:26:27,065
>> AUDIENCE: If someone
steals your identity,

621
00:26:27,065 --> 00:26:29,050
they might have access
to all this information.

622
00:26:29,050 --> 00:26:29,800
DAVID MALAN: Yeah.

623
00:26:29,800 --> 00:26:32,110
If someone steals your
identity-- and frankly,

624
00:26:32,110 --> 00:26:34,550
it's not all that hard,
especially when so many companies

625
00:26:34,550 --> 00:26:35,859
ask the same questions.

626
00:26:35,859 --> 00:26:37,650
What's your name, what's
your address, what

627
00:26:37,650 --> 00:26:40,191
are the last four digits of your
social security number, what

628
00:26:40,191 --> 00:26:43,680
was your favorite pet, what
was your favorite parent

629
00:26:43,680 --> 00:26:45,281
or whatever the questions might be.

630
00:26:45,281 --> 00:26:48,280
And in fact, as an aside, I've noticed,
having just the other day filled

631
00:26:48,280 --> 00:26:52,170
out questions like these, the
questions in a reasonable effort

632
00:26:52,170 --> 00:26:57,204
to become a little less well-known
are getting increasingly personal.

633
00:26:57,204 --> 00:27:00,370
And as soon as you start giving this
tidbit of information that might indeed

634
00:27:00,370 --> 00:27:03,392
be a secret to this company, and to
this company, and to this company,

635
00:27:03,392 --> 00:27:06,600
and to this company, it's not going to
be long before some company aggregates

636
00:27:06,600 --> 00:27:07,900
this kind of information.

637
00:27:07,900 --> 00:27:09,860
And so you've told
little simple secrets,

638
00:27:09,860 --> 00:27:13,330
like your best friend growing up, to
all of these individual companies.

639
00:27:13,330 --> 00:27:16,900
And soon enough, you have an attack
known as social engineering, whereby

640
00:27:16,900 --> 00:27:20,237
someone just masquerades as you on
the phone or spoofs your email address

641
00:27:20,237 --> 00:27:21,570
and somehow gets into the phone.

642
00:27:21,570 --> 00:27:22,790
>> So I'm not liking that.

643
00:27:22,790 --> 00:27:26,240
It's a possible solution, but
let's suppose I'm not liking that.

644
00:27:26,240 --> 00:27:29,730
Let's go back to the issue at
hand where phone is encrypted

645
00:27:29,730 --> 00:27:34,440
and we've not enabled some kind
of self-destruct mechanism.

646
00:27:34,440 --> 00:27:39,900
But I do-- rather, I have enabled
some self-destruct mechanism,

647
00:27:39,900 --> 00:27:42,510
but I nonetheless want
to appease a customer who

648
00:27:42,510 --> 00:27:44,970
accidentally wipes his or her phone.

649
00:27:44,970 --> 00:27:46,920
How else could we solve that problem?

650
00:27:46,920 --> 00:27:48,160
>> AUDIENCE: Make a backup.

651
00:27:48,160 --> 00:27:48,920
>> DAVID MALAN: Make a backup.

652
00:27:48,920 --> 00:27:50,990
And indeed, this is how
Apple happens to do this.

653
00:27:50,990 --> 00:27:53,964
One of the motivations of iCloud is
exactly this-- not only convenience

654
00:27:53,964 --> 00:27:56,380
and resting assured that all
of your photos and everything

655
00:27:56,380 --> 00:27:59,710
are backed up, but in this case--
because if your individual device,

656
00:27:59,710 --> 00:28:04,610
whether it's an iPod, or iPhone, or
iPad is lost, or stolen, or accidentally

657
00:28:04,610 --> 00:28:07,820
or deliberately wiped, at least
all of your data is somewhere else.

658
00:28:07,820 --> 00:28:10,130
And you can just go buy
or borrow another iPhone.

659
00:28:10,130 --> 00:28:13,180
You can restore from backup,
so to speak, from iCloud,

660
00:28:13,180 --> 00:28:14,640
and you're back up and running.

661
00:28:14,640 --> 00:28:16,260
>> Now, there's a trade-off there.

662
00:28:16,260 --> 00:28:19,660
Potentially, Apple now has
access to all of that same data.

663
00:28:19,660 --> 00:28:21,320
And we can come back to that some time.

664
00:28:21,320 --> 00:28:24,780
But at least now, we've solved
the problem in a different way.

665
00:28:24,780 --> 00:28:27,880
>> And if you visualize this
story line in your mind's eye,

666
00:28:27,880 --> 00:28:30,900
you can perhaps see that every
time we solve a problem--

667
00:28:30,900 --> 00:28:36,130
kind of covering up a leak in hose,
some other problem springs up elsewhere.

668
00:28:36,130 --> 00:28:38,530
We're really just pushing
the problem somewhere else.

669
00:28:38,530 --> 00:28:41,150
And in the case of the
adversary with the time delays,

670
00:28:41,150 --> 00:28:43,880
really what we're doing is
we're not keeping adversary out,

671
00:28:43,880 --> 00:28:47,190
we're just raising the bar over
which he or she has to jump in order

672
00:28:47,190 --> 00:28:49,310
to actually get access to our data.

673
00:28:49,310 --> 00:28:53,160
>> So any time, henceforth, you go to a
website, or you read some white paper,

674
00:28:53,160 --> 00:28:58,780
or some CTO or CSO tells you, oh, our
systems are secure-- it's baloney.

675
00:28:58,780 --> 00:29:02,480
There's nothing to be meant
by "our systems are secure"

676
00:29:02,480 --> 00:29:06,150
other than we take industry
standard probabilistic measures

677
00:29:06,150 --> 00:29:10,810
to keep people away from your
servers or away from your data.

678
00:29:10,810 --> 00:29:14,480
>> Now, the Apple situation has
gotten kind of interesting

679
00:29:14,480 --> 00:29:16,650
because they've been
asked to do something

680
00:29:16,650 --> 00:29:20,730
that's not quite as simple as
turn over the adversary's data.

681
00:29:20,730 --> 00:29:22,450
They've already done that from iCloud.

682
00:29:22,450 --> 00:29:25,320
But now, the FBI wants
to get into this phone.

683
00:29:25,320 --> 00:29:29,650
And the belief that it does in fact
have this self-destruct mechanism built

684
00:29:29,650 --> 00:29:33,030
in after 10 attempts-- and I
believe that's because they

685
00:29:33,030 --> 00:29:36,680
looked at the backups and realized
this feature seems to be enabled,

686
00:29:36,680 --> 00:29:38,700
and I assume they don't
want to necessarily

687
00:29:38,700 --> 00:29:41,720
try and waste one out
of their 10 attempts

688
00:29:41,720 --> 00:29:44,170
to confirm or deny this feature.

689
00:29:44,170 --> 00:29:49,510
>> And they also, unfortunately-- and
this is sort of the irony of it

690
00:29:49,510 --> 00:29:54,380
all, the county where this
fellow worked actually

691
00:29:54,380 --> 00:29:58,430
owned and was paying for special
software-- device management software--

692
00:29:58,430 --> 00:30:01,550
that had it been installed
on their employees' phones--

693
00:30:01,550 --> 00:30:04,820
so the phone in question is actually
state property or county property that

694
00:30:04,820 --> 00:30:06,240
was being used by an employee.

695
00:30:06,240 --> 00:30:10,580
Had they installed in advance
this device management software,

696
00:30:10,580 --> 00:30:13,660
they could have with a
simple click on a PC or Mac

697
00:30:13,660 --> 00:30:15,865
unlocked this phone trivially.

698
00:30:15,865 --> 00:30:18,740
But unfortunately, they didn't have
that software actually installed.

699
00:30:18,740 --> 00:30:21,510
>> So there are yet other ways
to address this kind of issue.

700
00:30:21,510 --> 00:30:24,590
It doesn't have to be a black
box in your employee's pocket.

701
00:30:24,590 --> 00:30:25,460
But they didn't.

702
00:30:25,460 --> 00:30:27,210
And so now we're stuck
with the situation

703
00:30:27,210 --> 00:30:31,290
with an encrypted iPhone that will
literally self-- will figuratively

704
00:30:31,290 --> 00:30:34,340
self-destruct after
10 incorrect attempts.

705
00:30:34,340 --> 00:30:37,240
And the FBI wants to get
data off of that phone.

706
00:30:37,240 --> 00:30:42,627
>> So let's take a look at what Tim
Cook has announced to the world

707
00:30:42,627 --> 00:30:43,710
and taken this bold stand.

708
00:30:43,710 --> 00:30:46,080
If you've not read it, let
me go ahead and do this.

709
00:30:46,080 --> 00:30:51,690
If you'd like either on your
computer to go to this you URL here,

710
00:30:51,690 --> 00:30:55,750
or I can grab for you some paper copies.

711
00:30:55,750 --> 00:30:59,090
Why don't we just take
two minutes, if you would,

712
00:30:59,090 --> 00:31:04,800
and read the actual letter that Tim
Cook wrote to Apple's customers.

713
00:31:04,800 --> 00:31:08,062
And we'll see if we can't then
tease apart what it actually means.

714
00:31:08,062 --> 00:31:09,770
And so I've circled
a few things in this.

715
00:31:09,770 --> 00:31:12,750
But let's see if we can't distill
what's actually being said here

716
00:31:12,750 --> 00:31:15,660
and where the real
interesting stuff is hidden.

717
00:31:15,660 --> 00:31:22,120
So I for instance, on the paragraph
starting-- under the San Bernardino

718
00:31:22,120 --> 00:31:26,747
case, the paragraph starting
"we have great respect for,"

719
00:31:26,747 --> 00:31:28,830
Tim Cook's last sentence
is this. "They have asked

720
00:31:28,830 --> 00:31:31,110
us to build a backdoor to the iPhone."

721
00:31:31,110 --> 00:31:33,510
This is a commonly used phrase,
"backdoor" to something.

722
00:31:33,510 --> 00:31:36,010
What does this actually
mean, as best you can tell,

723
00:31:36,010 --> 00:31:37,691
from what you've read here or elsewhere?

724
00:31:37,691 --> 00:31:38,440
AUDIENCE: Hack it.

725
00:31:38,440 --> 00:31:40,820
DAVID MALAN: They want to be able
to hack it, and what does that mean?

726
00:31:40,820 --> 00:31:42,110
What is a backdoor?

727
00:31:42,110 --> 00:31:43,710
>> AUDIENCE: An alternate entry point?

728
00:31:43,710 --> 00:31:44,460
DAVID MALAN: Yeah.

729
00:31:44,460 --> 00:31:45,800
So it's an alternate entry point.

730
00:31:45,800 --> 00:31:47,680
Much like an actual house
where you have a front door,

731
00:31:47,680 --> 00:31:50,721
and sometimes a back door where you're
supposed to come in the front door

732
00:31:50,721 --> 00:31:53,990
and maybe not so much the back
door, unless you belong there,

733
00:31:53,990 --> 00:31:57,510
the FBI is asking for a
figurative back door-- another way

734
00:31:57,510 --> 00:32:01,880
of getting into the phone that isn't
simply a human finger touching the code

735
00:32:01,880 --> 00:32:03,370
and getting in in the usual way.

736
00:32:03,370 --> 00:32:07,070
They want to somehow slurp the data off,
maybe with a cable, maybe wirelessly,

737
00:32:07,070 --> 00:32:10,830
or they want to somehow be able to
input the code, perhaps, to the phone

738
00:32:10,830 --> 00:32:13,330
without just using a raw human finger.

739
00:32:13,330 --> 00:32:16,285
>> So they allude to, in
the next paragraph,

740
00:32:16,285 --> 00:32:18,910
"the FBI wants us to make a new
version of the iPhone operating

741
00:32:18,910 --> 00:32:23,250
system, circumventing several
important security features."

742
00:32:23,250 --> 00:32:27,870
So why is the FBI asking Apple
to make a new operating system?

743
00:32:27,870 --> 00:32:30,410
That seems to be kind
of besides the point?

744
00:32:30,410 --> 00:32:31,130
No?

745
00:32:31,130 --> 00:32:33,476
Why do you think they
might be saying that?

746
00:32:33,476 --> 00:32:36,300

747
00:32:36,300 --> 00:32:37,960
How is the solution to the problem?

748
00:32:37,960 --> 00:32:40,430
>> AUDIENCE: [INAUDIBLE]

749
00:32:40,430 --> 00:32:41,180
DAVID MALAN: Yeah.

750
00:32:41,180 --> 00:32:42,715
AUDIENCE: [INAUDIBLE]

751
00:32:42,715 --> 00:32:43,590
DAVID MALAN: Exactly.

752
00:32:43,590 --> 00:32:46,881
The version of iOS, the operating system
that's running on the phone currently,

753
00:32:46,881 --> 00:32:50,020
has all of the security measures
that we were discussing earlier,

754
00:32:50,020 --> 00:32:52,610
for instance, the time
delay, potentially

755
00:32:52,610 --> 00:32:55,940
the self-destruct mechanism, all
of which are particularly bad.

756
00:32:55,940 --> 00:33:00,900
And so the data is on there
encrypted, and as best we can tell,

757
00:33:00,900 --> 00:33:05,610
Apple somehow stores data separate from
the underlying operation of the phone--

758
00:33:05,610 --> 00:33:07,030
the operating system.

759
00:33:07,030 --> 00:33:11,020
And so it would seem to be
possible to install a new operating

760
00:33:11,020 --> 00:33:14,660
system on the phone without
touching the user's data.

761
00:33:14,660 --> 00:33:18,090
>> In fact, if any of you have
ever updated Mac OS or Windows,

762
00:33:18,090 --> 00:33:20,340
it would be-- hopefully,
it's supposed to go this way.

763
00:33:20,340 --> 00:33:23,020
Hopefully, you've been able to update
your operating system from an older

764
00:33:23,020 --> 00:33:25,180
version to a newer version
without starting over,

765
00:33:25,180 --> 00:33:26,870
without deleting all of your files.

766
00:33:26,870 --> 00:33:28,410
Now, some of you have probably
had the experience where

767
00:33:28,410 --> 00:33:29,868
that does not go according to plan.

768
00:33:29,868 --> 00:33:32,920
But in theory, this should be
possible-- update the operating system,

769
00:33:32,920 --> 00:33:35,440
but do not delete or
touch the actual data.

770
00:33:35,440 --> 00:33:38,970
>> So the FBI is proposing that Apple
create an operating system that

771
00:33:38,970 --> 00:33:42,770
doesn't have these kinds of defense
mechanisms, installed onto the phone

772
00:33:42,770 --> 00:33:45,330
so that they can get access to the data.

773
00:33:45,330 --> 00:33:46,740
Now, how is that possible?

774
00:33:46,740 --> 00:33:49,510
Wouldn't the suspect
himself, who is now deceased,

775
00:33:49,510 --> 00:33:52,540
have to install that
software for the FBI?

776
00:33:52,540 --> 00:33:55,932
What is the FBI counting on here?

777
00:33:55,932 --> 00:33:57,828
>> AUDIENCE: Push it down somehow?

778
00:33:57,828 --> 00:33:59,250
That was my question.

779
00:33:59,250 --> 00:34:01,720
How do you install it if
you can't log in to it?

780
00:34:01,720 --> 00:34:02,470
DAVID MALAN: Yeah.

781
00:34:02,470 --> 00:34:02,970
Exactly.

782
00:34:02,970 --> 00:34:05,540
So you would seem to have a
chicken and the egg problem here,

783
00:34:05,540 --> 00:34:08,989
whereby, you would think,
to update the software,

784
00:34:08,989 --> 00:34:10,400
you need to log into the phone.

785
00:34:10,400 --> 00:34:13,380
But to log into the phone, you
need to update the software so as

786
00:34:13,380 --> 00:34:15,909
to avoid these defense mechanisms.

787
00:34:15,909 --> 00:34:17,330
>> So lets just reason backwards.

788
00:34:17,330 --> 00:34:22,010
So not necessarily being the programmer
at Apple, what must be the case?

789
00:34:22,010 --> 00:34:26,848
If the FBI thinks it can do this,
what must be the case logically?

790
00:34:26,848 --> 00:34:28,347
What is the answer to that question?

791
00:34:28,347 --> 00:34:31,420

792
00:34:31,420 --> 00:34:35,159
>> It must be possible to
do, presumably, somehow.

793
00:34:35,159 --> 00:34:37,600
So how might you do it?

794
00:34:37,600 --> 00:34:42,620
All you are is a user with an
iPhone, maybe a Mac or a PC,

795
00:34:42,620 --> 00:34:45,860
maybe a cable, maybe
a network connection.

796
00:34:45,860 --> 00:34:49,679
How might the FBI be
assuming Apple can do this?

797
00:34:49,679 --> 00:34:51,710
>> AUDIENCE: Maybe through
automatic updates?

798
00:34:51,710 --> 00:34:52,460
>> DAVID MALAN: Yeah.

799
00:34:52,460 --> 00:34:54,679
So very much in vogue these
days is automatic updates

800
00:34:54,679 --> 00:34:56,770
where an Android phone,
and iPhone, Windows Phone,

801
00:34:56,770 --> 00:34:59,400
whatnot will just
automatically download updates.

802
00:34:59,400 --> 00:35:02,750
So maybe Apple could just
update the operating system,

803
00:35:02,750 --> 00:35:05,280
as the FBI has requested, put
a copy of the new operating

804
00:35:05,280 --> 00:35:09,100
system in the cloud on
their servers, and just wait

805
00:35:09,100 --> 00:35:12,940
for the suspect's phone to connect
automatically, as it probably

806
00:35:12,940 --> 00:35:15,420
does nightly or every
five minutes or something,

807
00:35:15,420 --> 00:35:17,500
to pull down the new operating system.

808
00:35:17,500 --> 00:35:19,420
>> Now, let's pause for just a moment.

809
00:35:19,420 --> 00:35:22,262
You probably don't want to do
that for everyone in the world,

810
00:35:22,262 --> 00:35:23,970
otherwise we have an
even bigger problem.

811
00:35:23,970 --> 00:35:26,270
Well, maybe the FBI might like to
do that to everyone in the world,

812
00:35:26,270 --> 00:35:28,000
but probably won't go over so well.

813
00:35:28,000 --> 00:35:31,620
So just thinking logically
here, is that possible?

814
00:35:31,620 --> 00:35:32,660
Is that a deal breaker?

815
00:35:32,660 --> 00:35:37,370
Can you roll out software to
just one user in that scenario?

816
00:35:37,370 --> 00:35:38,896
How, would you think?

817
00:35:38,896 --> 00:35:41,830
>> AUDIENCE: You make it available
only for that device's address.

818
00:35:41,830 --> 00:35:42,580
DAVID MALAN: Yeah.

819
00:35:42,580 --> 00:35:44,110
Just for that device's address.

820
00:35:44,110 --> 00:35:46,180
And maybe that addresses
is some numeric address.

821
00:35:46,180 --> 00:35:47,730
Maybe it's the device's phone number.

822
00:35:47,730 --> 00:35:50,730
Maybe it's the device's
Apple ID, if you're familiar,

823
00:35:50,730 --> 00:35:52,560
like the email address
that the human uses

824
00:35:52,560 --> 00:35:56,210
to log in to that-- for automatic
updates to the App Store.

825
00:35:56,210 --> 00:35:58,210
So there's probably a way to do that.

826
00:35:58,210 --> 00:36:01,370
>> So you have the operating system
for everyone in the world,

827
00:36:01,370 --> 00:36:04,330
except for this one person who has
his own version of the operating

828
00:36:04,330 --> 00:36:05,520
system getting pulled down.

829
00:36:05,520 --> 00:36:07,250
Now, maybe it's not on the network.

830
00:36:07,250 --> 00:36:09,844
Maybe that's a little
easier said than done.

831
00:36:09,844 --> 00:36:11,010
So what's another mechanism?

832
00:36:11,010 --> 00:36:14,093
Well, it wasn't all that long ago that
most of us here, Android or iPhone,

833
00:36:14,093 --> 00:36:17,600
were updating our phones via cable--
some kind of USB cable connected

834
00:36:17,600 --> 00:36:18,479
to your Mac or PC.

835
00:36:18,479 --> 00:36:20,020
And that might very well be possible.

836
00:36:20,020 --> 00:36:25,070
>> And in fact, this is arguably a security
flaw in the current version of iOS,

837
00:36:25,070 --> 00:36:28,080
and iPhones more generally,
that that is in fact possible.

838
00:36:28,080 --> 00:36:32,480
You can update the software on the
phone without unlocking the phone,

839
00:36:32,480 --> 00:36:33,460
it would seem.

840
00:36:33,460 --> 00:36:35,310
Now, why is that a security flaw?

841
00:36:35,310 --> 00:36:39,660
Because they have opened themselves
to exactly this kind of request.

842
00:36:39,660 --> 00:36:44,786
>> So as an aside, the outcome that seems
inevitable from this whole process

843
00:36:44,786 --> 00:36:47,660
is there is no way that's going to
be possible with the next version,

844
00:36:47,660 --> 00:36:48,710
you would think, of iOS.

845
00:36:48,710 --> 00:36:49,210
Right?

846
00:36:49,210 --> 00:36:51,980
They could have deliberately
tied their hands-- Apple-- so

847
00:36:51,980 --> 00:36:53,470
that this isn't even possible.

848
00:36:53,470 --> 00:36:57,050
>> Now, they've probably been assuming
that because only they own the source

849
00:36:57,050 --> 00:37:00,420
code to iOS that this
isn't really a threat

850
00:37:00,420 --> 00:37:03,990
because no one's going to sit down
and build a whole operating system

851
00:37:03,990 --> 00:37:06,040
and figure out how to
install it on an iPhone.

852
00:37:06,040 --> 00:37:09,666
But it's certainly possible now to
just require a passcode moving forward

853
00:37:09,666 --> 00:37:11,040
to install this operating system.

854
00:37:11,040 --> 00:37:12,789
>> So that's the gist of
what they're asking.

855
00:37:12,789 --> 00:37:17,860
And the bigger picture that we can
defer to perhaps a lunchtime style

856
00:37:17,860 --> 00:37:20,200
chat or dinner table style
chat-- the government

857
00:37:20,200 --> 00:37:23,800
suggests that this tool could
be used only once on one phone.

858
00:37:23,800 --> 00:37:27,590
And that's where privacy
defendants really

859
00:37:27,590 --> 00:37:31,860
bring some strength to bear that
just seems very unreasonable.

860
00:37:31,860 --> 00:37:36,030
As soon as the software actually exists,
surely additional legal requests will

861
00:37:36,030 --> 00:37:39,870
come in, surely there's a risk of some
bad guy getting access to that kind

862
00:37:39,870 --> 00:37:42,560
of software, installing it
him or herself on phones,

863
00:37:42,560 --> 00:37:46,000
and so you're just opening, it
would seem , a can of worms.

864
00:37:46,000 --> 00:37:48,690
>> Now, even Obama recently,
if you've read or listened

865
00:37:48,690 --> 00:37:50,640
to one of his recent
speeches, commented,

866
00:37:50,640 --> 00:37:53,830
I think, that folks seemed to be
fetishizing their phones, so to speak,

867
00:37:53,830 --> 00:37:58,222
whereby we have accepted over 300 years
the fact that the police with a warrant

868
00:37:58,222 --> 00:38:00,180
hopefully can come into
your home or can search

869
00:38:00,180 --> 00:38:02,720
through the contents of
your drawers or whatnot,

870
00:38:02,720 --> 00:38:06,590
and yet, we seem to be putting
a phone on this pedestal whereby

871
00:38:06,590 --> 00:38:08,829
it should be immune to all prying eyes.

872
00:38:08,829 --> 00:38:11,870
But I would argue, frankly, from a
computer scientist's perspective, that

873
00:38:11,870 --> 00:38:15,370
is actually progress-- the fact that
we now have the mathematical means

874
00:38:15,370 --> 00:38:19,770
to actually keep data truly secure by
way of this thing called encryption,

875
00:38:19,770 --> 00:38:22,490
and we'll come back to
in just a little bit.

876
00:38:22,490 --> 00:38:27,510
>> So any questions about
any of that just yet?

877
00:38:27,510 --> 00:38:31,870
Well, let me show you just
how there is, in fact, one way

878
00:38:31,870 --> 00:38:34,200
to brute force your way into a phone.

879
00:38:34,200 --> 00:38:36,240
And in fact, this is
not out of the question.

880
00:38:36,240 --> 00:38:40,000
This is just a short YouTube video
of essentially a little robot someone

881
00:38:40,000 --> 00:38:45,570
built that does this with a little pad.

882
00:38:45,570 --> 00:38:46,830
>> And I forget what it is.

883
00:38:46,830 --> 00:38:49,830
This is using an Android phone because
an Android phone, in this case,

884
00:38:49,830 --> 00:38:51,210
is vulnerable to this attack.

885
00:38:51,210 --> 00:38:52,440
It will not timeout.

886
00:38:52,440 --> 00:38:55,530
It does not increase the
delay between attempts.

887
00:38:55,530 --> 00:38:59,490
And so you can just do this-- I
think for like three days, I think,

888
00:38:59,490 --> 00:39:01,390
was the caption in this video.

889
00:39:01,390 --> 00:39:03,730
After three days, this
funny looking device

890
00:39:03,730 --> 00:39:06,390
will hack into an Android
phone that has a four-- maybe

891
00:39:06,390 --> 00:39:07,900
it was a six digit passcode.

892
00:39:07,900 --> 00:39:12,530
So beware something like this--
you see this on the table near you.

893
00:39:12,530 --> 00:39:14,059
>> This though is one mechanism.

894
00:39:14,059 --> 00:39:15,600
So what is Apple actually asking for?

895
00:39:15,600 --> 00:39:16,710
This article's a little longer.

896
00:39:16,710 --> 00:39:19,550
And it's the only other article
we'll read today on paper or online.

897
00:39:19,550 --> 00:39:22,860
But let me invite you to take
probably four or so minutes

898
00:39:22,860 --> 00:39:25,770
to take a look at the following.

899
00:39:25,770 --> 00:39:28,160
This is a longer URL here.

900
00:39:28,160 --> 00:39:29,950
But if you have the
slides open in a tab,

901
00:39:29,950 --> 00:39:33,140
you can probably just copy and paste
this from the slides themselves.

902
00:39:33,140 --> 00:39:37,130
And I have a printout here, if you
would prefer actually looking on paper.

903
00:39:37,130 --> 00:39:40,470
>> This is a more technical article
that'll offer us an opportunity

904
00:39:40,470 --> 00:39:43,580
to actually tease apart
more technical jargon,

905
00:39:43,580 --> 00:39:47,575
and see what the authors actually mean.

906
00:39:47,575 --> 00:39:51,090
So if you need to keep finishing up--
but let me toss the question out there,

907
00:39:51,090 --> 00:39:54,680
based on what you've read, are there
any buzzwords, sentences, claims,

908
00:39:54,680 --> 00:39:57,350
that we should first
translate or distill

909
00:39:57,350 --> 00:40:00,252
that would make everything
more straightforward?

910
00:40:00,252 --> 00:40:03,090

911
00:40:03,090 --> 00:40:05,070
Anything at all?

912
00:40:05,070 --> 00:40:08,400
So if I started to pop quiz us
on what certain sentences mean,

913
00:40:08,400 --> 00:40:09,610
we should be OK?

914
00:40:09,610 --> 00:40:10,310
Oh, there we go.

915
00:40:10,310 --> 00:40:10,810
OK.

916
00:40:10,810 --> 00:40:13,165
AUDIENCE: [INAUDIBLE]
building some code into RAM.

917
00:40:13,165 --> 00:40:14,040
DAVID MALAN: Oh, RAM.

918
00:40:14,040 --> 00:40:14,540
OK.

919
00:40:14,540 --> 00:40:16,060
Yeah.

920
00:40:16,060 --> 00:40:19,124
RAM-- let me define it first and
we'll come back to that point.

921
00:40:19,124 --> 00:40:20,399
>> AUDIENCE: [INAUDIBLE]

922
00:40:20,399 --> 00:40:22,190
DAVID MALAN: What
they're asking for there.

923
00:40:22,190 --> 00:40:22,690
OK.

924
00:40:22,690 --> 00:40:25,164
So as a definition, RAM
is Random Access Memory.

925
00:40:25,164 --> 00:40:27,580
This is the type of memory
that all of our computers have.

926
00:40:27,580 --> 00:40:31,060
It is distinct from a hard
disk or a solid state disk.

927
00:40:31,060 --> 00:40:34,480
And a solid state disk or hard disk is
where your data is stored long term.

928
00:40:34,480 --> 00:40:37,490
So when you unplug the cord,
even when your battery dies,

929
00:40:37,490 --> 00:40:41,240
any data or programs that you have on
your hard drive or solid state drive

930
00:40:41,240 --> 00:40:42,240
remain there.

931
00:40:42,240 --> 00:40:45,380
>> RAM, meanwhile is the type of memory
that, when you double click an icon,

932
00:40:45,380 --> 00:40:47,530
or open some file, or
run some program, it's

933
00:40:47,530 --> 00:40:51,390
copied from the hard drive or
the solid state drive into RAM.

934
00:40:51,390 --> 00:40:54,440
RAM tends to be faster,
albeit more expensive.

935
00:40:54,440 --> 00:40:57,210
And that's where files and programs
live while they're being used.

936
00:40:57,210 --> 00:40:59,510
>> So we'll come back to the
implications of that in just a moment.

937
00:40:59,510 --> 00:41:01,840
But for those unfamiliar,
that's what that's all about.

938
00:41:01,840 --> 00:41:04,420
And phones have it as well.

939
00:41:04,420 --> 00:41:09,250
Any other definitions or
clarifications we can make?

940
00:41:09,250 --> 00:41:09,750
All right.

941
00:41:09,750 --> 00:41:13,710
So the pop quiz is what
are the three, at least,

942
00:41:13,710 --> 00:41:17,890
things that the FBI is specifically
asking Apple for technically?

943
00:41:17,890 --> 00:41:24,360

944
00:41:24,360 --> 00:41:26,064
One of them does indeed relate to RAM.

945
00:41:26,064 --> 00:41:27,230
So that's the spoiler there.

946
00:41:27,230 --> 00:41:28,925
And we'll come back to what that means.

947
00:41:28,925 --> 00:41:32,580

948
00:41:32,580 --> 00:41:35,679
But what does the government want?

949
00:41:35,679 --> 00:41:37,470
Yeah, Chris, you want
to give us one other?

950
00:41:37,470 --> 00:41:42,140
>> AUDIENCE: I think the ability to
electronically brute force a password,

951
00:41:42,140 --> 00:41:44,640
DAVID MALAN: Yeah, electronically
brute force the passwords.

952
00:41:44,640 --> 00:41:47,866
Again, brute force-- quick recap,
what does brute forcing mean?

953
00:41:47,866 --> 00:41:49,575
>> AUDIENCE: Try the
number of combinations.

954
00:41:49,575 --> 00:41:50,365
DAVID MALAN: Again.

955
00:41:50,365 --> 00:41:51,000
Exactly.

956
00:41:51,000 --> 00:41:53,730
Just try it again, and again,
and again, via brute force,

957
00:41:53,730 --> 00:41:55,550
not via intellect, not via cleverness.

958
00:41:55,550 --> 00:41:57,500
Just try every darn possibility.

959
00:41:57,500 --> 00:42:00,470
So the government wants a way
to avoid brute force-- they

960
00:42:00,470 --> 00:42:03,500
want a way to be able to
brute force it electronically,

961
00:42:03,500 --> 00:42:06,134
and electronically as opposed to what?

962
00:42:06,134 --> 00:42:07,030
>> AUDIENCE: Manually.

963
00:42:07,030 --> 00:42:07,946
>> DAVID MALAN: Manually.

964
00:42:07,946 --> 00:42:10,370
So as opposed to an FBI agent
physically typing things in,

965
00:42:10,370 --> 00:42:12,870
and as opposed to silly looking
devices like the one we just

966
00:42:12,870 --> 00:42:16,759
saw, automatically punching them, they
presumably want to do this wirelessly.

967
00:42:16,759 --> 00:42:19,300
And in fact, if you read the
government's request-- the court

968
00:42:19,300 --> 00:42:21,680
document-- via Bluetooth,
Wi-Fi, whatever

969
00:42:21,680 --> 00:42:23,910
is possible-- or maybe
via lightning cable

970
00:42:23,910 --> 00:42:25,820
that plugs into the
phone itself that would

971
00:42:25,820 --> 00:42:28,760
be connected via USB to some
hacking device that they have.

972
00:42:28,760 --> 00:42:32,510
>> So they want the ability to brute
force the phone electronically

973
00:42:32,510 --> 00:42:36,960
so that they can just do it faster
than a human or a robot could do it.

974
00:42:36,960 --> 00:42:40,100
They want somehow RAM--
let me read that sentence.

975
00:42:40,100 --> 00:42:43,360
"It wants Apple to design this
crippled software, the new operating

976
00:42:43,360 --> 00:42:48,280
system, to be loaded into memory,
AKA RAM, instead of on disk

977
00:42:48,280 --> 00:42:51,140
so that the data on the phone
remains forensically sound

978
00:42:51,140 --> 00:42:52,380
and won't be altered."

979
00:42:52,380 --> 00:42:57,920
>> So it's not clear to us, the readers,
exactly where the data is stored

980
00:42:57,920 --> 00:43:00,200
and where the operating
system is stored.

981
00:43:00,200 --> 00:43:02,800
But presumably, as a
matter of principle in law,

982
00:43:02,800 --> 00:43:06,020
the government doesn't
want to risk mutating

983
00:43:06,020 --> 00:43:10,180
any of the bits-- any of the 0s
and 1s, or the data on the drive--

984
00:43:10,180 --> 00:43:13,380
by putting a new operating
system onto the hard disk itself,

985
00:43:13,380 --> 00:43:16,024
lest that open them up
to a claim that wait

986
00:43:16,024 --> 00:43:19,190
a minute that file wasn't previously
there when the suspect owned the phone.

987
00:43:19,190 --> 00:43:21,600
>> Rather they want to put the
operating system in RAM,

988
00:43:21,600 --> 00:43:26,270
Random Access Memory, which is this
faster speed place that is distinct,

989
00:43:26,270 --> 00:43:28,720
physically, from the actual hard disk.

990
00:43:28,720 --> 00:43:32,320
Of course, the operating system doesn't
typically go there in its entirety,

991
00:43:32,320 --> 00:43:35,250
so that's a non-trivial request.

992
00:43:35,250 --> 00:43:39,280
So we've got this RAM request,
we've got this brute force request,

993
00:43:39,280 --> 00:43:42,425
and one other at least.

994
00:43:42,425 --> 00:43:44,050
What else is the government asking for?

995
00:43:44,050 --> 00:43:44,549
Ben?

996
00:43:44,549 --> 00:43:46,410
AUDIENCE: Remove the timing delay.

997
00:43:46,410 --> 00:43:47,160
DAVID MALAN: Yeah.

998
00:43:47,160 --> 00:43:49,040
Remove that timing
delay, which in this case

999
00:43:49,040 --> 00:43:55,470
is how many seconds, or
milliseconds, or-- 80 milliseconds?

1000
00:43:55,470 --> 00:43:57,310
Which sounds pretty fast.

1001
00:43:57,310 --> 00:44:02,870
I mean most humans can only notice
delays of 100 200 milliseconds

1002
00:44:02,870 --> 00:44:04,770
before something actually feels slow.

1003
00:44:04,770 --> 00:44:07,330
But 80 milliseconds is
roughly 100 milliseconds.

1004
00:44:07,330 --> 00:44:11,900
And 1,000 milliseconds is a second.

1005
00:44:11,900 --> 00:44:16,890
So that's like-- you can do 10
attempts per second, give or take.

1006
00:44:16,890 --> 00:44:20,230
>> So that feels pretty fast,
but not nearly fast enough

1007
00:44:20,230 --> 00:44:21,930
if you've got a six digit code.

1008
00:44:21,930 --> 00:44:24,330
And in fact, the article
makes mention of that too.

1009
00:44:24,330 --> 00:44:27,550
So if you've got a four digit
code, as we discussed before,

1010
00:44:27,550 --> 00:44:30,200
you might have one, two, three, four.

1011
00:44:30,200 --> 00:44:34,880
>> And each of these numbers can
be the number 0 through 9.

1012
00:44:34,880 --> 00:44:38,710
So that's 10 possibilities times 10
possibilities times 10 possibilities

1013
00:44:38,710 --> 00:44:39,630
times 10.

1014
00:44:39,630 --> 00:44:43,170
And this is where we
got that 10,000 from.

1015
00:44:43,170 --> 00:44:45,140
If you have a 6 digit
code, you of course

1016
00:44:45,140 --> 00:44:49,840
just add this here, which is
another 10, and another 10, which

1017
00:44:49,840 --> 00:44:52,360
means we can just add another 0.

1018
00:44:52,360 --> 00:44:55,600
And now, we're up to a
million possibilities.

1019
00:44:55,600 --> 00:45:01,520
>> So as an engineer, if 6 is too
few, a million-- that still

1020
00:45:01,520 --> 00:45:04,450
feels relatively low, especially
if you can do 10 per second.

1021
00:45:04,450 --> 00:45:07,600
It gets a little boring, but
you can do it via brute force.

1022
00:45:07,600 --> 00:45:12,717
What might be better
than a 6 digit passcode?

1023
00:45:12,717 --> 00:45:13,300
What's better?

1024
00:45:13,300 --> 00:45:15,900
>> AUDIENCE: [INAUDIBLE] digits or letters
and different combinations [INAUDIBLE].

1025
00:45:15,900 --> 00:45:16,330
>> DAVID MALAN: Yeah.

1026
00:45:16,330 --> 00:45:17,830
So let's take both of those in turn.

1027
00:45:17,830 --> 00:45:21,970
So slightly better than a six
digit passcode might be, of course,

1028
00:45:21,970 --> 00:45:25,710
a seven digit passcode, which gives
us 10 million possibilities, and just

1029
00:45:25,710 --> 00:45:26,707
an additional digit.

1030
00:45:26,707 --> 00:45:29,790
Better than that though would be an 8
digit passcode, 9 digit passcode, 10

1031
00:45:29,790 --> 00:45:30,630
digit passcode.

1032
00:45:30,630 --> 00:45:31,630
>> But push back, now.

1033
00:45:31,630 --> 00:45:34,460
Now, you're not the engineer
or the security person.

1034
00:45:34,460 --> 00:45:37,220
Now you're the product manager
or the marketing person.

1035
00:45:37,220 --> 00:45:42,270
Why is a seven digit passcode not
better than a six digit passcode

1036
00:45:42,270 --> 00:45:46,001
for some definition of "better"?

1037
00:45:46,001 --> 00:45:47,626
AUDIENCE: It takes longer for the user.

1038
00:45:47,626 --> 00:45:48,130
DAVID MALAN: Yeah.

1039
00:45:48,130 --> 00:45:49,560
It takes longer for the user.

1040
00:45:49,560 --> 00:45:51,690
It takes an additional click.

1041
00:45:51,690 --> 00:45:55,237
And slightly more compellingly
too, I would say, is what?

1042
00:45:55,237 --> 00:45:56,112
AUDIENCE: [INAUDIBLE]

1043
00:45:56,112 --> 00:45:56,862
DAVID MALAN: Yeah.

1044
00:45:56,862 --> 00:45:59,712
It's slightly harder to remember
the longer and longer it gets.

1045
00:45:59,712 --> 00:46:01,920
We humans, at least in the
US, have kind of maxed out

1046
00:46:01,920 --> 00:46:03,239
at 10 digits for phone numbers.

1047
00:46:03,239 --> 00:46:06,030
And even that, I know like three
people's phone numbers these days.

1048
00:46:06,030 --> 00:46:07,740
So that's kind of a wash.

1049
00:46:07,740 --> 00:46:11,870
>> So there's a point where it's just not
a good user experience-- or UX would

1050
00:46:11,870 --> 00:46:13,600
be the trendy way of saying that.

1051
00:46:13,600 --> 00:46:15,620
So what's better than just using digits?

1052
00:46:15,620 --> 00:46:18,320
Well, instead of 10
possibilities, why don't we just

1053
00:46:18,320 --> 00:46:23,830
get more clever-- and instead
of using 10 digits, 0 through 9.

1054
00:46:23,830 --> 00:46:27,340
>> How else could we make
a 6 digit passcode--

1055
00:46:27,340 --> 00:46:30,470
a 6 symbol passcode more secure?

1056
00:46:30,470 --> 00:46:31,680
What did you propose?

1057
00:46:31,680 --> 00:46:32,380
Letters.

1058
00:46:32,380 --> 00:46:37,890
So instead of maybe digits, why don't we
just do letters, like 26 times 26 times

1059
00:46:37,890 --> 00:46:41,130
26-- and wow, this is actually
getting pretty big fast.

1060
00:46:41,130 --> 00:46:43,760
>> So if I go here-- this
is my little calculator.

1061
00:46:43,760 --> 00:46:48,690
And if I do 10 times 10 times
10 times 10 times 10 times 10.

1062
00:46:48,690 --> 00:46:50,850
That's where we got the
million possibilities

1063
00:46:50,850 --> 00:46:52,450
from for a 6 digit passcode.

1064
00:46:52,450 --> 00:46:58,280
But if instead we're doing 26
times 26 times 26 times another 26,

1065
00:46:58,280 --> 00:47:05,980
26, 26-- this is now giving
us 308 million possibilities.

1066
00:47:05,980 --> 00:47:10,116
>> And is that reasonable to
switch from numbers to letters

1067
00:47:10,116 --> 00:47:11,240
and still have it 6 digits?

1068
00:47:11,240 --> 00:47:14,030
This means you need a 6 letter word.

1069
00:47:14,030 --> 00:47:17,241
Most of us could probably remember a
six digit English or some other language

1070
00:47:17,241 --> 00:47:17,740
word.

1071
00:47:17,740 --> 00:47:19,010
That's pretty reasonable.

1072
00:47:19,010 --> 00:47:22,220
>> But we don't need to restrict
ourselves to just letters.

1073
00:47:22,220 --> 00:47:23,930
Why don't I get a little more ambitious?

1074
00:47:23,930 --> 00:47:29,080
What might be slightly
better than letters here?

1075
00:47:29,080 --> 00:47:31,667
Be the engineer proposing
an even better solution.

1076
00:47:31,667 --> 00:47:32,574
>> AUDIENCE: [INAUDIBLE]

1077
00:47:32,574 --> 00:47:34,240
DAVID MALAN: A combination-- characters.

1078
00:47:34,240 --> 00:47:39,920
So not just 26 letters, but if I add
back those numbers from before-- well,

1079
00:47:39,920 --> 00:47:42,760
everything's going wrong-- that's 36.

1080
00:47:42,760 --> 00:47:43,990
That's still 26.

1081
00:47:43,990 --> 00:47:47,900
That's 36 times 36 times-- and so forth.

1082
00:47:47,900 --> 00:47:49,360
So that's getting bigger.

1083
00:47:49,360 --> 00:47:54,980
>> How much bigger can we get this
address space, as someone might say?

1084
00:47:54,980 --> 00:47:57,300
What else could you add in
besides letters and numbers?

1085
00:47:57,300 --> 00:47:58,065
I'm up to 36.

1086
00:47:58,065 --> 00:48:00,196
I'm 26, a through z.

1087
00:48:00,196 --> 00:48:01,120
>> AUDIENCE: [INAUDIBLE]

1088
00:48:01,120 --> 00:48:01,870
DAVID MALAN: Yeah.

1089
00:48:01,870 --> 00:48:03,510
So we can really go
crazy with the keyboard.

1090
00:48:03,510 --> 00:48:05,270
Or even more simply,
we can keep it simpler.

1091
00:48:05,270 --> 00:48:05,660
>> AUDIENCE: [INAUDIBLE]

1092
00:48:05,660 --> 00:48:06,430
>> DAVID MALAN: Yeah.

1093
00:48:06,430 --> 00:48:10,070
If we go uppercase and
lowercase, now I have 26 plus 26.

1094
00:48:10,070 --> 00:48:13,300
So that's 52 already-- plus another 10.

1095
00:48:13,300 --> 00:48:15,080
That's 62.

1096
00:48:15,080 --> 00:48:17,200
And just to see the
implications of this,

1097
00:48:17,200 --> 00:48:18,970
now, let's just do another bit of math.

1098
00:48:18,970 --> 00:48:24,640
So 62 times 62 times 62
times 62 times 62 times 62.

1099
00:48:24,640 --> 00:48:30,850
That now is giving me 56
billion possibilities.

1100
00:48:30,850 --> 00:48:32,960
>> And it's still kind of reasonable.

1101
00:48:32,960 --> 00:48:36,230
Could a human remember
a 6 symbol-- where

1102
00:48:36,230 --> 00:48:39,380
a symbol is just a letter or
a number character password?

1103
00:48:39,380 --> 00:48:39,880
Probably.

1104
00:48:39,880 --> 00:48:42,600
That doesn't feel all that unreasonable.

1105
00:48:42,600 --> 00:48:44,010
So what more can we add in?

1106
00:48:44,010 --> 00:48:48,600
And as an aside, has anyone here,
ever heard the phrase base 64?

1107
00:48:48,600 --> 00:48:49,470
Base 64?

1108
00:48:49,470 --> 00:48:52,386
>> So we'll come back to this tomorrow
when we talk about representation.

1109
00:48:52,386 --> 00:48:54,910
Long story short, all of us
humans in the room most likely

1110
00:48:54,910 --> 00:48:57,560
understand base 10, the
so-called decimal system.

1111
00:48:57,560 --> 00:49:00,509
And all of us in this room
count using 0s through 9s.

1112
00:49:00,509 --> 00:49:03,300
We're going to see tomorrow, in
more detail, that a computer counts

1113
00:49:03,300 --> 00:49:06,410
using only 0s and 1s, the
so-called binary system.

1114
00:49:06,410 --> 00:49:08,410
So dec-- decimal-- is 10.

1115
00:49:08,410 --> 00:49:09,980
Bi-- binary-- is 2.

1116
00:49:09,980 --> 00:49:13,250
>> Turns out there's also base 64 for
which there isn't, to my knowledge,

1117
00:49:13,250 --> 00:49:14,340
a really fancy word.

1118
00:49:14,340 --> 00:49:19,030
But that means that you have
not 0 through 1 or 0 through 9,

1119
00:49:19,030 --> 00:49:22,090
you essentially have 0 through 64.

1120
00:49:22,090 --> 00:49:24,450
But you use letters in that mix.

1121
00:49:24,450 --> 00:49:28,750
And so we'll actually see that the
means by which computers, for instance,

1122
00:49:28,750 --> 00:49:32,620
attach files in an email these
days-- an email, of course,

1123
00:49:32,620 --> 00:49:37,130
might have an image on it-- maybe
even a sound or a movie file.

1124
00:49:37,130 --> 00:49:38,630
But email is just text.

1125
00:49:38,630 --> 00:49:43,360
>> It turns out that you can represent
things like music, and videos,

1126
00:49:43,360 --> 00:49:46,210
and pictures and the like
as text using something

1127
00:49:46,210 --> 00:49:49,940
called base 64 where you use not only
lowercase letters, and upper case

1128
00:49:49,940 --> 00:49:55,770
letters, and numbers, but
also the underscore character

1129
00:49:55,770 --> 00:49:58,030
and the slash on a keyboard.

1130
00:49:58,030 --> 00:49:59,322
So more on that to come.

1131
00:49:59,322 --> 00:50:00,780
So this is just getting really big.

1132
00:50:00,780 --> 00:50:02,970
And now, as the security
researcher, how could you

1133
00:50:02,970 --> 00:50:06,327
make a pass code even more secure?

1134
00:50:06,327 --> 00:50:09,160
We're now using lower case letters,
upper case letters, and numbers.

1135
00:50:09,160 --> 00:50:11,118
And you proposed, Victoria,
just a moment ago--

1136
00:50:11,118 --> 00:50:12,890
AUDIENCE: [INAUDIBLE]

1137
00:50:12,890 --> 00:50:14,140
DAVID MALAN: Dots are symbols.

1138
00:50:14,140 --> 00:50:15,750
And now, we're really just
kind of getting crazy.

1139
00:50:15,750 --> 00:50:17,900
We're using all of the
keys on the keyboard.

1140
00:50:17,900 --> 00:50:21,080
And let me estimate
that there are 128, give

1141
00:50:21,080 --> 00:50:23,990
or take, possibilities
on a typical keyboard,

1142
00:50:23,990 --> 00:50:25,490
depending on your language and such.

1143
00:50:25,490 --> 00:50:27,140
And there might even be more than that.

1144
00:50:27,140 --> 00:50:31,090
>> So now, let's still assume that
we're only using a 6 digit passcode

1145
00:50:31,090 --> 00:50:33,900
and that's why I have 6 of those 128.

1146
00:50:33,900 --> 00:50:36,261
Let's see if I can pronounce this now.

1147
00:50:36,261 --> 00:50:36,760
All right.

1148
00:50:36,760 --> 00:50:42,370
So that's millions, billions--
four quadrillion possibilities,

1149
00:50:42,370 --> 00:50:44,840
if I counted this
correctly-- four quadrillion.

1150
00:50:44,840 --> 00:50:47,850
Let me just double check, lest
I be exaggerating our security.

1151
00:50:47,850 --> 00:50:51,151
>> So that's hundreds of thousands,
millions-- sorry, trillions.

1152
00:50:51,151 --> 00:50:52,900
I overestimated by a
factor of a thousand.

1153
00:50:52,900 --> 00:50:54,350
My apologies.

1154
00:50:54,350 --> 00:50:56,480
4 trillion possibilities.

1155
00:50:56,480 --> 00:50:57,760
So that's more secure, right?

1156
00:50:57,760 --> 00:51:02,530
Especially when we began this discussion
with 1 out of 10,000 possible codes.

1157
00:51:02,530 --> 00:51:04,200
Now, we're up to 4 trillion.

1158
00:51:04,200 --> 00:51:08,080
>> Now, does this mean a
phone is "secure" if it

1159
00:51:08,080 --> 00:51:12,640
is using a passcode that is 6
characters long, each of which

1160
00:51:12,640 --> 00:51:15,980
can be a number, or a letter, or
some funky symbol on the keyboard?

1161
00:51:15,980 --> 00:51:22,070
Is a phone secure now if this is
in fact what the suspect was using?

1162
00:51:22,070 --> 00:51:23,030
>> AUDIENCE: [INAUDIBLE]

1163
00:51:23,030 --> 00:51:23,780
DAVID MALAN: Yeah.

1164
00:51:23,780 --> 00:51:24,946
And that's a perfect answer.

1165
00:51:24,946 --> 00:51:28,210
You conditionally explained
that by reasonable standards--

1166
00:51:28,210 --> 00:51:33,090
probabilistically, you're not going
to get into this phone anytime soon.

1167
00:51:33,090 --> 00:51:37,240
However there is a chance, if small
chance-- one out of 4 trillion--

1168
00:51:37,240 --> 00:51:40,450
that you might actually get the
answer right on the first time.

1169
00:51:40,450 --> 00:51:43,680
>> And the reality is, too, that if
this suspect is like most humans--

1170
00:51:43,680 --> 00:51:45,930
probably many of us in
his room-- he probably

1171
00:51:45,930 --> 00:51:50,620
did not choose some crazy password with
funky symbols on the key because why?

1172
00:51:50,620 --> 00:51:53,940
Most of us wouldn't remember
something that's so funky as that.

1173
00:51:53,940 --> 00:51:58,590
And so it probably is maybe someone's
birthday, or some word, or some phrase,

1174
00:51:58,590 --> 00:52:00,550
or something more memorable.

1175
00:52:00,550 --> 00:52:05,270
>> So it's probably not even as "secure"
as it might be mathematically.

1176
00:52:05,270 --> 00:52:07,130
So where does this leave things?

1177
00:52:07,130 --> 00:52:11,490
It remains to be seen what
Apple is going to agree to here.

1178
00:52:11,490 --> 00:52:15,480
But it certainly has implications
more broadly for society.

1179
00:52:15,480 --> 00:52:18,380
But the takeaways for today
are not so much the legalities,

1180
00:52:18,380 --> 00:52:22,360
not so much the ethics, or any of that,
but really the understanding of what's

1181
00:52:22,360 --> 00:52:23,340
actually going on.

1182
00:52:23,340 --> 00:52:25,880
>> And when you read something
like this to think to yourself,

1183
00:52:25,880 --> 00:52:28,520
is this an author just using
buzzwords, is there actually

1184
00:52:28,520 --> 00:52:32,797
technical meat to this comment, and
what might I go and Google in this case?

1185
00:52:32,797 --> 00:52:35,380
And in fact, probably one of the
more technical things in here

1186
00:52:35,380 --> 00:52:39,850
was this mention of RAM or memory,
and that was simply for, presumably,

1187
00:52:39,850 --> 00:52:43,450
the legal issue to which they allude.

1188
00:52:43,450 --> 00:52:45,500
>> A secure enclave is,
I think, Kareem, you

1189
00:52:45,500 --> 00:52:48,310
mentioned earlier this idea of
exponential backoff, so to speak--

1190
00:52:48,310 --> 00:52:49,768
or I put those words in your mouth.

1191
00:52:49,768 --> 00:52:51,410
And that's a feature not in this phone.

1192
00:52:51,410 --> 00:52:54,600
It apparently just has
the 80 millisecond delay

1193
00:52:54,600 --> 00:52:59,190
so it doesn't get worse, and
worse, and worse, over time.

1194
00:52:59,190 --> 00:52:59,760
All right.

1195
00:52:59,760 --> 00:53:01,828
Any questions?

1196
00:53:01,828 --> 00:53:02,950
Yeah, Dan.

1197
00:53:02,950 --> 00:53:07,877
>> AUDIENCE: If you don't mind me asking,
where do you stand on the issue?

1198
00:53:07,877 --> 00:53:09,960
DAVID MALAN: I would side,
absolutely, with Apple.

1199
00:53:09,960 --> 00:53:12,990
I think math is not something
that you should poke holes in.

1200
00:53:12,990 --> 00:53:17,700
And I think the reality is,
as even this article cites,

1201
00:53:17,700 --> 00:53:20,620
you can poke holes in the iPhone,
you can poke holes in the Android,

1202
00:53:20,620 --> 00:53:22,770
but there will always
be some alternative

1203
00:53:22,770 --> 00:53:25,320
that a particularly
smart adversary can use.

1204
00:53:25,320 --> 00:53:27,960
>> So these kinds of measures
really just protect us

1205
00:53:27,960 --> 00:53:32,340
against the dummies-- the lesser
adversaries, which has value,

1206
00:53:32,340 --> 00:53:34,150
but the reality is a
determined adversary

1207
00:53:34,150 --> 00:53:38,250
will absolutely keep encrypting his
or her data by some other mechanism,

1208
00:53:38,250 --> 00:53:40,887
whether it's via a mobile
application, a desktop application.

1209
00:53:40,887 --> 00:53:42,720
I think this is inevitable,
and I also think

1210
00:53:42,720 --> 00:53:44,937
this is a good thing as
a matter of principle.

1211
00:53:44,937 --> 00:53:49,807
>> AUDIENCE: My question is, at
the end of the day, [INAUDIBLE]

1212
00:53:49,807 --> 00:53:52,250
there is the one guy who
can access everything.

1213
00:53:52,250 --> 00:53:53,360
>> DAVID MALAN: Yeah.

1214
00:53:53,360 --> 00:53:55,624
>> AUDIENCE: So is it easy
for FBI [INAUDIBLE]

1215
00:53:55,624 --> 00:54:01,340
or somebody else instead of
other companies [INAUDIBLE]?

1216
00:54:01,340 --> 00:54:02,260
>> DAVID MALAN: Yeah.

1217
00:54:02,260 --> 00:54:04,770
And I think, especially in
this country, at least where

1218
00:54:04,770 --> 00:54:08,840
there were the recent revelations
as to just how far the NSA has

1219
00:54:08,840 --> 00:54:12,910
been going that I, especially nowadays,
don't buy the argument that we'll just

1220
00:54:12,910 --> 00:54:15,284
use it in this particular case.

1221
00:54:15,284 --> 00:54:16,700
I think that sets a bad precedent.

1222
00:54:16,700 --> 00:54:22,150
>> And already, there is a fundamental
paranoia we should have.

1223
00:54:22,150 --> 00:54:25,900
All of us, like chumps, if
you will, are walking around

1224
00:54:25,900 --> 00:54:32,370
with cameras, and microphones, and GPS
responders in our pockets, willingly,

1225
00:54:32,370 --> 00:54:36,830
telling someone potentially, even
if it's just Apple or just Google,

1226
00:54:36,830 --> 00:54:38,560
where we are at all times.

1227
00:54:38,560 --> 00:54:41,900
And there really is nothing stopping
Apple or some malicious engineer

1228
00:54:41,900 --> 00:54:46,420
at Apple from somehow
embedding in iOS a feature that

1229
00:54:46,420 --> 00:54:52,030
only turns on David Malan's microphone
24/7, and sends that data up to Apple.

1230
00:54:52,030 --> 00:54:54,800
>> And in fact, an interesting
side note here this

1231
00:54:54,800 --> 00:54:57,270
is kind of sort of already
happening as a "feature."

1232
00:54:57,270 --> 00:55:02,670
If you read the news about a year
ago, Samsung started, rightfully so,

1233
00:55:02,670 --> 00:55:06,142
to take some flak in the press because
they have these "smart TVs," where

1234
00:55:06,142 --> 00:55:08,100
as best I can tell "smart
TV" really just means

1235
00:55:08,100 --> 00:55:09,680
"TV with bad user interface."

1236
00:55:09,680 --> 00:55:12,910
But a "smart TV," as a
function of hardware, typically

1237
00:55:12,910 --> 00:55:15,850
has a microphone and
a camera these days.

1238
00:55:15,850 --> 00:55:16,360
And why?

1239
00:55:16,360 --> 00:55:19,434
Why does a TV need a
microphone or a camera?

1240
00:55:19,434 --> 00:55:20,360
>> AUDIENCE: Skype.

1241
00:55:20,360 --> 00:55:22,302
>> DAVID MALAN: Skype, which
is reasonable if you

1242
00:55:22,302 --> 00:55:25,260
want to use it in a conference room
or at home for video conferencing--

1243
00:55:25,260 --> 00:55:26,610
pretty reasonable, pretty compelling.

1244
00:55:26,610 --> 00:55:27,770
>> AUDIENCE: Voice commands.

1245
00:55:27,770 --> 00:55:28,760
>> DAVID MALAN: Voice
commands-- if you want

1246
00:55:28,760 --> 00:55:31,780
to say change channel, lower
volume, raise volume, turn off.

1247
00:55:31,780 --> 00:55:35,390
That's not unreasonable, a la
Siri, and Google Now, and such.

1248
00:55:35,390 --> 00:55:36,280
Why else?

1249
00:55:36,280 --> 00:55:38,281
>> AUDIENCE: To spy on you.

1250
00:55:38,281 --> 00:55:39,030
DAVID MALAN: Yeah.

1251
00:55:39,030 --> 00:55:42,080
So that's what the
paranoid in us might say.

1252
00:55:42,080 --> 00:55:45,990
And the reality is, whether
by a bug or deliberate intent,

1253
00:55:45,990 --> 00:55:47,380
this is absolutely possible.

1254
00:55:47,380 --> 00:55:48,600
Let's give them some credit.

1255
00:55:48,600 --> 00:55:51,610

1256
00:55:51,610 --> 00:55:54,260
Why might you, as a user, actually
want a camera in your TV--

1257
00:55:54,260 --> 00:55:55,940
or what's the proposed feature there?

1258
00:55:55,940 --> 00:55:59,480

1259
00:55:59,480 --> 00:56:01,990
Why is there a camera
in your living room

1260
00:56:01,990 --> 00:56:04,480
or in your bedroom
staring down at you all--

1261
00:56:04,480 --> 00:56:06,084
>> AUDIENCE: Security [INAUDIBLE].

1262
00:56:06,084 --> 00:56:06,750
DAVID MALAN: OK.

1263
00:56:06,750 --> 00:56:08,570
Security.

1264
00:56:08,570 --> 00:56:09,620
You could argue that.

1265
00:56:09,620 --> 00:56:12,180
In this case, it's not
so much the consumer TVs

1266
00:56:12,180 --> 00:56:13,800
that are in the business of security.

1267
00:56:13,800 --> 00:56:17,550
In this case it's, because
of a [INAUDIBLE] feature.

1268
00:56:17,550 --> 00:56:20,267
Why is there a camera in a TV?

1269
00:56:20,267 --> 00:56:22,100
AUDIENCE: Video games
detecting [INAUDIBLE].

1270
00:56:22,100 --> 00:56:23,080
DAVID MALAN: OK.

1271
00:56:23,080 --> 00:56:23,710
Yeah.

1272
00:56:23,710 --> 00:56:24,330
Closer.

1273
00:56:24,330 --> 00:56:27,120
And some TVs probably do
that-- have built in games.

1274
00:56:27,120 --> 00:56:29,884
This-- and I frankly think is a
little stupid-- gesture control.

1275
00:56:29,884 --> 00:56:31,800
I think stupid insofar
as I don't really think

1276
00:56:31,800 --> 00:56:34,940
we're there yet where we're living
in the Jetsons where it just works.

1277
00:56:34,940 --> 00:56:38,160
Now, I think you probably look like an
idiot to your TV when it doesn't work.

1278
00:56:38,160 --> 00:56:41,110
>> But gesture control, whereby
the world is getting better,

1279
00:56:41,110 --> 00:56:44,650
incrementing a la Xbox Kinect, if you're
familiar with the video game system,

1280
00:56:44,650 --> 00:56:45,860
being able to detect motion.

1281
00:56:45,860 --> 00:56:48,610
So maybe this means lower the
volume, this means raise the volume,

1282
00:56:48,610 --> 00:56:51,140
maybe this means swipe left
to change channel, swipe right

1283
00:56:51,140 --> 00:56:52,180
to change channels.

1284
00:56:52,180 --> 00:56:55,000
>> This is one of the reasons-- this
is the purported reason that they

1285
00:56:55,000 --> 00:56:56,000
have the thing in there.

1286
00:56:56,000 --> 00:56:58,572
But what Samsung took some
flak for just a few months ago

1287
00:56:58,572 --> 00:57:01,280
was that if you read their privacy
policy, which no one of course

1288
00:57:01,280 --> 00:57:05,000
is going to do, they encourage
you in their privacy policy

1289
00:57:05,000 --> 00:57:08,410
not to have private conversations
in the vicinity of your TV.

1290
00:57:08,410 --> 00:57:09,980
>> [LAUGHTER]

1291
00:57:09,980 --> 00:57:12,500
>> And we laugh, but like
it's actually there.

1292
00:57:12,500 --> 00:57:15,440
And that is because in order
to implement this feature,

1293
00:57:15,440 --> 00:57:17,320
the TV is always listening.

1294
00:57:17,320 --> 00:57:19,810
It has to be-- or it's always watching.

1295
00:57:19,810 --> 00:57:22,900
And even if you have some
defense mechanism in place--

1296
00:57:22,900 --> 00:57:26,360
kind of like Siri where you have
to say, hey, Siri, or OK, Google,

1297
00:57:26,360 --> 00:57:31,050
or whatever-- the TV still has
to be listening 24/7 for you

1298
00:57:31,050 --> 00:57:33,920
to say, hey, Siri, or OK, Google.

1299
00:57:33,920 --> 00:57:36,350
So hopefully, that's all staying local.

1300
00:57:36,350 --> 00:57:39,570
And there's no technical reason why
it couldn't stay local, software

1301
00:57:39,570 --> 00:57:40,580
updates aside.

1302
00:57:40,580 --> 00:57:42,790
>> But in reality, very
often, Siri and Google

1303
00:57:42,790 --> 00:57:45,849
alike are sending these
data to the cloud,

1304
00:57:45,849 --> 00:57:48,140
so to speak, where they get
processed there by smarter,

1305
00:57:48,140 --> 00:57:52,280
faster, constantly updated computers,
and then send the responses back down

1306
00:57:52,280 --> 00:57:54,600
to the TV.

1307
00:57:54,600 --> 00:57:59,730
Oh and the fun thing here-- we took a
look at this for another class I teach.

1308
00:57:59,730 --> 00:58:01,760
We'll see this a little later today.

1309
00:58:01,760 --> 00:58:05,420
>> There's something in the world
called security and encryption, which

1310
00:58:05,420 --> 00:58:06,670
we're getting to right now.

1311
00:58:06,670 --> 00:58:11,710
And in theory, there's something called
HTTP and HTTPS, the latter of which

1312
00:58:11,710 --> 00:58:12,310
is secure.

1313
00:58:12,310 --> 00:58:13,790
The S is for Security, and
we'll come back to that.

1314
00:58:13,790 --> 00:58:15,498
And then they operate
on something called

1315
00:58:15,498 --> 00:58:19,510
different ports, different numeric
values inside of a computer signifies

1316
00:58:19,510 --> 00:58:21,520
if this is secure or
not secure typically.

1317
00:58:21,520 --> 00:58:26,890
>> Samsung, I believe, in this case, was
using the "secure port," so to speak.

1318
00:58:26,890 --> 00:58:29,170
They were using the
secure address, but they

1319
00:58:29,170 --> 00:58:31,420
were using it to send encrypted data.

1320
00:58:31,420 --> 00:58:35,000
So some security researchers essentially
connected a device to their TV

1321
00:58:35,000 --> 00:58:37,800
and realized when they
spoke commands to their TV,

1322
00:58:37,800 --> 00:58:41,740
it was being uploaded to the
cloud through the correct channel,

1323
00:58:41,740 --> 00:58:44,500
so to speak, but completely
unencrypted, which

1324
00:58:44,500 --> 00:58:48,250
meant anyone in the vicinity or anyone
on the internet between points A and B

1325
00:58:48,250 --> 00:58:50,830
could be seeing and listening
to your voice commands

1326
00:58:50,830 --> 00:58:53,110
from your living room or your bedroom.

1327
00:58:53,110 --> 00:58:56,710
>> So there too, not only are we vulnerable
potentially to maliciousness, also

1328
00:58:56,710 --> 00:58:59,680
just stupidity and bugs, in this case.

1329
00:58:59,680 --> 00:59:01,840
So these are the kinds
of things to be ware.

1330
00:59:01,840 --> 00:59:03,925
And again, the goals
for today and tomorrow

1331
00:59:03,925 --> 00:59:06,300
are to understand not necessarily
how you would implement

1332
00:59:06,300 --> 00:59:08,860
that underneath the hood,
but just reason backwards,

1333
00:59:08,860 --> 00:59:12,400
if my TV is responding to
gesture control and my words,

1334
00:59:12,400 --> 00:59:15,220
I'm guessing my TV is
not so sophisticated

1335
00:59:15,220 --> 00:59:17,630
as to have the entire
English or the entire Spanish

1336
00:59:17,630 --> 00:59:21,540
or whatever language I speak dictionary
built into it constantly updated.

1337
00:59:21,540 --> 00:59:23,830
It's probably easier just
to send those commands up

1338
00:59:23,830 --> 00:59:26,870
to some server-- Google, or
Apple, or Samsung, or the like.

1339
00:59:26,870 --> 00:59:28,810
And indeed, that's what's
typically happening.

1340
00:59:28,810 --> 00:59:34,580
So mind what you say in front of
your TVs starting tonight perhaps.

1341
00:59:34,580 --> 00:59:35,290
>> All right.

1342
00:59:35,290 --> 00:59:38,900
So that leads us then to encryption
with a more technical look.

1343
00:59:38,900 --> 00:59:42,300
And we won't go too deep a dive into
this, but this article we looked at

1344
00:59:42,300 --> 00:59:45,770
did mention something called
AES-- Advanced Encryption

1345
00:59:45,770 --> 00:59:47,640
Standard, is what it stands for.

1346
00:59:47,640 --> 00:59:53,361
And it made mention of something
juicy, a 256-bit AES key-- secret key.

1347
00:59:53,361 --> 00:59:55,985
And I'll just pull it up if you're
curious to see where it was.

1348
00:59:55,985 --> 01:00:01,500
It was in the-- How Would They Do That.

1349
01:00:01,500 --> 01:00:05,670
So somewhere inside of an
iPhone and an Android phone,

1350
01:00:05,670 --> 01:00:09,010
presumably, is some kind of secret key.

1351
01:00:09,010 --> 01:00:12,630
And it's this secret key
that keeps data secure.

1352
01:00:12,630 --> 01:00:15,780
>> And in fact, have any of you
and your iPhones ever gone

1353
01:00:15,780 --> 01:00:22,180
to Settings-- I think, Settings,
maybe General, and then Erase iPhone?

1354
01:00:22,180 --> 01:00:23,430
It's somewhere under Settings.

1355
01:00:23,430 --> 01:00:25,950
You can erase your
iPhone, and it tells you

1356
01:00:25,950 --> 01:00:27,910
that you're going to erase it securely.

1357
01:00:27,910 --> 01:00:32,465
And what does it mean, typically, to
erase a phone or a computer securely?

1358
01:00:32,465 --> 01:00:34,779

1359
01:00:34,779 --> 01:00:37,570
And actually, let me see if I can
give you just a quick screenshot.

1360
01:00:37,570 --> 01:00:38,653
We can probably find this.

1361
01:00:38,653 --> 01:00:42,700
So, iphone erase securely
setting screenshot.

1362
01:00:42,700 --> 01:00:45,285
Let's see if we can
just find a quick photo.

1363
01:00:45,285 --> 01:00:48,070

1364
01:00:48,070 --> 01:00:50,500
Erase data-- that's not-- here it is.

1365
01:00:50,500 --> 01:00:53,042
>> So this is the screen I was thinking of.

1366
01:00:53,042 --> 01:00:54,750
You can generally, on
an iPhone, navigate

1367
01:00:54,750 --> 01:00:56,680
to a screen that looks like this.

1368
01:00:56,680 --> 01:00:59,730
And Erase All Content and Settings--
if you click that, it tells you it's

1369
01:00:59,730 --> 01:01:00,730
going to do it securely.

1370
01:01:00,730 --> 01:01:04,881
What does securely mean
in a phone or a computer?

1371
01:01:04,881 --> 01:01:05,380
Ben?

1372
01:01:05,380 --> 01:01:08,890
>> AUDIENCE: In a way that's difficult
to then go back and actually find it.

1373
01:01:08,890 --> 01:01:09,640
DAVID MALAN: Good.

1374
01:01:09,640 --> 01:01:12,780
So in a way that's difficult to go
back and find what you've erased.

1375
01:01:12,780 --> 01:01:15,430
So erasing it truly means erasing it.

1376
01:01:15,430 --> 01:01:18,330
And the industry does not
have a good history with this.

1377
01:01:18,330 --> 01:01:21,400
>> Back in the day, most of us
probably had PCs in some form.

1378
01:01:21,400 --> 01:01:22,640
Some of you still might.

1379
01:01:22,640 --> 01:01:27,230
Back in the day, when we still had
floppy disks and certain other media,

1380
01:01:27,230 --> 01:01:31,730
it was very common to run a format
command, or an erase command,

1381
01:01:31,730 --> 01:01:35,010
or a partition command, which are
all generally related to getting

1382
01:01:35,010 --> 01:01:37,750
a drive-- a disk ready for use.

1383
01:01:37,750 --> 01:01:42,520
>> And back in the day, I
can even visualize it now,

1384
01:01:42,520 --> 01:01:46,100
the DOS-- if you're familiar, the
command-- the black and white prompt

1385
01:01:46,100 --> 01:01:49,560
in Windows-- or even before Windows--
would yell at you in all capital

1386
01:01:49,560 --> 01:01:54,780
letters, ALL DATA WILL BE DESTROYED or
ALL DATA WILL BE ERASED-- complete lie.

1387
01:01:54,780 --> 01:01:58,370
It was a complete technical and
actual lie because, typically,

1388
01:01:58,370 --> 01:02:01,260
what a computer does-- even to
this day in most contexts is

1389
01:02:01,260 --> 01:02:04,300
that when you drag a file to
your Recycle bin or to your trash

1390
01:02:04,300 --> 01:02:08,500
can on Mac OS, or Windows, or what not--
we all probably know that it hasn't

1391
01:02:08,500 --> 01:02:09,950
actually been deleted yet, right?

1392
01:02:09,950 --> 01:02:13,807
You have to actually do what
to actually delete a file?

1393
01:02:13,807 --> 01:02:14,890
AUDIENCE: Empty the trash.

1394
01:02:14,890 --> 01:02:17,050
DAVID MALAN: You have to empty the
trash can or empty the Recycle bin.

1395
01:02:17,050 --> 01:02:19,437
We've all been taught that,
and that's the mental model

1396
01:02:19,437 --> 01:02:20,520
we have in the real world.

1397
01:02:20,520 --> 01:02:21,880
That is also a lie.

1398
01:02:21,880 --> 01:02:25,140
Almost always, by default these
days, when you empty your trash

1399
01:02:25,140 --> 01:02:27,890
or empty your Recycle bin, even
by going to the right menu option,

1400
01:02:27,890 --> 01:02:31,550
or right clicking, or Control clicking
and following good human intuition,

1401
01:02:31,550 --> 01:02:32,500
it's a lie.

1402
01:02:32,500 --> 01:02:36,730
>> All the computer is doing
is "forgetting" your file.

1403
01:02:36,730 --> 01:02:38,830
In other words, somewhere
inside of your computer,

1404
01:02:38,830 --> 01:02:41,000
you can think of there as
being a big cheat sheet,

1405
01:02:41,000 --> 01:02:44,780
a big Excel file, a big
table with rows and columns

1406
01:02:44,780 --> 01:02:49,580
that says a file called resume.doc
is at this location on my hard drive,

1407
01:02:49,580 --> 01:02:53,040
and a file called friends.text
is in this location,

1408
01:02:53,040 --> 01:02:56,780
and profilephoto.jpeg is at
this location in my hard drive.

1409
01:02:56,780 --> 01:02:59,890
>> So whole bunch of file names--
whole bunch of physical locations

1410
01:02:59,890 --> 01:03:01,160
inside of your computer.

1411
01:03:01,160 --> 01:03:04,900
And when a computer "erases"
a file, typically all it does

1412
01:03:04,900 --> 01:03:07,510
is it deletes that row
or crosses that out.

1413
01:03:07,510 --> 01:03:09,540
It leaves the file on the disk.

1414
01:03:09,540 --> 01:03:10,857
It just forgets where it is.

1415
01:03:10,857 --> 01:03:13,190
And that's useful because if
it's forgotten where it is,

1416
01:03:13,190 --> 01:03:15,730
it can reuse that space later on.

1417
01:03:15,730 --> 01:03:17,480
It can just put another
file on top of it.

1418
01:03:17,480 --> 01:03:19,479
And tomorrow again, we'll
talk about 0s and 1s--

1419
01:03:19,479 --> 01:03:21,900
that just means changing
some 0s to 1s, some 1s to 0s,

1420
01:03:21,900 --> 01:03:26,840
leaving some alone-- but generally,
reconstituting a file out of bits, 0s

1421
01:03:26,840 --> 01:03:28,010
and 1s.

1422
01:03:28,010 --> 01:03:30,390
>> So what is this actually doing?

1423
01:03:30,390 --> 01:03:34,410
Thankfully, in iOS's case, since Apple
actually is quite good at security,

1424
01:03:34,410 --> 01:03:38,340
even on Mac OS, erasing your
files does in fact do it securely.

1425
01:03:38,340 --> 01:03:39,010
But how?

1426
01:03:39,010 --> 01:03:42,520
Well in Mac OS and Windows, if
you have the right software,

1427
01:03:42,520 --> 01:03:45,750
what it will-- to erase
something securely

1428
01:03:45,750 --> 01:03:47,532
does have some technical meaning.

1429
01:03:47,532 --> 01:03:49,990
And again, we'll come back to
this in more detail tomorrow.

1430
01:03:49,990 --> 01:03:53,000
>> But to erase a file securely
does mean doing something to it

1431
01:03:53,000 --> 01:03:54,121
so it can't be recovered.

1432
01:03:54,121 --> 01:03:55,120
But what does that mean?

1433
01:03:55,120 --> 01:03:58,440
Well, if a file, for today's purposes,
is represented with 0s and 1s

1434
01:03:58,440 --> 01:04:00,680
somehow-- I have no idea
how, more on that tomorrow.

1435
01:04:00,680 --> 01:04:04,270
But 0s and 1s-- the way
you erase a file securely

1436
01:04:04,270 --> 01:04:08,350
is you maybe change all of
those 0s and 1s to just all 0s

1437
01:04:08,350 --> 01:04:13,060
or just all the 1s-- just scramble
them up completely randomly

1438
01:04:13,060 --> 01:04:16,170
so that if someone thereafter looks
at those 0s and 1s, it's meaningless.

1439
01:04:16,170 --> 01:04:18,378
And it's not recoverable
because you did it randomly,

1440
01:04:18,378 --> 01:04:20,730
or you made them all 0's or all 1's.

1441
01:04:20,730 --> 01:04:22,540
>> That's not actually what Apple does.

1442
01:04:22,540 --> 01:04:24,940
Because it turns out when
you erased your iPhone,

1443
01:04:24,940 --> 01:04:26,980
it doesn't take all that long.

1444
01:04:26,980 --> 01:04:31,560
No, in fact, if you erase a computer
hard drive, it might take an hour,

1445
01:04:31,560 --> 01:04:35,990
it might take three days to
literally change every 0 and 1

1446
01:04:35,990 --> 01:04:36,889
to some other value.

1447
01:04:36,889 --> 01:04:38,930
There's just a lot of bits
these days, especially

1448
01:04:38,930 --> 01:04:40,580
if you have a hard drive
that's one terabyte,

1449
01:04:40,580 --> 01:04:43,121
so to speak-- four terrabytes--
will take a really long time.

1450
01:04:43,121 --> 01:04:46,050
But Apple does it within a few
seconds-- maybe a couple minutes,

1451
01:04:46,050 --> 01:04:48,610
but reasonably quickly.

1452
01:04:48,610 --> 01:04:49,530
>> Now, why is that?

1453
01:04:49,530 --> 01:04:51,220
It all relates to the same discussion.

1454
01:04:51,220 --> 01:04:55,230
Apple, by default, keeps all of
the data on your phone encrypted--

1455
01:04:55,230 --> 01:04:57,480
scrambled in some way.

1456
01:04:57,480 --> 01:05:00,550
And so to erase a phone,
you don't necessarily

1457
01:05:00,550 --> 01:05:03,850
have to change the data--
because the general principle

1458
01:05:03,850 --> 01:05:05,970
of encryption-- the art
of scrambling information

1459
01:05:05,970 --> 01:05:08,540
or cryptography as the
science itself is called,

1460
01:05:08,540 --> 01:05:13,570
is that to an adversary
looking at encrypted data,

1461
01:05:13,570 --> 01:05:18,210
it should look random-- he or she should
not be able to glean any insights.

1462
01:05:18,210 --> 01:05:21,660
They should not be able to realize--
this person seems to use the word "the"

1463
01:05:21,660 --> 01:05:22,160
a lot.

1464
01:05:22,160 --> 01:05:24,860
Just because I see some pattern
emerging again and again--

1465
01:05:24,860 --> 01:05:29,100
it should look completely random
statistically to an adversary.

1466
01:05:29,100 --> 01:05:33,560
>> So by that logic, when Apple allows
you to erase all content in settings,

1467
01:05:33,560 --> 01:05:37,310
the data already looks random
to any person on the streets who

1468
01:05:37,310 --> 01:05:39,290
might look at your phone's data.

1469
01:05:39,290 --> 01:05:41,480
So they don't have to change your data.

1470
01:05:41,480 --> 01:05:45,577
All they have to do to erase your
phone is do what do you think?

1471
01:05:45,577 --> 01:05:47,410
AUDIENCE: [INAUDIBLE]
your code incorrectly.

1472
01:05:47,410 --> 01:05:47,630
DAVID MALAN: Yeah.

1473
01:05:47,630 --> 01:05:48,870
Well, you could do-- yes.

1474
01:05:48,870 --> 01:05:54,210
Physically, they could just type your
code in, 7 7 7 7, 10 times incorrectly.

1475
01:05:54,210 --> 01:05:57,980
But rather, you can just
forget the secret key.

1476
01:05:57,980 --> 01:06:02,620
So encryption is all about
having, generally, some secrets.

1477
01:06:02,620 --> 01:06:05,457
>> So much like you can't get into a
bank vault without a combination,

1478
01:06:05,457 --> 01:06:08,290
much like you can't get into your
front door without a physical key,

1479
01:06:08,290 --> 01:06:10,206
you can't get into your
hotel room without one

1480
01:06:10,206 --> 01:06:14,390
of those magnetic cards
or such, in theory,

1481
01:06:14,390 --> 01:06:17,810
there is something special something
that only you know or have that

1482
01:06:17,810 --> 01:06:19,564
allows you access to
some secure resource.

1483
01:06:19,564 --> 01:06:21,730
In the case of your phone,
it's the four digit code.

1484
01:06:21,730 --> 01:06:23,620
In the case of your hotel,
it's the little card key.

1485
01:06:23,620 --> 01:06:25,680
In the case of your home,
it's the physical key.

1486
01:06:25,680 --> 01:06:27,650
Any number of things can be a key.

1487
01:06:27,650 --> 01:06:30,520
>> But in computing, it's
almost always a number, where

1488
01:06:30,520 --> 01:06:32,260
a number is just a sequence of bits.

1489
01:06:32,260 --> 01:06:35,460
And again, a bit is just a 0 or
1, but more on that tomorrow.

1490
01:06:35,460 --> 01:06:42,840
So when Apple claims to be
using 256-bit AES secret key.

1491
01:06:42,840 --> 01:06:47,120
That just means that the secret
key inside of your computer

1492
01:06:47,120 --> 01:06:50,260
is something like 1011001100000.

1493
01:06:50,260 --> 01:06:53,480
>> I'm just making this up as
we go, and I won't bother

1494
01:06:53,480 --> 01:06:57,680
writing out 256 possible 0s and 1s.

1495
01:06:57,680 --> 01:07:01,660
And we'll see tomorrow how
this maps to an actual number.

1496
01:07:01,660 --> 01:07:04,620
But for now, just know it's a
really long pattern of 0s and 1s.

1497
01:07:04,620 --> 01:07:08,780
And that secret-- that's like a really
big magnetic card key for your hotel

1498
01:07:08,780 --> 01:07:12,790
room that only you have, or it's like
a really special metal key with lots

1499
01:07:12,790 --> 01:07:15,530
of little teeth that only you have.

1500
01:07:15,530 --> 01:07:18,040
>> How is this useful?

1501
01:07:18,040 --> 01:07:20,910
How was it useful to use a key?

1502
01:07:20,910 --> 01:07:22,050
Well, let's do this.

1503
01:07:22,050 --> 01:07:24,910
Let's start with truly a clean slate.

1504
01:07:24,910 --> 01:07:30,070
And let me propose, just as in a
little experiment here for say,

1505
01:07:30,070 --> 01:07:37,170
a moment-- how about we
take the word "hello."

1506
01:07:37,170 --> 01:07:39,560
>> And suppose that you're
back in middle school

1507
01:07:39,560 --> 01:07:42,080
and you want to send the
boy or girl across the aisle

1508
01:07:42,080 --> 01:07:45,090
that you have a crush on
a secret message, "hello,"

1509
01:07:45,090 --> 01:07:47,820
but you don't want to be
embarrassed if the teacher picks up

1510
01:07:47,820 --> 01:07:52,500
the scrap of paper that intercepts the
note that you're passing to him or her.

1511
01:07:52,500 --> 01:07:54,170
>> You want to encrypt this information.

1512
01:07:54,170 --> 01:07:57,360
You want to scramble it so it just
looks like you're writing nonsense.

1513
01:07:57,360 --> 01:07:59,490
And probably, it's something
juicier than "hello,"

1514
01:07:59,490 --> 01:08:01,650
but we'll just take the word "hello."

1515
01:08:01,650 --> 01:08:07,470
>> How could we go about encrypting this
message between the two little kids

1516
01:08:07,470 --> 01:08:09,010
on a piece of paper?

1517
01:08:09,010 --> 01:08:12,060
What should he or she
write instead of "hello"?

1518
01:08:12,060 --> 01:08:13,080
>> AUDIENCE: [INAUDIBLE]

1519
01:08:13,080 --> 01:08:13,717
>> DAVID MALAN: What's that?

1520
01:08:13,717 --> 01:08:14,970
>> AUDIENCE: Number of
letter in the alphabet.

1521
01:08:14,970 --> 01:08:17,053
>> DAVID MALAN: The number
of letter in the alphabet.

1522
01:08:17,053 --> 01:08:22,340
OK, so if it's a b c d e f g h, I could
maybe do something like 8 for that.

1523
01:08:22,340 --> 01:08:26,270
And a b c d e-- and I
can do the 5 for that.

1524
01:08:26,270 --> 01:08:30,970
And similarly, I can just come up with
a numeric mapping that would presumably

1525
01:08:30,970 --> 01:08:32,290
just confuse the teacher.

1526
01:08:32,290 --> 01:08:34,490
And he or she probably
doesn't have enough--

1527
01:08:34,490 --> 01:08:37,060
doesn't care enough to
actually figure out what it is.

1528
01:08:37,060 --> 01:08:40,569
So let's consider though, is it secure?

1529
01:08:40,569 --> 01:08:41,406
Why not?

1530
01:08:41,406 --> 01:08:43,310
>> AUDIENCE: Because it's easy to guess it.

1531
01:08:43,310 --> 01:08:45,220
If in case someone is really interested.

1532
01:08:45,220 --> 01:08:45,460
>> DAVID MALAN: Yeah.

1533
01:08:45,460 --> 01:08:47,918
If they are really interested
and if they have more numbers

1534
01:08:47,918 --> 01:08:51,060
to go with than just five-- if
there's like a whole paragraph--

1535
01:08:51,060 --> 01:08:54,779
and it just so happens that all of
the numbers are between 1 and 26,

1536
01:08:54,779 --> 01:08:56,350
that's kind of an interesting clue.

1537
01:08:56,350 --> 01:08:58,240
And you could brute force that.

1538
01:08:58,240 --> 01:09:01,062
Let's see if a is 1,
and b is 2, and c is 3.

1539
01:09:01,062 --> 01:09:03,020
And if not, maybe let's
try some other pairing.

1540
01:09:03,020 --> 01:09:05,310
But a determined teacher--
an adversarial teacher--

1541
01:09:05,310 --> 01:09:07,369
could certainly figure this out.

1542
01:09:07,369 --> 01:09:08,410
So what else could we do?

1543
01:09:08,410 --> 01:09:12,200
A simple encoding-- and this
truly is called a code--

1544
01:09:12,200 --> 01:09:16,020
not to be confused with programming
code or programming languages-- a code.

1545
01:09:16,020 --> 01:09:19,620
And in fact, if you recall stories
from yesteryear, especially

1546
01:09:19,620 --> 01:09:22,750
in the military, a code book--
a code book might literally

1547
01:09:22,750 --> 01:09:26,000
be a physical book that's got
two columns, one is a letter,

1548
01:09:26,000 --> 01:09:29,430
one is a number-- or some other such
symbol-- that just maps to the other.

1549
01:09:29,430 --> 01:09:32,140
And a code is a mapping
from one thing to another.

1550
01:09:32,140 --> 01:09:33,430
>> So that would be a code.

1551
01:09:33,430 --> 01:09:38,890
Encryption though-- or a cipher as you
might say-- is more of an algorithm.

1552
01:09:38,890 --> 01:09:39,960
It's a process.

1553
01:09:39,960 --> 01:09:41,529
It's not just something you look up.

1554
01:09:41,529 --> 01:09:46,899
You have to apply some logic to apply
encryption, or a cipher in this case.

1555
01:09:46,899 --> 01:09:50,010
So what's slightly more sophisticated,
do you think, than that?

1556
01:09:50,010 --> 01:09:53,560

1557
01:09:53,560 --> 01:09:57,786
What else could we do to send
the word "hello" semisecretly?

1558
01:09:57,786 --> 01:10:00,224
>> AUDIENCE: [INAUDIBLE]

1559
01:10:00,224 --> 01:10:00,890
DAVID MALAN: OK.

1560
01:10:00,890 --> 01:10:02,181
So we could write it backwards.

1561
01:10:02,181 --> 01:10:06,890
So we could do something
like o-l-l-e-h or such,

1562
01:10:06,890 --> 01:10:08,890
and it starts to look a
little more complicated.

1563
01:10:08,890 --> 01:10:10,302
So it's kind of scrambled.

1564
01:10:10,302 --> 01:10:12,260
And you have to know the
secret, and the secret

1565
01:10:12,260 --> 01:10:16,050
is "backwords" or "reverse" or
some sort of mechanism there.

1566
01:10:16,050 --> 01:10:17,300
>> But that is an algorithm.

1567
01:10:17,300 --> 01:10:20,450
That is a process where you have
to move this letter over here,

1568
01:10:20,450 --> 01:10:22,360
this letter over here,
this letter over here,

1569
01:10:22,360 --> 01:10:24,110
and you have to repeat
it again and again.

1570
01:10:24,110 --> 01:10:26,800
And we'll see tomorrow that this
repetition is something called

1571
01:10:26,800 --> 01:10:28,600
a loop, which is fairly
intuitive, but it's

1572
01:10:28,600 --> 01:10:30,570
very common in computer programming.

1573
01:10:30,570 --> 01:10:34,260
What else might we do?

1574
01:10:34,260 --> 01:10:37,592
>> AUDIENCE: You could increase the
first letter by 1, second letter by 2,

1575
01:10:37,592 --> 01:10:39,422
third letter by 3 [INAUDIBLE].

1576
01:10:39,422 --> 01:10:40,380
DAVID MALAN: Very nice.

1577
01:10:40,380 --> 01:10:42,720
So we could do something
like-- and increase them--

1578
01:10:42,720 --> 01:10:44,570
you mean like h becomes i.

1579
01:10:44,570 --> 01:10:46,480
And let me keep it
simple for the moment.

1580
01:10:46,480 --> 01:10:48,190
Maybe e becomes f.

1581
01:10:48,190 --> 01:10:52,150
And this becomes m m, and this is p.

1582
01:10:52,150 --> 01:10:55,850
>> Now, I'm kind of liking this
because now it doesn't jump out

1583
01:10:55,850 --> 01:10:57,070
at you what has happened.

1584
01:10:57,070 --> 01:10:58,920
And it looks like nonsense.

1585
01:10:58,920 --> 01:11:03,280
But in terms of the security
of this cipher, and the cipher

1586
01:11:03,280 --> 01:11:06,470
here is kind of like a plus
1 algorithm of just adding

1587
01:11:06,470 --> 01:11:08,500
1 letter to each of my own letters.

1588
01:11:08,500 --> 01:11:13,530
And just as a corner case,
what should I do if I hit z?

1589
01:11:13,530 --> 01:11:14,030
>> AUDIENCE: A.

1590
01:11:14,030 --> 01:11:14,270
>> DAVID MALAN: Yeah.

1591
01:11:14,270 --> 01:11:15,840
Probably just go back to a.

1592
01:11:15,840 --> 01:11:17,550
But what if I want an exclamation point?

1593
01:11:17,550 --> 01:11:19,340
Well, we'll have to come
back to that sometime.

1594
01:11:19,340 --> 01:11:21,131
So there's some corner
cases, so to speak--

1595
01:11:21,131 --> 01:11:24,100
things you need to anticipate if
you want to support those features.

1596
01:11:24,100 --> 01:11:28,996
But what is attackable about this?

1597
01:11:28,996 --> 01:11:31,620
It's obviously not that secure
because we sort of thought of it

1598
01:11:31,620 --> 01:11:33,010
and wrote it down super fast.

1599
01:11:33,010 --> 01:11:35,750
So presumably, a smart
adversary could do the opposite.

1600
01:11:35,750 --> 01:11:40,630
But what information is leaked
in this particular ciphertext?

1601
01:11:40,630 --> 01:11:43,010
Computer scientists
would call this cleartext

1602
01:11:43,010 --> 01:11:49,510
and this ciphertext-- ciphertext
meaning just scrambled or encrypted.

1603
01:11:49,510 --> 01:11:53,570
We're leaking information, so
to speak, with this ciphertext.

1604
01:11:53,570 --> 01:11:58,978
I know something about the
original word, right now.

1605
01:11:58,978 --> 01:12:00,370
>> AUDIENCE: Same number of letter.

1606
01:12:00,370 --> 01:12:01,870
DAVID MALAN: Same number of letters.

1607
01:12:01,870 --> 01:12:03,510
So that's leaking information.

1608
01:12:03,510 --> 01:12:07,170
I have sent my crush a five
letter word, it would seem.

1609
01:12:07,170 --> 01:12:08,080
And what else?

1610
01:12:08,080 --> 01:12:09,000
>> AUDIENCE: Yeah.

1611
01:12:09,000 --> 01:12:09,870
There are letters.

1612
01:12:09,870 --> 01:12:11,676
>> DAVID MALAN: They're still letters.

1613
01:12:11,676 --> 01:12:14,049
>> AUDIENCE: Third and
fourth characters repeat.

1614
01:12:14,049 --> 01:12:16,340
DAVID MALAN: Yeah, the third
and fourth letters repeat.

1615
01:12:16,340 --> 01:12:20,010
And this is very
common-- this realization

1616
01:12:20,010 --> 01:12:22,770
for what's called a frequency analysis.

1617
01:12:22,770 --> 01:12:25,860
And I used the word "the,"
anticipating this earlier.

1618
01:12:25,860 --> 01:12:27,802
"The" is a very common English word.

1619
01:12:27,802 --> 01:12:30,260
And so if we actually had a
paragraph or a whole essay that

1620
01:12:30,260 --> 01:12:34,160
was somehow encrypted, and I kept seeing
the same patterns of three letters, not

1621
01:12:34,160 --> 01:12:38,720
t-h-e, but like x-y-z or something like
that, I might just guess, on a hunch,

1622
01:12:38,720 --> 01:12:41,410
based on the popularity
of "the" in English

1623
01:12:41,410 --> 01:12:47,030
that maybe I should start replacing
every x-y-z with t-h-e, respectively--

1624
01:12:47,030 --> 01:12:48,750
and you chip away at the problem.

1625
01:12:48,750 --> 01:12:51,830
>> And in fact, if you've ever seen
a movie about cryptographers,

1626
01:12:51,830 --> 01:12:54,420
especially during military times,
cracking codes-- a lot of it

1627
01:12:54,420 --> 01:12:59,070
is this trial and error, and leveraging
assumptions, and taking guesses,

1628
01:12:59,070 --> 01:13:00,150
and seeing where it goes.

1629
01:13:00,150 --> 01:13:04,430
And in fact, m-m-- we sometimes
see m-m in the English word,

1630
01:13:04,430 --> 01:13:06,320
so maybe this is unchanged.

1631
01:13:06,320 --> 01:13:14,830
We see e-e, we see o-o, we see
l-l, we don't really see y-x.

1632
01:13:14,830 --> 01:13:18,230
And there's bunches of others I could
probably contrive that we never see.

1633
01:13:18,230 --> 01:13:20,757
So we've narrowed our
search space, so to speak.

1634
01:13:20,757 --> 01:13:23,090
In other words, if the problem
initially feels this big,

1635
01:13:23,090 --> 01:13:25,420
as soon as you start ruling
out possibilities or ruling

1636
01:13:25,420 --> 01:13:29,840
in possibilities, starts to get a little
more tenable, a little more solvable.

1637
01:13:29,840 --> 01:13:33,170
And in fact, this is an
example actually of something

1638
01:13:33,170 --> 01:13:38,530
called a Caesar cipher, where a Caesar
cipher is a rotational cipher where

1639
01:13:38,530 --> 01:13:40,450
one letter becomes
another and you just add

1640
01:13:40,450 --> 01:13:44,670
uniformly the same number
of changes to each letter.

1641
01:13:44,670 --> 01:13:48,330
And Dan actually hinted at something
slightly more sophisticated earlier,

1642
01:13:48,330 --> 01:13:52,700
which we might add, for instance,
1 letter to the first letter.

1643
01:13:52,700 --> 01:13:55,390
e-f-- Maybe this becomes g, two away.

1644
01:13:55,390 --> 01:14:00,320
Maybe this becomes m-n-o--
this time it becomes p.

1645
01:14:00,320 --> 01:14:01,500
And then so forth.

1646
01:14:01,500 --> 01:14:05,840
>> We add incrementing values
to each of the letters, which

1647
01:14:05,840 --> 01:14:09,880
is harder because, now notice, l-l
does not look like m-m, anymore.

1648
01:14:09,880 --> 01:14:11,860
We now need to be a little fancier.

1649
01:14:11,860 --> 01:14:16,750
And this is what's called, after
a French guy, a Vigenere cipher,

1650
01:14:16,750 --> 01:14:19,550
where you're using disparate
keys, different values.

1651
01:14:19,550 --> 01:14:21,300
And in fact, let's tie
that back together.

1652
01:14:21,300 --> 01:14:24,040
>> We used the word "key" before,
both in the physical sense,

1653
01:14:24,040 --> 01:14:25,490
for hotels and homes.

1654
01:14:25,490 --> 01:14:30,290
But in the electronic sense, a key
is just a secret value, typically.

1655
01:14:30,290 --> 01:14:36,540
And a secret value in this
earlier case, i-f-m-m-p--

1656
01:14:36,540 --> 01:14:43,862
what is the secret key I'm using for
this cipher that Dan proposed earlier?

1657
01:14:43,862 --> 01:14:45,070
AUDIENCE: Plus 1 [INAUDIBLE].

1658
01:14:45,070 --> 01:14:45,820
DAVID MALAN: Yeah.

1659
01:14:45,820 --> 01:14:49,930
The key is just the number 1-- not
the most secure, but it's simple.

1660
01:14:49,930 --> 01:14:53,560
But all of these security mechanisms
require what-- that not only I

1661
01:14:53,560 --> 01:14:57,660
know the secret is 1, but also what?

1662
01:14:57,660 --> 01:14:58,860
Who else has to know it?

1663
01:14:58,860 --> 01:15:00,360
AUDIENCE: The recipient [INAUDIBLE].

1664
01:15:00,360 --> 01:15:02,110
DAVID MALAN: The
recipient has to know it.

1665
01:15:02,110 --> 01:15:04,384
And just for clarity,
who must not know it?

1666
01:15:04,384 --> 01:15:05,300
AUDIENCE: The teacher.

1667
01:15:05,300 --> 01:15:06,675
DAVID MALAN: The teacher-- right?

1668
01:15:06,675 --> 01:15:10,360
Unless he or she has the time and energy
to brute force it or figure it out.

1669
01:15:10,360 --> 01:15:13,930
So super simple idea, but it
maps to what you're reading about

1670
01:15:13,930 --> 01:15:16,060
and hearing about every day in the news.

1671
01:15:16,060 --> 01:15:19,240
But the 256-- this is essentially 1 bit.

1672
01:15:19,240 --> 01:15:21,330
256 bits is much bigger.

1673
01:15:21,330 --> 01:15:26,080
And again, we'll get a quantitative
sense of that tomorrow.

1674
01:15:26,080 --> 01:15:32,390
Any questions then on Apple, security,
encryption, in these building blocks?

1675
01:15:32,390 --> 01:15:34,390
Yeah, Roman.

1676
01:15:34,390 --> 01:15:35,849
>> AUDIENCE: [INAUDIBLE].

1677
01:15:35,849 --> 01:15:37,390
Do you have any insights [INAUDIBLE]?

1678
01:15:37,390 --> 01:15:40,050

1679
01:15:40,050 --> 01:15:41,550
DAVID MALAN: Oh, it's good question.

1680
01:15:41,550 --> 01:15:44,750
I don't know internally--
and Apple, of all companies

1681
01:15:44,750 --> 01:15:49,340
is particularly quiet when it comes to
those kinds of implementation details.

1682
01:15:49,340 --> 01:15:53,410
But I can say more generally, a
fundamental tenet of security,

1683
01:15:53,410 --> 01:15:55,910
at least in the academic
community, is that you should never

1684
01:15:55,910 --> 01:15:58,420
have what's called
security through obscurity.

1685
01:15:58,420 --> 01:16:03,510
You should never do something to
protect data, or users, or information,

1686
01:16:03,510 --> 01:16:07,490
whose security and
privacy is all grounded

1687
01:16:07,490 --> 01:16:11,520
on no one knowing how it works.

1688
01:16:11,520 --> 01:16:13,590
>> In other words, what
the article alludes to,

1689
01:16:13,590 --> 01:16:17,864
AES, Advanced Encryption Standard-- that
is actually a global, public, standard

1690
01:16:17,864 --> 01:16:20,530
that you can open up a math book
or go on Wikipedia and actually

1691
01:16:20,530 --> 01:16:22,300
read what the algorithm is.

1692
01:16:22,300 --> 01:16:25,570
And much like the algorithm
here is the super simple plus 1,

1693
01:16:25,570 --> 01:16:29,010
this is more complicated mathematics,
but it's public knowledge.

1694
01:16:29,010 --> 01:16:30,510
And this has a number of upsides.

1695
01:16:30,510 --> 01:16:33,020
One, it means anyone can
use it and implement it.

1696
01:16:33,020 --> 01:16:35,320
But two, it also means that
millions of smart people

1697
01:16:35,320 --> 01:16:40,460
can review it and make sure to
speak up if it's flawed in some way.

1698
01:16:40,460 --> 01:16:44,190
>> And so in fact, one of the best defenses
against governmental back doors,

1699
01:16:44,190 --> 01:16:47,750
whether in this country or any
other, is to just publicly discuss

1700
01:16:47,750 --> 01:16:50,700
these kinds of algorithms
because it's very unlikely

1701
01:16:50,700 --> 01:16:53,380
that the entire world
of academic researchers

1702
01:16:53,380 --> 01:16:58,120
is going to collude and actually
tolerate hidden back doors

1703
01:16:58,120 --> 01:16:59,840
in algorithms like that.

1704
01:16:59,840 --> 01:17:01,940
>> However, you do need to
keep something secret.

1705
01:17:01,940 --> 01:17:06,270
And so just to be clear, when
you're using a cipher, like AES,

1706
01:17:06,270 --> 01:17:10,040
or something like Caesar, or Vigenere
that we alluded to there, what does

1707
01:17:10,040 --> 01:17:11,038
have to be kept secret?

1708
01:17:11,038 --> 01:17:15,030

1709
01:17:15,030 --> 01:17:16,969
Not the algorithm, not the process.

1710
01:17:16,969 --> 01:17:17,760
AUDIENCE: The code.

1711
01:17:17,760 --> 01:17:20,890
DAVID MALAN: The code, right--
and the key, to be clear.

1712
01:17:20,890 --> 01:17:27,190
And so to be super clear, even though
this is a trivial example, the cipher,

1713
01:17:27,190 --> 01:17:30,130
or algorithm, that we've generally
been using in this discussion

1714
01:17:30,130 --> 01:17:31,630
is this thing here, the plus.

1715
01:17:31,630 --> 01:17:37,190
So addition is our super
simple cipher or algorithm.

1716
01:17:37,190 --> 01:17:41,367
AES would be a much more
complex equivalent of the plus.

1717
01:17:41,367 --> 01:17:43,950
You do a lot more math, a lot
more additions, multiplications,

1718
01:17:43,950 --> 01:17:45,530
and so forth.

1719
01:17:45,530 --> 01:17:48,620
>> But the key is not the
same as the cipher.

1720
01:17:48,620 --> 01:17:51,400
In this case, it's also super
simple-- just the number 1.

1721
01:17:51,400 --> 01:17:58,490
In Apple's case, it's some
256-bit pattern of 0s and 1s.

1722
01:17:58,490 --> 01:18:01,460
So I'm not really answering your
own question because I can't really

1723
01:18:01,460 --> 01:18:06,690
speak to what Apple knows,
but the Apple engineers

1724
01:18:06,690 --> 01:18:10,000
have disclosed that they implement
this algorithm to some extent.

1725
01:18:10,000 --> 01:18:11,836
We have to trust that
they're being true,

1726
01:18:11,836 --> 01:18:14,210
and we have to trust that they
didn't, nonetheless, build

1727
01:18:14,210 --> 01:18:16,830
in some secret backdoor for the NSA.

1728
01:18:16,830 --> 01:18:18,330
And that's fundamentally hard to do.

1729
01:18:18,330 --> 01:18:21,040
>> In fact, the frightening
thought I can leave you

1730
01:18:21,040 --> 01:18:24,615
with on this particular subject is,
much as we might all talk about this

1731
01:18:24,615 --> 01:18:27,490
and much as Tim Cook might assure
us that these phones do not already

1732
01:18:27,490 --> 01:18:33,030
do what the FBI wants them to do, it's
nearly impossible to verify or audit

1733
01:18:33,030 --> 01:18:34,320
as much.

1734
01:18:34,320 --> 01:18:37,140
Do we even know that my
camera's not on right now?

1735
01:18:37,140 --> 01:18:40,580
Do you know that your own Macbook's
camera's not on right now?

1736
01:18:40,580 --> 01:18:43,460
Well, most of you might know
intuitively or from experience,

1737
01:18:43,460 --> 01:18:46,634
well, if the green light's
off, what does that mean?

1738
01:18:46,634 --> 01:18:47,550
AUDIENCE: It's not on.

1739
01:18:47,550 --> 01:18:48,020
DAVID MALAN: It's not on.

1740
01:18:48,020 --> 01:18:48,520
OK.

1741
01:18:48,520 --> 01:18:51,650
You've been taught that,
but why couldn't you

1742
01:18:51,650 --> 01:18:56,236
write software that turns off the
light but turns on the camera?

1743
01:18:56,236 --> 01:18:59,810

1744
01:18:59,810 --> 01:19:03,100
There's really no fundamental
defense against something like that.

1745
01:19:03,100 --> 01:19:05,510
>> So even we humans can
be socially engineered

1746
01:19:05,510 --> 01:19:09,530
by our computers to trust
one truth-- one reality,

1747
01:19:09,530 --> 01:19:11,340
when really we can
then be taken advantage

1748
01:19:11,340 --> 01:19:14,904
of because of that exact same
assumption that a green light means

1749
01:19:14,904 --> 01:19:15,570
the camera's on.

1750
01:19:15,570 --> 01:19:16,980
That's not necessarily true.

1751
01:19:16,980 --> 01:19:17,340
>> AUDIENCE: [INAUDIBLE]

1752
01:19:17,340 --> 01:19:18,090
>> DAVID MALAN: Yeah.

1753
01:19:18,090 --> 01:19:20,600
So actually, I always kind
of smile, but I appreciate it

1754
01:19:20,600 --> 01:19:24,490
when you see real diehards around
campus-- you have tape on yours.

1755
01:19:24,490 --> 01:19:25,980
OK.

1756
01:19:25,980 --> 01:19:28,622
So putting tape on it is
a more surefire thing.

1757
01:19:28,622 --> 01:19:31,080
Of course, there's still a
microphone that you can't really

1758
01:19:31,080 --> 01:19:32,707
tape over as effectively.

1759
01:19:32,707 --> 01:19:34,290
But these are the kinds of trade-offs.

1760
01:19:34,290 --> 01:19:36,400
>> And in fact, one take
away hopefully for today

1761
01:19:36,400 --> 01:19:39,306
should be absolute terror when
it comes to these kinds of things

1762
01:19:39,306 --> 01:19:41,680
because, at the end of the
day, we have to trust someone.

1763
01:19:41,680 --> 01:19:44,417
And that too is a fundamental
tenet of security.

1764
01:19:44,417 --> 01:19:46,000
Ultimately, you have to trust someone.

1765
01:19:46,000 --> 01:19:48,208
You have to trust that the
person you have a crush on

1766
01:19:48,208 --> 01:19:51,280
is not going to tell his or her best
friend what that secret code is,

1767
01:19:51,280 --> 01:19:54,820
and then disclose that same information
that you're trying to keep secret.

1768
01:19:54,820 --> 01:19:55,320
All right.

1769
01:19:55,320 --> 01:19:58,640
Let's take a look-- yeah, Dan.

1770
01:19:58,640 --> 01:20:03,101
>> AUDIENCE: What does the acronym
CBC stand for under the latest AES?

1771
01:20:03,101 --> 01:20:10,422
>> DAVID MALAN: Oh, CBC is block-- what's
it stand for-- block [INAUDIBLE] CBC.

1772
01:20:10,422 --> 01:20:15,090

1773
01:20:15,090 --> 01:20:16,340
Cypher Block Chaining.

1774
01:20:16,340 --> 01:20:20,230
So Cypher Block Chaining is
an acronym that refers to,

1775
01:20:20,230 --> 01:20:23,120
I believe, the process
of what goes on inside

1776
01:20:23,120 --> 01:20:26,980
of an algorithm for cryptography, in
this case, whereby it's iterative.

1777
01:20:26,980 --> 01:20:29,240
You do something again,
and again, and again.

1778
01:20:29,240 --> 01:20:31,740
And you take a previous
output, and feed it

1779
01:20:31,740 --> 01:20:34,660
into your process as a subsequent input.

1780
01:20:34,660 --> 01:20:38,410
So you keep feeding the
results back into themselves.

1781
01:20:38,410 --> 01:20:44,850
>> And an analog here might be-- I should
be able to come up with a good metaphor

1782
01:20:44,850 --> 01:20:45,410
here.

1783
01:20:45,410 --> 01:20:48,060
Let me try to think of
a better example here.

1784
01:20:48,060 --> 01:20:51,484
Let's see if we can come
up with a quick picture.

1785
01:20:51,484 --> 01:20:53,400
Let's see if Wikipedia
gives us a picture that

1786
01:20:53,400 --> 01:20:55,370
would explain-- no, that's good.

1787
01:20:55,370 --> 01:20:58,610

1788
01:20:58,610 --> 01:21:02,840
This is a more pictorial
complexity than we really want.

1789
01:21:02,840 --> 01:21:08,770
But the idea here is that if you are
enciphering something, it gets fed in,

1790
01:21:08,770 --> 01:21:12,070
then the output gets fed in
again, then it gets fed in again,

1791
01:21:12,070 --> 01:21:15,870
so that your iteratively
scrambling information using

1792
01:21:15,870 --> 01:21:20,047
previous output as a subsequent input.

1793
01:21:20,047 --> 01:21:22,130
Let me see if I come up
with a better explanation.

1794
01:21:22,130 --> 01:21:25,600
Give me lunch time to
noodle on that one.

1795
01:21:25,600 --> 01:21:26,380
>> All right.

1796
01:21:26,380 --> 01:21:28,640
Let's come back here.

1797
01:21:28,640 --> 01:21:32,537
I want to encourage you-- your only
homework for tonight, if you'd like,

1798
01:21:32,537 --> 01:21:35,120
and you haven't seen it, is to
watch a 20 minute video, if you

1799
01:21:35,120 --> 01:21:36,850
have internet access and go on YouTube.

1800
01:21:36,850 --> 01:21:40,840
Last Week Tonight is a brilliant show
by John Oliver from The Daily Show.

1801
01:21:40,840 --> 01:21:45,010
>> And at this URL here, you can actually
look at his look-- his humorous,

1802
01:21:45,010 --> 01:21:48,020
but simultaneously serious
look at the same issue.

1803
01:21:48,020 --> 01:21:51,629
And hopefully, even more of
that video will make sense.

1804
01:21:51,629 --> 01:21:52,920
And this is in the slides, too.

1805
01:21:52,920 --> 01:21:56,070
So if you have the URL up with
the slides, this is here, too.

1806
01:21:56,070 --> 01:21:59,620
And we'll get you online
during the break as well.

1807
01:21:59,620 --> 01:22:02,420
>> So in our concluding minutes,
let's take a quick look

1808
01:22:02,420 --> 01:22:06,530
at one other example of a technology
that's ever present these days,

1809
01:22:06,530 --> 01:22:13,700
file sharing, both in consumer
and in corporate contexts.

1810
01:22:13,700 --> 01:22:22,370
And that is by way of, for our
purposes, something called Dropbox.

1811
01:22:22,370 --> 01:22:26,000
So for those unfamiliar, in a
sentence or two, what problem

1812
01:22:26,000 --> 01:22:28,388
does Dropbox solve?

1813
01:22:28,388 --> 01:22:32,390
>> AUDIENCE: [INAUDIBLE] and then get
it on your iPhone or iPad anywhere.

1814
01:22:32,390 --> 01:22:33,290
>> DAVID MALAN: Yeah.

1815
01:22:33,290 --> 01:22:33,790
Exactly.

1816
01:22:33,790 --> 01:22:36,290
It allows you to share
files often with yourself

1817
01:22:36,290 --> 01:22:39,290
so that if you do have an iPhone,
an Android phone, a Mac, a PC,

1818
01:22:39,290 --> 01:22:42,220
multiple Macs, multiple PCs,
home computers, work computers,

1819
01:22:42,220 --> 01:22:46,480
you can have a folder that in turn has
its own sub folders that automatically

1820
01:22:46,480 --> 01:22:48,324
get synchronized across
all your devices.

1821
01:22:48,324 --> 01:22:49,490
And it's wonderfully useful.

1822
01:22:49,490 --> 01:22:52,350
>> For instance, in the morning,
if I'm preparing for class,

1823
01:22:52,350 --> 01:22:54,560
I might get my slides, or
videos, or pictures ready,

1824
01:22:54,560 --> 01:22:57,760
drop them in a folder on a home
computer, then walk to school,

1825
01:22:57,760 --> 01:23:01,187
and open up a work computer here,
and voila, it's magically there--

1826
01:23:01,187 --> 01:23:03,270
unless I screwed up, which
has happened sometimes,

1827
01:23:03,270 --> 01:23:05,830
and there's nothing more stressful than
having done all that work hours prior

1828
01:23:05,830 --> 01:23:08,610
and you have nothing to show for
it when it comes time for class.

1829
01:23:08,610 --> 01:23:10,640
So it fails sometimes,
or the human fails,

1830
01:23:10,640 --> 01:23:13,210
but in theory that's exactly
what it's supposed to do.

1831
01:23:13,210 --> 01:23:16,780
>> More compellingly, for other
users, is that I can very often

1832
01:23:16,780 --> 01:23:19,400
then Control click or
right click a folder

1833
01:23:19,400 --> 01:23:22,080
or file that I'm using
with this service,

1834
01:23:22,080 --> 01:23:26,080
and I can send a URL that results
from that click to a friend,

1835
01:23:26,080 --> 01:23:28,270
and he or she can then
download a copy of that file.

1836
01:23:28,270 --> 01:23:31,630
Or better yet, we can share
folders so that if I make a change,

1837
01:23:31,630 --> 01:23:35,080
then Victoria can see my changes in
her folder, and Kareem later in the day

1838
01:23:35,080 --> 01:23:39,120
can edit it and see that
same file and folder as well.

1839
01:23:39,120 --> 01:23:40,741
So there's a lot of implications here.

1840
01:23:40,741 --> 01:23:42,740
And we'll just scratch
the surface, and try here

1841
01:23:42,740 --> 01:23:46,610
to spook you a bit into not taking
for granted how all of this works

1842
01:23:46,610 --> 01:23:50,210
and what the actual implications
are for things that you're using.

1843
01:23:50,210 --> 01:23:55,300
>> In particular, let's consider
how Dropbox must surely work.

1844
01:23:55,300 --> 01:24:00,510
So if I'm over here-- let's
draw a quick picture of me.

1845
01:24:00,510 --> 01:24:10,760
If this is little old me-- this is
little old me on my laptop here.

1846
01:24:10,760 --> 01:24:19,330
And let's say this is Victoria
with her tape on her camera.

1847
01:24:19,330 --> 01:24:24,080
And here we have Kareem,
with his laptop here.

1848
01:24:24,080 --> 01:24:27,660
And then somewhere is this thing called
the cloud, more on that this afternoon

1849
01:24:27,660 --> 01:24:28,660
as well.

1850
01:24:28,660 --> 01:24:30,880
>> So how does Dropbox work?

1851
01:24:30,880 --> 01:24:32,629
Suppose I create a
folder on my computer,

1852
01:24:32,629 --> 01:24:34,420
and I install this
software called Dropbox.

1853
01:24:34,420 --> 01:24:37,500
But we could also be talking
about OneDrive from Microsoft,

1854
01:24:37,500 --> 01:24:41,160
or we could talk about the Google
Drive, or any number of other products.

1855
01:24:41,160 --> 01:24:43,270
It's all fundamentally the same.

1856
01:24:43,270 --> 01:24:45,780
>> If I've got a folder called
Dropbox on this computer,

1857
01:24:45,780 --> 01:24:48,790
and I've just created a PowerPoint
presentation, or an Excel file,

1858
01:24:48,790 --> 01:24:52,220
or an essay, and I drag
it into that folder, what

1859
01:24:52,220 --> 01:24:56,630
must happen in order to get it to
Victoria's computer or Kareem's

1860
01:24:56,630 --> 01:24:57,210
computer?

1861
01:24:57,210 --> 01:24:58,085
>> AUDIENCE: [INAUDIBLE]

1862
01:24:58,085 --> 01:25:00,430

1863
01:25:00,430 --> 01:25:01,180
DAVID MALAN: Yeah.

1864
01:25:01,180 --> 01:25:05,100
So somewhere in here, there's a company.

1865
01:25:05,100 --> 01:25:08,140
And we'll call this Dropbox.

1866
01:25:08,140 --> 01:25:09,530
And this is David.

1867
01:25:09,530 --> 01:25:10,710
This is Victoria.

1868
01:25:10,710 --> 01:25:12,350
And this is Kareem here.

1869
01:25:12,350 --> 01:25:17,220
>> So somehow, I must have an internet
connection that leads to the internet--

1870
01:25:17,220 --> 01:25:19,680
more on that after our
break-- that gets stored

1871
01:25:19,680 --> 01:25:23,590
on servers in Dropbox's headquarters,
or data center, wherever it is.

1872
01:25:23,590 --> 01:25:27,890
And then Victoria's computer and
Kareem's computer get that data how?

1873
01:25:27,890 --> 01:25:29,639
>> AUDIENCE: [INAUDIBLE]

1874
01:25:29,639 --> 01:25:30,930
DAVID MALAN: Have to say again?

1875
01:25:30,930 --> 01:25:31,650
AUDIENCE: [INAUDIBLE]

1876
01:25:31,650 --> 01:25:31,890
DAVID MALAN: Yeah.

1877
01:25:31,890 --> 01:25:33,098
I have to share it with them.

1878
01:25:33,098 --> 01:25:35,310
So I have to have sent
Kareem to Victoria a URL,

1879
01:25:35,310 --> 01:25:38,210
or I have to click some menu option
and type in their e-mail address

1880
01:25:38,210 --> 01:25:39,560
so it automatically gets shared.

1881
01:25:39,560 --> 01:25:40,768
Let's suppose I've done that.

1882
01:25:40,768 --> 01:25:43,247
What then happens in
terms of this picture?

1883
01:25:43,247 --> 01:25:47,570
>> AUDIENCE: You need a user account
and a way to authenticate--

1884
01:25:47,570 --> 01:25:48,320
DAVID MALAN: Yeah.

1885
01:25:48,320 --> 01:25:50,720
We're going to need a priori
some kind of user account.

1886
01:25:50,720 --> 01:25:52,220
So I've got to register for Dropbox.

1887
01:25:52,220 --> 01:25:55,910
Each of you probably has to register
for Dropbox, at least in this scenario.

1888
01:25:55,910 --> 01:26:00,790
But then ultimately, that file gets
transmitted down in this direction,

1889
01:26:00,790 --> 01:26:03,250
just as it went up from
my direction there.

1890
01:26:03,250 --> 01:26:05,800
>> Similarly, if we've used a
certain feature of Dropbox,

1891
01:26:05,800 --> 01:26:08,930
you can either make copies of files
or actually share the originals.

1892
01:26:08,930 --> 01:26:11,090
If you guys start to make
copies, then in theory

1893
01:26:11,090 --> 01:26:13,450
those should propagate back to me.

1894
01:26:13,450 --> 01:26:17,660
>> So if you're a
particularly paranoid user,

1895
01:26:17,660 --> 01:26:21,600
or you're the CTO or chief
security officer at a company,

1896
01:26:21,600 --> 01:26:25,020
what kinds of questions should you be
asking here about this whole process?

1897
01:26:25,020 --> 01:26:26,280
>> AUDIENCE: [INAUDIBLE]

1898
01:26:26,280 --> 01:26:26,620
>> DAVID MALAN: Yeah.

1899
01:26:26,620 --> 01:26:27,120
OK.

1900
01:26:27,120 --> 01:26:29,520
So I am now the head of Dropbox.

1901
01:26:29,520 --> 01:26:30,060
Yes.

1902
01:26:30,060 --> 01:26:34,960
We use industry standard
encryption to secure your data.

1903
01:26:34,960 --> 01:26:36,381
Satisfied?

1904
01:26:36,381 --> 01:26:36,880
Why not?

1905
01:26:36,880 --> 01:26:40,457

1906
01:26:40,457 --> 01:26:41,540
OK, I'll be more specific.

1907
01:26:41,540 --> 01:26:46,940
I use 256-bit AES encryption just
like our friends at Apple do.

1908
01:26:46,940 --> 01:26:49,940
>> AUDIENCE: But all that data
exists on all those machines,

1909
01:26:49,940 --> 01:26:52,654
and all those machines
are a vulnerability.

1910
01:26:52,654 --> 01:26:53,320
DAVID MALAN: OK.

1911
01:26:53,320 --> 01:26:53,819
True.

1912
01:26:53,819 --> 01:26:56,920
So suppose there's a whole bunch of
servers in Dropbox's headquarters,

1913
01:26:56,920 --> 01:27:02,760
or data center, or all of their
data centers, and it's the data--

1914
01:27:02,760 --> 01:27:06,200
and this is a feature-- is replicated--
copied-- across multiple servers

1915
01:27:06,200 --> 01:27:09,160
because, God forbid, one
computer, one hard drive dies.

1916
01:27:09,160 --> 01:27:11,830
These days very common
is to replicate data

1917
01:27:11,830 --> 01:27:14,530
across at least two computers
or two hard drives--

1918
01:27:14,530 --> 01:27:17,350
sometimes as many as
five or more so that,

1919
01:27:17,350 --> 01:27:20,140
statistically, even
though, yes, lightning

1920
01:27:20,140 --> 01:27:23,660
might strike all of Dropbox's
data centers simultaneously,

1921
01:27:23,660 --> 01:27:27,020
or they might get physically attacked,
or compromised all at the same time,

1922
01:27:27,020 --> 01:27:30,660
the probability of that happening
is very, very, very low.

1923
01:27:30,660 --> 01:27:34,730
So for all intents and purposes,
my data is backed up safely.

1924
01:27:34,730 --> 01:27:35,970
>> But it's encrypted.

1925
01:27:35,970 --> 01:27:36,480
So what?

1926
01:27:36,480 --> 01:27:39,390
Doesn't matter if every copy gets
stolen, doesn't matter if the data

1927
01:27:39,390 --> 01:27:42,760
center gets infiltrated,
my data is still encrypted

1928
01:27:42,760 --> 01:27:46,140
so no one can see what it is.

1929
01:27:46,140 --> 01:27:47,956
What questions should
you continue asking?

1930
01:27:47,956 --> 01:27:51,169

1931
01:27:51,169 --> 01:27:53,752
AUDIENCE: Is it all encrypted
the same way across [INAUDIBLE]?

1932
01:27:53,752 --> 01:27:56,650

1933
01:27:56,650 --> 01:27:58,750
>> DAVID MALAN: Embarrassingly, yes.

1934
01:27:58,750 --> 01:28:01,962
We use the same key to encrypt
all of our customer's data.

1935
01:28:01,962 --> 01:28:04,920
AUDIENCE: But then it's very easy to
unencrypt and decrypt [INAUDIBLE].

1936
01:28:04,920 --> 01:28:05,400
DAVID MALAN: It is.

1937
01:28:05,400 --> 01:28:06,274
And that's a feature.

1938
01:28:06,274 --> 01:28:09,800
We can do this super fast for you,
which is why the file syncs so quickly.

1939
01:28:09,800 --> 01:28:12,550
We use the same encryption--
the same key for everyone.

1940
01:28:12,550 --> 01:28:14,060
It's a feature.

1941
01:28:14,060 --> 01:28:16,320
>> And I said it sheepishly--
and this actually,

1942
01:28:16,320 --> 01:28:18,430
I believe, is still
actually technically true.

1943
01:28:18,430 --> 01:28:22,900
They do use the same secret key,
whether it's 256 bits or more,

1944
01:28:22,900 --> 01:28:24,530
for all customer's data.

1945
01:28:24,530 --> 01:28:26,820
And this is partly
for technical reasons.

1946
01:28:26,820 --> 01:28:31,989
One, if I am sharing a file
with Victoria and Kareem,

1947
01:28:31,989 --> 01:28:33,780
and they want to be
able to access it, I've

1948
01:28:33,780 --> 01:28:35,840
got to somehow decrypt it for them.

1949
01:28:35,840 --> 01:28:41,110
But I don't really have a mechanism to
give Victoria and Karim a secret key.

1950
01:28:41,110 --> 01:28:45,610
>> If I email it to them, I'm compromising
it because anyone on the internet

1951
01:28:45,610 --> 01:28:46,960
could intercept my email.

1952
01:28:46,960 --> 01:28:50,480
I certainly am not going to call
them with a sequence of 256 0s

1953
01:28:50,480 --> 01:28:53,150
and 1s or more, and
tell them to type it in.

1954
01:28:53,150 --> 01:28:55,964
>> It could just be a password,
but I'd still have to call them.

1955
01:28:55,964 --> 01:28:58,130
And in business, this isn't
going to work very well.

1956
01:28:58,130 --> 01:28:59,570
If you want to share
a file with 30 people,

1957
01:28:59,570 --> 01:29:01,030
I'm not going to make
30 darn phone calls.

1958
01:29:01,030 --> 01:29:03,100
And I can't send out an email
because that's insecure.

1959
01:29:03,100 --> 01:29:05,475
>> So there's really this
fundamental problem of sharing it.

1960
01:29:05,475 --> 01:29:09,440
So you know what, it's just easier if
Dropbox does the encryption for us.

1961
01:29:09,440 --> 01:29:12,510
But if they do it for us,
only they know the key.

1962
01:29:12,510 --> 01:29:17,550
And if they reuse the key,
that means that all of the data

1963
01:29:17,550 --> 01:29:21,596
could be compromised if that
key itself is compromised.

1964
01:29:21,596 --> 01:29:24,220
Now, having asked at least one
buddy at Dropbox, they do have--

1965
01:29:24,220 --> 01:29:26,178
and I think they have
white papers that testify

1966
01:29:26,178 --> 01:29:30,180
to this fact-- they do have very, very
few people who have access to that key.

1967
01:29:30,180 --> 01:29:32,560
The computers have to have
it in memory, and it's

1968
01:29:32,560 --> 01:29:34,960
got to be locked up in
some vault somewhere so

1969
01:29:34,960 --> 01:29:38,090
that, God forbid, the computers
crash or need to be rebooted,

1970
01:29:38,090 --> 01:29:40,990
someone does have to type
in that key at some point.

1971
01:29:40,990 --> 01:29:43,820
>> So that is really the secret
sauce if there were any.

1972
01:29:43,820 --> 01:29:47,670
But this definitely has
implications for my data.

1973
01:29:47,670 --> 01:29:52,020
It's disclosable, if someone compromises
that key or that data center.

1974
01:29:52,020 --> 01:29:55,150
>> But it also allows
Dropbox another feature.

1975
01:29:55,150 --> 01:29:57,620
It turns out-- and this is
kind of a business cost--

1976
01:29:57,620 --> 01:30:01,140
if you used a different
key for every customer,

1977
01:30:01,140 --> 01:30:04,980
or even more so for every
file, mathematically,

1978
01:30:04,980 --> 01:30:09,270
every file, when encrypted, would
look different from every other file.

1979
01:30:09,270 --> 01:30:12,450
>> So even if I had two copies of
the same PowerPoint presentation

1980
01:30:12,450 --> 01:30:16,280
on Kareem's computer and on my
computer, if those files were encrypted

1981
01:30:16,280 --> 01:30:20,030
with different keys, the
ciphertext-- the scrambled thing--

1982
01:30:20,030 --> 01:30:21,610
would look different.

1983
01:30:21,610 --> 01:30:24,100
This is not a good
thing because it doesn't

1984
01:30:24,100 --> 01:30:27,510
let Dropbox realize that those
files are the same, as we've

1985
01:30:27,510 --> 01:30:28,920
kind of discussed earlier.

1986
01:30:28,920 --> 01:30:33,940
Why might Dropbox want to know
when two users or more are

1987
01:30:33,940 --> 01:30:35,640
sharing the exact same file?

1988
01:30:35,640 --> 01:30:39,140
Why is that useful information for
Dropbox from a business perspective?

1989
01:30:39,140 --> 01:30:39,860
>> AUDIENCE: Space.

1990
01:30:39,860 --> 01:30:40,651
>> DAVID MALAN: Space.

1991
01:30:40,651 --> 01:30:42,400
A PowerPoint presentation's
not that big,

1992
01:30:42,400 --> 01:30:45,760
but people commonly share big
movie files, video files--

1993
01:30:45,760 --> 01:30:48,120
maybe really big
PowerPoint presentations.

1994
01:30:48,120 --> 01:30:51,410
And if you have two users with
the same file, or 10 users,

1995
01:30:51,410 --> 01:30:53,970
or maybe a million users
with the same popular

1996
01:30:53,970 --> 01:30:56,340
illegally downloaded
movie file, it's kind of

1997
01:30:56,340 --> 01:31:01,820
wasteful to store a million copies
of the same gigabytes of information,

1998
01:31:01,820 --> 01:31:05,160
the same gigabyte sized video, and
so Dropbox, like a lot of companies,

1999
01:31:05,160 --> 01:31:08,940
have a feature called "deduplication--
deduplication, which is just

2000
01:31:08,940 --> 01:31:14,040
a fancy way of saying store one
copy of the same file, not multiple,

2001
01:31:14,040 --> 01:31:17,290
and just keep track of the fact
that a million people, or whatever,

2002
01:31:17,290 --> 01:31:18,890
have that same file.

2003
01:31:18,890 --> 01:31:22,710
>> So just point all million
people or so to that same file.

2004
01:31:22,710 --> 01:31:24,330
And you still back it up a few times.

2005
01:31:24,330 --> 01:31:26,690
So this is separate from
the issue of redundancy

2006
01:31:26,690 --> 01:31:28,980
in case you have hardware
failures or the like.

2007
01:31:28,980 --> 01:31:33,920
But deduplication requires that
you not encrypt files individually

2008
01:31:33,920 --> 01:31:36,260
if you want to be able to
determine after the fact

2009
01:31:36,260 --> 01:31:38,320
if they're still in fact the same.

2010
01:31:38,320 --> 01:31:39,970
>> So there's some trade-offs here.

2011
01:31:39,970 --> 01:31:43,370
And it's not necessarily
clear what the right call is.

2012
01:31:43,370 --> 01:31:46,500
Personally with Dropbox, I'll use
it for anything related to work,

2013
01:31:46,500 --> 01:31:49,500
certainly anything related to class,
certainly for any files that I know

2014
01:31:49,500 --> 01:31:51,984
are going to end up on the
internet anyway by choice.

2015
01:31:51,984 --> 01:31:53,900
But I don't really use
it for financial stuff,

2016
01:31:53,900 --> 01:31:56,500
nothing particularly private
or family related because,

2017
01:31:56,500 --> 01:31:59,710
as a matter of principle, not
super comfortable with the fact

2018
01:31:59,710 --> 01:32:02,170
that it might be
encrypted on my own Mac,

2019
01:32:02,170 --> 01:32:05,590
but as soon as it goes out of the cloud,
it's on little old Dropbox's Servers.

2020
01:32:05,590 --> 01:32:08,740
And I'm pretty sure no one
at Dropbox has it out for me

2021
01:32:08,740 --> 01:32:10,490
and is going to go
poking around my files,

2022
01:32:10,490 --> 01:32:15,450
but they absolutely could in theory,
no matter what policies or defense

2023
01:32:15,450 --> 01:32:16,710
mechanisms they put in place.

2024
01:32:16,710 --> 01:32:19,980
It just must be
technologically possible.

2025
01:32:19,980 --> 01:32:23,120
>> And God forbid they are
compromised, I'd rather my file not

2026
01:32:23,120 --> 01:32:27,360
end up in some big zip that some hacker
puts online for the whole world to see.

2027
01:32:27,360 --> 01:32:28,640
So let's push back on that.

2028
01:32:28,640 --> 01:32:29,700
What's a solution then?

2029
01:32:29,700 --> 01:32:33,850
Could you continue using a
service like Dropbox comfortably

2030
01:32:33,850 --> 01:32:35,580
and assuage my kinds of concerns?

2031
01:32:35,580 --> 01:32:36,580
>> AUDIENCE: Private cloud.

2032
01:32:36,580 --> 01:32:37,704
DAVID MALAN: Private cloud.

2033
01:32:37,704 --> 01:32:38,816
What does that mean?

2034
01:32:38,816 --> 01:32:41,524
>> AUDIENCE: Well, you secure it
somehow so that it's available only

2035
01:32:41,524 --> 01:32:42,680
for a particular group.

2036
01:32:42,680 --> 01:32:43,430
>> DAVID MALAN: Yeah.

2037
01:32:43,430 --> 01:32:47,080
So you need to partition the cloud
into something a little more narrowly

2038
01:32:47,080 --> 01:32:47,580
defined.

2039
01:32:47,580 --> 01:32:48,496
And we'll talk about--

2040
01:32:48,496 --> 01:32:49,400
AUDIENCE: Internet.

2041
01:32:49,400 --> 01:32:50,441
>> DAVID MALAN: An internet.

2042
01:32:50,441 --> 01:32:55,330
So I could just backup locally to my own
home, backup server, or cloud server,

2043
01:32:55,330 --> 01:32:56,007
so to speak.

2044
01:32:56,007 --> 01:32:58,090
Unfortunately, that means
that Victoria and Kareem

2045
01:32:58,090 --> 01:33:00,960
need to visit more often if I
want to share files with them but.

2046
01:33:00,960 --> 01:33:02,070
That might be one way.

2047
01:33:02,070 --> 01:33:04,150
>> There are also third
party software that I

2048
01:33:04,150 --> 01:33:09,040
could use on my Mac or my PC that
encrypts the contents of a folder,

2049
01:33:09,040 --> 01:33:12,520
but then I do have to call
Victoria or Karim, or email them,

2050
01:33:12,520 --> 01:33:15,070
or something to tell them that secret.

2051
01:33:15,070 --> 01:33:20,200
And that's a bit of a white lie because
there are types of cryptography that

2052
01:33:20,200 --> 01:33:22,470
do allow me and Kareem,
and me and Victoria,

2053
01:33:22,470 --> 01:33:26,180
to exchange secret messages without
having to, in advance, share

2054
01:33:26,180 --> 01:33:28,090
a private key-- a secret
key with each other.

2055
01:33:28,090 --> 01:33:30,450
It's actually something called
public key cryptography.

2056
01:33:30,450 --> 01:33:33,630
>> And we won't go into technical
detail, but whereas we today

2057
01:33:33,630 --> 01:33:36,090
have been talking about
secret key cryptography,

2058
01:33:36,090 --> 01:33:39,000
where both sender and recipient
have to know the same secret,

2059
01:33:39,000 --> 01:33:41,470
there's something called
public key cryptography, which

2060
01:33:41,470 --> 01:33:44,500
has a public key and a private
key, which long story short have

2061
01:33:44,500 --> 01:33:48,245
a fancy mathematical relationship
whereby if I want to send Victoria

2062
01:33:48,245 --> 01:33:52,310
a secret message, I ask her for her
public key, which by definition she

2063
01:33:52,310 --> 01:33:53,300
can email me.

2064
01:33:53,300 --> 01:33:54,680
She can post it on her website.

2065
01:33:54,680 --> 01:33:56,950
>> It is meant mathematically to be public.

2066
01:33:56,950 --> 01:34:00,580
But it has a relationship with
another really big number called

2067
01:34:00,580 --> 01:34:05,070
the private key such that when I
encrypt my message to her, "hello,"

2068
01:34:05,070 --> 01:34:07,104
with her public key,
you can perhaps guess

2069
01:34:07,104 --> 01:34:09,270
what's the only key
mathematically in the world that

2070
01:34:09,270 --> 01:34:16,500
can decrypt my message-- her private
key or corresponding private key.

2071
01:34:16,500 --> 01:34:19,040
>> It's fancier math than we've
been to talking about here.

2072
01:34:19,040 --> 01:34:21,710
It's not just addition
certainly, but that too exists.

2073
01:34:21,710 --> 01:34:25,240
And in fact, and we'll come back
to this when we talk about the web,

2074
01:34:25,240 --> 01:34:27,897
odds are you've never
called someone at amazon.com

2075
01:34:27,897 --> 01:34:30,980
when you want to check out with your
shopping cart and type in your credit

2076
01:34:30,980 --> 01:34:34,260
card number, and yet somehow or other
that padlock symbol is telling you

2077
01:34:34,260 --> 01:34:35,560
your connection is secure.

2078
01:34:35,560 --> 01:34:37,730
Somehow or other your
little old Mac or PC

2079
01:34:37,730 --> 01:34:39,900
does have an encrypted
connection to Amazon

2080
01:34:39,900 --> 01:34:43,010
even though you've never
arranged with them for a secret.

2081
01:34:43,010 --> 01:34:47,020
And that's because the web is
using public key cryptography.

2082
01:34:47,020 --> 01:34:50,327
Why don't we pause here, take our 15
minute break after Olivier's question.

2083
01:34:50,327 --> 01:34:51,910
AUDIENCE: I just have a dumb question.

2084
01:34:51,910 --> 01:34:53,076
DAVID MALAN: No, not at all.

2085
01:34:53,076 --> 01:34:57,320
AUDIENCE: If you have the original file,
and the key's the same for Dropbox,

2086
01:34:57,320 --> 01:34:59,244
for everyone, and you
have the encrypted file.

2087
01:34:59,244 --> 01:35:02,317
Can you [INAUDIBLE] the key?

2088
01:35:02,317 --> 01:35:03,650
DAVID MALAN: Say that once more.

2089
01:35:03,650 --> 01:35:06,970
AUDIENCE: If you have the original
file and the encrypted file,

2090
01:35:06,970 --> 01:35:11,274
and you have both of them,
can't you just [INAUDIBLE]?

2091
01:35:11,274 --> 01:35:11,940
DAVID MALAN: Oh.

2092
01:35:11,940 --> 01:35:13,850
A good question.

2093
01:35:13,850 --> 01:35:16,850
If you have the plaintext
and the ciphertext,

2094
01:35:16,850 --> 01:35:19,950
can you infer the secret key?

2095
01:35:19,950 --> 01:35:21,640
Depends on the cipher.

2096
01:35:21,640 --> 01:35:23,110
Sometimes yes, sometimes no.

2097
01:35:23,110 --> 01:35:28,840
It depends on how complex
the actual algorithm is.

2098
01:35:28,840 --> 01:35:30,820
>> But that does not help your situation.

2099
01:35:30,820 --> 01:35:33,220
It is a fundamental
tenet to, if you have

2100
01:35:33,220 --> 01:35:35,750
access to the original file
and the resulting file,

2101
01:35:35,750 --> 01:35:38,440
you should no longer use
that key because now you

2102
01:35:38,440 --> 01:35:39,700
have leaked information.

2103
01:35:39,700 --> 01:35:41,700
And an adversary could
use that and exploit

2104
01:35:41,700 --> 01:35:46,590
that to do what you're alluding to,
and reverse engineer what that key is.

2105
01:35:46,590 --> 01:35:50,260
>> But in this case, presumably when you're
sending something to the recipient,

2106
01:35:50,260 --> 01:35:52,590
you already have a trust
relationship with them.

2107
01:35:52,590 --> 01:35:56,420
And so by definition, they should
have or know that key already.

2108
01:35:56,420 --> 01:35:59,430
It's when someone in the
middle gets in the way.

2109
01:35:59,430 --> 01:36:00,480
Good question.

2110
01:36:00,480 --> 01:36:02,020
>> All right, Why don't we
pause, take a 15 minute break.

2111
01:36:02,020 --> 01:36:03,000
Rest rooms are that way.

2112
01:36:03,000 --> 01:36:05,375
I think there's probably some
drinks and snacks that way.

2113
01:36:05,375 --> 01:36:07,900
And we'll resume at 5
after 11, how about?

2114
01:36:07,900 --> 01:36:09,450
11:05.

2115
01:36:09,450 --> 01:36:11,736