1
00:00:00,000 --> 00:00:02,976


2
00:00:02,976 --> 00:00:06,448
[MUSIC PLAYING]

3
00:00:06,448 --> 00:01:12,307


4
00:01:12,307 --> 00:01:13,390
DAVID J. MALAN: All right.

5
00:01:13,390 --> 00:01:15,130
So this is CS50.

6
00:01:15,130 --> 00:01:18,520
My name is David Malan, and this is
Harvard University's introduction

7
00:01:18,520 --> 00:01:20,740
to the intellectual
enterprises of computer science

8
00:01:20,740 --> 00:01:22,450
and the art of programming.

9
00:01:22,450 --> 00:01:25,060
And this, of course, is
our special family weekend,

10
00:01:25,060 --> 00:01:28,180
wherein not only our CS50's own
students here in the audience, but also

11
00:01:28,180 --> 00:01:29,990
some family members as well.

12
00:01:29,990 --> 00:01:32,560
Now, you're showing up in the
semester a little bit late.

13
00:01:32,560 --> 00:01:36,130
We've just tackled week eight, which
is really our ninth week since computer

14
00:01:36,130 --> 00:01:37,630
scientists start counting from zero.

15
00:01:37,630 --> 00:01:40,047
So we've done a whole lot of
work over the past few weeks,

16
00:01:40,047 --> 00:01:42,890
as you might have heard via
emails or text messages home,

17
00:01:42,890 --> 00:01:45,858
including a language
known here as binary.

18
00:01:45,858 --> 00:01:48,400
So on the screen here, of course,
is a lot of zeros and ones.

19
00:01:48,400 --> 00:01:52,600
And suffice it to say, let me sum
up the past nine weeks with this

20
00:01:52,600 --> 00:01:54,340
is what's going on underneath the hood.

21
00:01:54,340 --> 00:01:57,465
But of course, today, we thought we'd
make things a little more accessible,

22
00:01:57,465 --> 00:01:58,870
a little more broadly applicable.

23
00:01:58,870 --> 00:02:01,480
And indeed, our focus
today will not be on what

24
00:02:01,480 --> 00:02:04,060
these patterns of zeros
and ones represent,

25
00:02:04,060 --> 00:02:07,390
which in astute, I might
notice are replicated visually

26
00:02:07,390 --> 00:02:10,190
with these light bulbs being
in a pattern on and off.

27
00:02:10,190 --> 00:02:14,140
And as your child might have hinted
before class, or perhaps, now,

28
00:02:14,140 --> 00:02:17,200
this might very well spell a
word up to eight characters

29
00:02:17,200 --> 00:02:20,890
long because you can encode, even in
the real world, things digital too.

30
00:02:20,890 --> 00:02:23,470
But today, we'll focus on
things much more high level,

31
00:02:23,470 --> 00:02:26,170
this notion of cybersecurity,
like the security

32
00:02:26,170 --> 00:02:31,210
of our data, our privacy of our
systems, particularly, on the internet

33
00:02:31,210 --> 00:02:34,330
nowadays because presumably, all of
us are carrying technologies around

34
00:02:34,330 --> 00:02:36,770
in our pocket using laptops
and desktops every day.

35
00:02:36,770 --> 00:02:40,685
And so the goal today is
to stipulate that this

36
00:02:40,685 --> 00:02:42,310
is what's going on underneath the hood.

37
00:02:42,310 --> 00:02:44,350
But let's solve some
problems at a higher level,

38
00:02:44,350 --> 00:02:47,770
so that your homework, when you go
back to wherever you're visiting from,

39
00:02:47,770 --> 00:02:50,740
can actually be to apply some
of today's lessons learned.

40
00:02:50,740 --> 00:02:55,780
So with that said, perhaps, the most
common familiar defense of one's

41
00:02:55,780 --> 00:02:58,720
systems, and data, phone, and
laptops, and desktops would just

42
00:02:58,720 --> 00:03:00,070
be these simple passwords.

43
00:03:00,070 --> 00:03:02,440
Unfortunately, you and I
are-- frankly, as humans,

44
00:03:02,440 --> 00:03:05,150
not all that good at choosing passwords.

45
00:03:05,150 --> 00:03:09,010
And this is in itself a relatively weak
form of defense, even though each of us

46
00:03:09,010 --> 00:03:12,430
has dozens, hundreds,
of passwords nowadays

47
00:03:12,430 --> 00:03:14,920
or at least dozens or
hundreds of accounts,

48
00:03:14,920 --> 00:03:19,653
maybe fives or tens of
dozens of passwords.

49
00:03:19,653 --> 00:03:21,820
Indeed, if you're in the
habit of reusing passwords,

50
00:03:21,820 --> 00:03:25,160
we'll see today, probably among
our first lessons learned.

51
00:03:25,160 --> 00:03:29,050
So for instance, if we look back
at the past year, 2021, thanks

52
00:03:29,050 --> 00:03:33,310
to security researchers who take a look
at data that has been hacked or leaked

53
00:03:33,310 --> 00:03:37,990
online by way of public databases, we
have a sense as computer scientists

54
00:03:37,990 --> 00:03:40,720
of what the most
popular, or equivalently,

55
00:03:40,720 --> 00:03:43,270
what some of the worst
passwords are that you and I are

56
00:03:43,270 --> 00:03:44,300
choosing for our system.

57
00:03:44,300 --> 00:03:46,750
So as of this past year,
according to one measure,

58
00:03:46,750 --> 00:03:53,470
the most commonly used password
in systems everywhere was 123456.

59
00:03:53,470 --> 00:03:54,220
All right?

60
00:03:54,220 --> 00:03:56,770
The number two password
in our top 10 here list

61
00:03:56,770 --> 00:04:00,670
was only slightly longer, 123456789.

62
00:04:00,670 --> 00:04:05,590
After that, we took a turn in
the other direction, 12345 alone.

63
00:04:05,590 --> 00:04:08,440
After that, it got a little
more interesting, qwerty,

64
00:04:08,440 --> 00:04:10,360
which might sound
pretty cryptic, but not

65
00:04:10,360 --> 00:04:12,100
if you look down at your US keyboard.

66
00:04:12,100 --> 00:04:16,750
And it's the top left hand row of the
keys on an American keyboard, so also,

67
00:04:16,750 --> 00:04:18,310
not all that hard.

68
00:04:18,310 --> 00:04:21,790
Perhaps, not surprisingly
a little disconcertingly,

69
00:04:21,790 --> 00:04:25,270
number five was password.

70
00:04:25,270 --> 00:04:29,200
Meanwhile, number six returns
us to digits, 12345678.

71
00:04:29,200 --> 00:04:33,310
After that, really, less effort, 111111.

72
00:04:33,310 --> 00:04:37,780
After that, a little more variation,
but not all that much 123123.

73
00:04:37,780 --> 00:04:41,830
After that, it's getting even
less interesting, 1234567890.

74
00:04:41,830 --> 00:04:45,520
And then lastly topping
the list is just 1234567.

75
00:04:45,520 --> 00:04:48,470
So this is not a good
top 10 list to be on.

76
00:04:48,470 --> 00:04:52,600
So among today's first takeaways is if
you see your password on the screen,

77
00:04:52,600 --> 00:04:54,490
you didn't make the list in a good way.

78
00:04:54,490 --> 00:04:57,550
This means hundreds, thousands,
millions of other people

79
00:04:57,550 --> 00:04:59,380
probably have that password of yours.

80
00:04:59,380 --> 00:05:02,590
Now, in and of itself, that's
not necessarily worrisome

81
00:05:02,590 --> 00:05:05,920
because I don't know who has these
passwords in a room as large as this.

82
00:05:05,920 --> 00:05:09,460
But just intuitively,
why is this a bad thing?

83
00:05:09,460 --> 00:05:14,120
Either parent or child is
welcome to raise a hand here.

84
00:05:14,120 --> 00:05:16,210
Why might this be--

85
00:05:16,210 --> 00:05:17,275
intuitively, yeah?

86
00:05:17,275 --> 00:05:18,707
AUDIENCE: Access to it.

87
00:05:18,707 --> 00:05:20,040
DAVID J. MALAN: So access to it.

88
00:05:20,040 --> 00:05:22,290
I mean, we literally, as
computer scientists, now have

89
00:05:22,290 --> 00:05:24,720
a database of really common passwords.

90
00:05:24,720 --> 00:05:25,680
And your thoughts?

91
00:05:25,680 --> 00:05:28,408
AUDIENCE: [INAUDIBLE]
find it out quickly.

92
00:05:28,408 --> 00:05:30,700
DAVID J. MALAN: Yeah, you
can just find it out quickly.

93
00:05:30,700 --> 00:05:33,910
I mean, you could imagine trying to
guess someone's password by just typing

94
00:05:33,910 --> 00:05:36,280
in random letters, random
numbers, random words,

95
00:05:36,280 --> 00:05:38,290
but not if you have a top 10 list.

96
00:05:38,290 --> 00:05:41,920
The adversaries in the world might
as well just start with this list.

97
00:05:41,920 --> 00:05:44,727
Now, you'll notice that even absent
from this are slight variance.

98
00:05:44,727 --> 00:05:46,810
Some of you might be
thinking, I'm not on the list

99
00:05:46,810 --> 00:05:50,470
because I do something clever like I
use an exclamation point for the number

100
00:05:50,470 --> 00:05:55,240
one, or a three for
an E, or a 5 for an S.

101
00:05:55,240 --> 00:05:57,160
And based on the smiles
in the room right now,

102
00:05:57,160 --> 00:06:00,460
you're not all that clever, it turns
out, because other people are smiling

103
00:06:00,460 --> 00:06:04,390
too, which is to say that an adversary
can take those same heuristics that you

104
00:06:04,390 --> 00:06:06,700
might think are making
things more secure by just

105
00:06:06,700 --> 00:06:08,752
tweaking some letters to
numbers or vise versa.

106
00:06:08,752 --> 00:06:10,960
But if you're doing it and
other people are doing it,

107
00:06:10,960 --> 00:06:14,150
the bad guys, so to speak, are
going to be doing it as well.

108
00:06:14,150 --> 00:06:16,570
So unfortunately, when
it comes to passwords,

109
00:06:16,570 --> 00:06:20,020
better is longer, and random,
and really unguessable.

110
00:06:20,020 --> 00:06:21,520
But that's not what most of us have.

111
00:06:21,520 --> 00:06:23,562
In fact, case in point on
our phones, whether you

112
00:06:23,562 --> 00:06:25,540
have an Android device
or an iPhone nowadays,

113
00:06:25,540 --> 00:06:28,900
odds are you have something
relatively simplistic protecting it,

114
00:06:28,900 --> 00:06:30,070
if you have anything at all.

115
00:06:30,070 --> 00:06:31,778
But at least, Apple
and Google are pretty

116
00:06:31,778 --> 00:06:34,630
good at at least nudging us to
choose these kinds of passcodes now.

117
00:06:34,630 --> 00:06:38,360
And a four-digit passcode
is quite common nowadays.

118
00:06:38,360 --> 00:06:42,190
And so here's where we have an
opportunity, thanks to the URL

119
00:06:42,190 --> 00:06:45,490
that you saw on the screen
earlier, to conjecture as a group

120
00:06:45,490 --> 00:06:49,060
just how long might it take an
adversary, someone out there

121
00:06:49,060 --> 00:06:51,460
who's out to get us or get one of us--

122
00:06:51,460 --> 00:06:54,820
how long might it take an
adversary to figure out

123
00:06:54,820 --> 00:06:57,580
your phone's four-digit passcode?

124
00:06:57,580 --> 00:06:58,930
This is CS50's own Carter.

125
00:06:58,930 --> 00:07:01,510
Carter, if you could switch over
and pull the audience here--

126
00:07:01,510 --> 00:07:04,552
if you take out your phone or laptop,
whatever device you might have used

127
00:07:04,552 --> 00:07:10,960
a few minutes ago, to scan that
QR code or to visit that same URL,

128
00:07:10,960 --> 00:07:14,320
you can see these
questions on your browser.

129
00:07:14,320 --> 00:07:15,790
And if you can't, that's fine.

130
00:07:15,790 --> 00:07:18,217
We'll share some aggregate
data, nonetheless.

131
00:07:18,217 --> 00:07:20,800
But you should have an opportunity
to tap one of your answers.

132
00:07:20,800 --> 00:07:26,920
And we'll give folks a few more seconds
if you'd like to play along at home.

133
00:07:26,920 --> 00:07:30,670
And here in just a moment--

134
00:07:30,670 --> 00:07:32,287
probably have many people reporting.

135
00:07:32,287 --> 00:07:34,870
But why don't we go ahead and
take a look at some percentages?

136
00:07:34,870 --> 00:07:36,910
It looks like most of you--

137
00:07:36,910 --> 00:07:39,340
60% to 70%-- are proposing
just a few seconds,

138
00:07:39,340 --> 00:07:42,340
so that's not all that good news
if it's a four-digit passcode.

139
00:07:42,340 --> 00:07:44,620
Some of you are hoping
it's a few minutes.

140
00:07:44,620 --> 00:07:46,510
8% are hoping a few hours.

141
00:07:46,510 --> 00:07:50,560
More than 4% of you are really
hoping, perhaps, it's a few days.

142
00:07:50,560 --> 00:07:53,192
Well, let's actually consider
how we can answer this question

143
00:07:53,192 --> 00:07:55,900
and make today not just conceptual,
but a little quantitative too

144
00:07:55,900 --> 00:07:58,483
and see if we can't slap some
numbers on questions like these,

145
00:07:58,483 --> 00:08:00,610
so ultimately, you can make
more informed decisions

146
00:08:00,610 --> 00:08:01,850
with your system's security.

147
00:08:01,850 --> 00:08:05,140
So for instance, when it comes
to four-digit pass codes,

148
00:08:05,140 --> 00:08:08,560
rather than just consider
how secure it is, well,

149
00:08:08,560 --> 00:08:11,630
let's make it a more precise question
like, what are the forms of attack?

150
00:08:11,630 --> 00:08:14,755
Well, the simplest attack might be just
someone grabbing your phone, be it,

151
00:08:14,755 --> 00:08:17,320
in your family, or maybe at
Starbucks, or the airport,

152
00:08:17,320 --> 00:08:21,760
or the like and just starting all
possible combinations, maybe 0000,

153
00:08:21,760 --> 00:08:24,490
then 0001, and 0002.

154
00:08:24,490 --> 00:08:26,660
We could maybe automate
this a little bit.

155
00:08:26,660 --> 00:08:30,220
So for instance, I might
potentially be able to do something

156
00:08:30,220 --> 00:08:32,720
like robotosize this here.

157
00:08:32,720 --> 00:08:35,140
Let me go ahead and full
screen a quick video here

158
00:08:35,140 --> 00:08:38,679
that's just going to paint a picture
in just a moment on the screen of how,

159
00:08:38,679 --> 00:08:41,470
if we're a really clever adversary
and know how to build things,

160
00:08:41,470 --> 00:08:44,270
well, at least, maybe we could
automate some of that process.

161
00:08:44,270 --> 00:08:46,420
So here's an Android phone
sitting on a counter.

162
00:08:46,420 --> 00:08:50,920
Here's a very simple tripod and a little
touch device robotically doing all

163
00:08:50,920 --> 00:08:56,260
of that hacking for you starting at
0000 probably all the way up to 9999.

164
00:08:56,260 --> 00:09:00,340
Now, that too wasn't necessarily
all that fast, but at least,

165
00:09:00,340 --> 00:09:02,530
the adversary can step
away and doesn't actually

166
00:09:02,530 --> 00:09:05,140
have to be bothered
with the time involved,

167
00:09:05,140 --> 00:09:08,390
the cost involved, in actually
hacking that particular device.

168
00:09:08,390 --> 00:09:11,800
Well, let's go one level deeper,
a little more interestingly,

169
00:09:11,800 --> 00:09:17,950
and consider here how much time
really this so-called brute force

170
00:09:17,950 --> 00:09:18,700
attack would take.

171
00:09:18,700 --> 00:09:21,825
And that's actually a term of art, much
like in yesteryear when maybe there

172
00:09:21,825 --> 00:09:24,820
was a battering ram trying to
brute force their way into a castle

173
00:09:24,820 --> 00:09:25,870
or something like that.

174
00:09:25,870 --> 00:09:29,380
A brute force attack digitally
is just someone trying manually

175
00:09:29,380 --> 00:09:32,950
all possible codes or maybe
robotically trying all possible codes,

176
00:09:32,950 --> 00:09:35,500
but generally automating
the process in some way

177
00:09:35,500 --> 00:09:37,270
to go through all possibilities.

178
00:09:37,270 --> 00:09:42,820
Well, if you've got, for
instance, a four-digit passcode--

179
00:09:42,820 --> 00:09:46,150
let's ask maybe a follow-up question
here, not how long it will take,

180
00:09:46,150 --> 00:09:51,190
but how many possible
four-digit passcodes are there?

181
00:09:51,190 --> 00:09:53,420
Because then maybe, we
can do some quick math.

182
00:09:53,420 --> 00:09:57,132
And if every passcode takes me a second,
or a few milliseconds, or the like,

183
00:09:57,132 --> 00:10:00,340
then I think we can try to extrapolate
from that whether the first answer was

184
00:10:00,340 --> 00:10:03,920
seconds, or minutes, or days,
or hours, or something else.

185
00:10:03,920 --> 00:10:06,815
So how many four-digit
passcodes are possible?

186
00:10:06,815 --> 00:10:09,940
If you take out your same device, it
should have just changed automatically

187
00:10:09,940 --> 00:10:13,750
if it doesn't seem to have maybe reload
your browser with some menu option.

188
00:10:13,750 --> 00:10:17,740
And then tap in here, how many
four-digit passcodes are possible?

189
00:10:17,740 --> 00:10:20,860
Four total, 40, 9,999, 10,000.

190
00:10:20,860 --> 00:10:24,610
Or unsure is OK too.

191
00:10:24,610 --> 00:10:25,510
So let's see.

192
00:10:25,510 --> 00:10:27,900
We'll give you a few more moments.

193
00:10:27,900 --> 00:10:30,600
How many four-digit
passcodes are possible?

194
00:10:30,600 --> 00:10:33,160
And shall we reveal the results?

195
00:10:33,160 --> 00:10:37,230
So now, it looks like a few of you--

196
00:10:37,230 --> 00:10:42,540
2% of you are saying just
for passcodes, 40, 9,999.

197
00:10:42,540 --> 00:10:44,340
There's definitely some contention here.

198
00:10:44,340 --> 00:10:45,900
And 6% are unsure.

199
00:10:45,900 --> 00:10:48,850
Well, how do we wrap
our minds around this?

200
00:10:48,850 --> 00:10:50,760
Well, let's just do
this real simple here.

201
00:10:50,760 --> 00:10:54,160
Let me switch back over
to doing a bit of math.

202
00:10:54,160 --> 00:10:57,932
And if we have here 10 possibilities
for each digit, if there's four digits,

203
00:10:57,932 --> 00:11:01,140
each digit can be zero, one, two, three,
four, five, six, seven, eight, nine.

204
00:11:01,140 --> 00:11:02,590
So that's 10 possibilities.

205
00:11:02,590 --> 00:11:05,490
So if you think about the
number of permutations,

206
00:11:05,490 --> 00:11:09,540
that's 10 possibilities for the first
digit times 10 for the next, times 10

207
00:11:09,540 --> 00:11:11,290
for the next, times 10 for the next.

208
00:11:11,290 --> 00:11:14,130
And so if we do that out, 10
times, 10 times, 10 times, 10 or 10

209
00:11:14,130 --> 00:11:20,220
to the fourth, there are, indeed-- and
66% of you found 10,000 possibilities.

210
00:11:20,220 --> 00:11:22,620
And so now we can kind of
work backwards and decide,

211
00:11:22,620 --> 00:11:25,800
how long is it going to take for an
adversary to hack into this phone?

212
00:11:25,800 --> 00:11:28,800
Because if it's one attack,
one guess per second,

213
00:11:28,800 --> 00:11:31,470
well, that's going to map
out to 10,000 seconds,

214
00:11:31,470 --> 00:11:34,560
but maybe not if the adversary
isn't a roboticist or a human.

215
00:11:34,560 --> 00:11:37,170
What if they're a software
programmer or someone who

216
00:11:37,170 --> 00:11:40,560
has taken even a class
introductory, like CS50 and learned

217
00:11:40,560 --> 00:11:41,760
a little bit of programming?

218
00:11:41,760 --> 00:11:43,590
Well, a little bit
frighteningly, it's not

219
00:11:43,590 --> 00:11:46,470
all that hard to hack
into systems if you just

220
00:11:46,470 --> 00:11:49,890
know how to code, too, and really have
the computer do your work for you.

221
00:11:49,890 --> 00:11:53,190
So in fact, let me go ahead and change
over to another screen on my computer

222
00:11:53,190 --> 00:11:53,910
here.

223
00:11:53,910 --> 00:11:56,307
This is different for students
in the group from VS Code.

224
00:11:56,307 --> 00:11:58,140
This is just a black
and white version of it

225
00:11:58,140 --> 00:11:59,670
that we've used briefly in the past.

226
00:11:59,670 --> 00:12:03,090
And I'm just going to go ahead and
create a program called crack.py.

227
00:12:03,090 --> 00:12:05,160
To crack something
just technically means

228
00:12:05,160 --> 00:12:08,380
to figure out what it is, figure
out a password in this case.

229
00:12:08,380 --> 00:12:11,730
And .py means I'm going to use a
programming language that we here

230
00:12:11,730 --> 00:12:15,390
in CS50 have been dabbling in the past
couple of weeks with more to come next

231
00:12:15,390 --> 00:12:16,300
week as well.

232
00:12:16,300 --> 00:12:20,220
So it turns out, and you need not
understand each of these lines of code,

233
00:12:20,220 --> 00:12:25,518
if I want to try, maybe, generating
all 10,000 possible codes,

234
00:12:25,518 --> 00:12:27,060
I'm not going to bother with a robot.

235
00:12:27,060 --> 00:12:29,250
I've got all these cables
coming out of my computer.

236
00:12:29,250 --> 00:12:32,250
And odds are one of them is a
USB cable or a lightning cable.

237
00:12:32,250 --> 00:12:36,120
Surely, we could figure out how to
connect laptop or desktop to phone

238
00:12:36,120 --> 00:12:39,810
and just automate the process nowadays
by just sending all of the numbers

239
00:12:39,810 --> 00:12:44,220
into the phone until one unlocks the
trick just like in the movies or TV.

240
00:12:44,220 --> 00:12:47,410
Well, in Python, I could write a
program that does this as follows.

241
00:12:47,410 --> 00:12:51,510
I can import, so to speak, all of the
decimal digits, zero through nine.

242
00:12:51,510 --> 00:12:54,990
And this, for students in the room,
is just a slightly better version

243
00:12:54,990 --> 00:12:57,580
of typing out 10 different
numbers manually.

244
00:12:57,580 --> 00:13:00,540
I can also import from
a library, so to speak,

245
00:13:00,540 --> 00:13:02,910
called itertools for
iteration tools, which

246
00:13:02,910 --> 00:13:04,720
means to do something again and again.

247
00:13:04,720 --> 00:13:09,210
I can import a function called
product, which means the cross product.

248
00:13:09,210 --> 00:13:11,850
Combine this with this
some number of times.

249
00:13:11,850 --> 00:13:13,680
And then it's just two
more lines of code.

250
00:13:13,680 --> 00:13:15,850
I can use what's called
a loop in programming.

251
00:13:15,850 --> 00:13:19,770
So for every pass code in
the cross product of all 10

252
00:13:19,770 --> 00:13:24,420
of those digits repeated,
a total of four times--

253
00:13:24,420 --> 00:13:27,480
let me go ahead and-- rather
than bother connecting my phone

254
00:13:27,480 --> 00:13:29,520
and hacking my own phone,
let me just print out

255
00:13:29,520 --> 00:13:31,680
every one of those 10,000
codes on the screen,

256
00:13:31,680 --> 00:13:33,833
and we'll see how fast
the hacker could do this.

257
00:13:33,833 --> 00:13:34,500
Let me go ahead.

258
00:13:34,500 --> 00:13:37,750
And print and with an asterisk, which
is a little trick to format it nicely,

259
00:13:37,750 --> 00:13:39,750
I'm going to print out
each of those pass codes.

260
00:13:39,750 --> 00:13:43,350
And that's it, four lines of
code, maybe 40 seconds of talking,

261
00:13:43,350 --> 00:13:45,960
but maybe really four seconds
of coding if I actually

262
00:13:45,960 --> 00:13:47,250
did this without the audience.

263
00:13:47,250 --> 00:13:49,470
And now let me go ahead
and save the file.

264
00:13:49,470 --> 00:13:52,410
And I'm going to run, as we
do every day in class of late.

265
00:13:52,410 --> 00:13:54,370
Python of crack.py.

266
00:13:54,370 --> 00:13:57,840
And when I hit Enter, I should see on
the screen all 10,000 possibilities

267
00:13:57,840 --> 00:14:00,180
from 0000 9999.

268
00:14:00,180 --> 00:14:00,810
So let's see.

269
00:14:00,810 --> 00:14:04,560
Is it a few seconds,
minutes, hours, or days?

270
00:14:04,560 --> 00:14:05,790
Done.

271
00:14:05,790 --> 00:14:08,518
So barely even seconds plural if that.

272
00:14:08,518 --> 00:14:11,310
So that should be a little
disconcerting because all that adversary

273
00:14:11,310 --> 00:14:14,393
needs to do is grab your phone off the
counter, plug in a cable, and boom.

274
00:14:14,393 --> 00:14:15,390
They're done.

275
00:14:15,390 --> 00:14:18,720
There's no ticking clock or
worries as in the movies or TV

276
00:14:18,720 --> 00:14:20,700
that maybe you're going
to come into the room.

277
00:14:20,700 --> 00:14:22,900
You don't need that much
of a window of time.

278
00:14:22,900 --> 00:14:25,000
So what would be better than this?

279
00:14:25,000 --> 00:14:27,630
Well, let's consider
what our options might

280
00:14:27,630 --> 00:14:30,392
be if we don't want to just
use four-digit pass code.

281
00:14:30,392 --> 00:14:32,850
Some of you, indeed, might have
better passcodes than that.

282
00:14:32,850 --> 00:14:36,930
And maybe, you use four-letter
passcodes instead, so A through Z,

283
00:14:36,930 --> 00:14:38,820
maybe uppercase and lowercase.

284
00:14:38,820 --> 00:14:41,110
That starts to make things
a little more interesting.

285
00:14:41,110 --> 00:14:43,080
So should we poll this question too?

286
00:14:43,080 --> 00:14:46,592
If we upgrade from four
digits to just four letters,

287
00:14:46,592 --> 00:14:49,800
English letters, A through Z, uppercase
and lowercase-- why don't we go ahead

288
00:14:49,800 --> 00:14:55,260
and pol the group here and ask how
many four-letter passcodes are there

289
00:14:55,260 --> 00:14:57,850
instead?

290
00:14:57,850 --> 00:15:00,510
So this time, the range starts at four.

291
00:15:00,510 --> 00:15:03,090
Still not the right
answer, though, this time.

292
00:15:03,090 --> 00:15:05,800
How many four-letter
pass codes are possible?

293
00:15:05,800 --> 00:15:08,870


294
00:15:08,870 --> 00:15:11,090
[INAUDIBLE]

295
00:15:11,090 --> 00:15:12,290
Take a couple more seconds.

296
00:15:12,290 --> 00:15:16,590


297
00:15:16,590 --> 00:15:18,030
All right.

298
00:15:18,030 --> 00:15:20,110
Almost a couple hundred
responses in already.

299
00:15:20,110 --> 00:15:21,465
A few more seconds.

300
00:15:21,465 --> 00:15:24,920


301
00:15:24,920 --> 00:15:31,070
And why don't we go ahead and
reveal now the answers, which are--

302
00:15:31,070 --> 00:15:35,690
OK, so we solved a couple
of problems at least.

303
00:15:35,690 --> 00:15:37,430
OK, someone's just messing with us now.

304
00:15:37,430 --> 00:15:37,930
All right.

305
00:15:37,930 --> 00:15:39,620
So it looks like most of you--

306
00:15:39,620 --> 00:15:43,350
76% of you have claimed it's
seven million plus possibilities.

307
00:15:43,350 --> 00:15:46,040
So that's encouraging because
that's a whole order of magnitude

308
00:15:46,040 --> 00:15:46,792
more than before.

309
00:15:46,792 --> 00:15:49,250
Well, let's figure out how we
might do this mathematically.

310
00:15:49,250 --> 00:15:52,100
So if we've got 26
lowercase, 26 uppercase,

311
00:15:52,100 --> 00:15:55,020
that's 52 possibilities now
for each of those four digits.

312
00:15:55,020 --> 00:15:57,680
So that's 52 times
itself four times, which,

313
00:15:57,680 --> 00:16:00,530
indeed, either off the top
of your head a good guess,

314
00:16:00,530 --> 00:16:03,500
a calculator on the same device
you're using right now, indeed,

315
00:16:03,500 --> 00:16:05,790
gives us seven million instead.

316
00:16:05,790 --> 00:16:08,250
Well, what might be
slightly better than that?

317
00:16:08,250 --> 00:16:09,837
Well, maybe four characters.

318
00:16:09,837 --> 00:16:12,170
And this, indeed, is what
your Macs, PCs, and phones are

319
00:16:12,170 --> 00:16:15,860
urging us to do nowadays, not just
numbers, not just letters, but really

320
00:16:15,860 --> 00:16:19,940
annoying punctuation, so it really
looks cryptic not just to the adversary,

321
00:16:19,940 --> 00:16:22,030
but also to you and me, unfortunately.

322
00:16:22,030 --> 00:16:23,030
And that's the downside.

323
00:16:23,030 --> 00:16:27,500
But here now, we have a mental model,
and really, a computational framework

324
00:16:27,500 --> 00:16:29,737
via which we can evaluate
the security of these.

325
00:16:29,737 --> 00:16:31,820
And I'll go ahead and spoil
some of the math here.

326
00:16:31,820 --> 00:16:37,080
If we've got 52 letters of the alphabet,
uppercase and lowercase, 10 digits,

327
00:16:37,080 --> 00:16:40,850
and if I count them out on my
keyboard, about 32 punctuation symbols

328
00:16:40,850 --> 00:16:43,190
in typical English grammar,
that actually gives us

329
00:16:43,190 --> 00:16:49,230
94 possibilities now, which is
up from 52, which is up from 10.

330
00:16:49,230 --> 00:16:50,780
So now, we're really moving.

331
00:16:50,780 --> 00:16:53,700
And now that would give us
78 million possibilities,

332
00:16:53,700 --> 00:16:55,280
so another order of magnitude.

333
00:16:55,280 --> 00:16:58,202
Now, it's still going to be
relatively fast because you know what?

334
00:16:58,202 --> 00:16:59,160
I can actually do this.

335
00:16:59,160 --> 00:17:01,220
Let me go back into my code here.

336
00:17:01,220 --> 00:17:03,500
Let me reopen this same program.

337
00:17:03,500 --> 00:17:06,750
And I can point out just how
easy it is to make these changes.

338
00:17:06,750 --> 00:17:11,630
Instead of importing digits as before,
I can import, as your child might know,

339
00:17:11,630 --> 00:17:14,940
ascii letters, which are A
through Z, uppercase, lowercase.

340
00:17:14,940 --> 00:17:18,319
And I can just change
this here, ascii letters.

341
00:17:18,319 --> 00:17:21,349
And so this was that first version
where we just changed to letters.

342
00:17:21,349 --> 00:17:22,640
Let me now rerun the code.

343
00:17:22,640 --> 00:17:26,220
And instead of seeing numbers, we'll
see letters flying across the screen.

344
00:17:26,220 --> 00:17:29,060
And if I walk over here to
the screen, we'll see that.

345
00:17:29,060 --> 00:17:33,170
By the time I get here, we're halfway
through the entire alphabet lowercase.

346
00:17:33,170 --> 00:17:35,120
If I now start walking
away, I think, yeah,

347
00:17:35,120 --> 00:17:37,700
we're already done now
with uppercase as well.

348
00:17:37,700 --> 00:17:39,990
If I upgrade this slightly further--

349
00:17:39,990 --> 00:17:44,300
let's go ahead and take it one more
level and, perhaps, do, let's say,

350
00:17:44,300 --> 00:17:47,870
ascii letters, and
digits, and punctuation.

351
00:17:47,870 --> 00:17:50,550
And this would be the
Pythonic way to say that.

352
00:17:50,550 --> 00:17:53,600
And I'm going to add to those
letters those same digits,

353
00:17:53,600 --> 00:17:55,940
those same punctuation symbols.

354
00:17:55,940 --> 00:17:58,700
Let me shrink my font just so the
code still fits on the screen.

355
00:17:58,700 --> 00:18:01,880
And what we now have is with
a two seconds of changes,

356
00:18:01,880 --> 00:18:04,490
a program that if I run this version--

357
00:18:04,490 --> 00:18:08,100
whoops-- without the
typographical error--

358
00:18:08,100 --> 00:18:10,520
this is what we call in CS50 a bug.

359
00:18:10,520 --> 00:18:12,780
So now, we run the same--

360
00:18:12,780 --> 00:18:17,030
this is what we call
in CS50 a second bug--

361
00:18:17,030 --> 00:18:17,846
punctuation.

362
00:18:17,846 --> 00:18:20,490


363
00:18:20,490 --> 00:18:21,865
This is where I cross my fingers.

364
00:18:21,865 --> 00:18:22,365
OK.

365
00:18:22,365 --> 00:18:25,440
So now it's going to be a little hard
to see as flies across the screen.

366
00:18:25,440 --> 00:18:28,700
But you probably are seeing glimpses
of some weird punctuation characters

367
00:18:28,700 --> 00:18:29,250
as well.

368
00:18:29,250 --> 00:18:32,270
And I won't waste our time trying
to talk through this because this

369
00:18:32,270 --> 00:18:33,470
is going to take longer.

370
00:18:33,470 --> 00:18:34,700
We're still in the lowercase.

371
00:18:34,700 --> 00:18:35,930
I'm still over here already.

372
00:18:35,930 --> 00:18:39,290
We've not even gotten
to N, now O, then P.

373
00:18:39,290 --> 00:18:40,790
So this is going to run longer.

374
00:18:40,790 --> 00:18:44,630
But let's end with one final question
on the security of all these systems.

375
00:18:44,630 --> 00:18:47,540
I'm going to cancel that by
hitting Control C on my keyboard.

376
00:18:47,540 --> 00:18:51,950
And let's ask the question instead,
if we use eight-character passwords,

377
00:18:51,950 --> 00:18:53,360
so twice as many characters.

378
00:18:53,360 --> 00:18:55,790
But even that is not terribly long.

379
00:18:55,790 --> 00:18:58,670
This is eight characters alone
on the stage, eight characters.

380
00:18:58,670 --> 00:19:01,100
Using letters, numbers, and
punctuation might be better.

381
00:19:01,100 --> 00:19:03,530
Let's do one final
vote here, if we could.

382
00:19:03,530 --> 00:19:07,700
On your same device, how many
eight-character possibilities

383
00:19:07,700 --> 00:19:12,800
are there now for these passcodes?

384
00:19:12,800 --> 00:19:17,000
And now four didn't even
make the list this time.

385
00:19:17,000 --> 00:19:21,363
All right, a few more seconds,
about 100 responses so far.

386
00:19:21,363 --> 00:19:23,780
How about we go ahead-- and,
Carter, if you wouldn't mind,

387
00:19:23,780 --> 00:19:28,220
let's reveal the results based on the
vote, a pretty decent spread here.

388
00:19:28,220 --> 00:19:30,317
Although the quadrillions
are quickly buzzing in.

389
00:19:30,317 --> 00:19:32,150
And they're contending
with the others here.

390
00:19:32,150 --> 00:19:34,700
Looks like 44% of you said quintillion.

391
00:19:34,700 --> 00:19:36,740
34% said quadrillion.

392
00:19:36,740 --> 00:19:39,800
And this time, for the
first time, you overbid.

393
00:19:39,800 --> 00:19:44,270
So indeed, if we go back to the math
here, at least, the majority over bid.

394
00:19:44,270 --> 00:19:46,160
If we have eight-character
passcodes that

395
00:19:46,160 --> 00:19:50,600
gives us 94 times itself eight
times or 94 to the eighth power.

396
00:19:50,600 --> 00:20:01,430
And in fact, that gives us
roughly 6,095,689,385,410,816

397
00:20:01,430 --> 00:20:02,630
possible passcodes.

398
00:20:02,630 --> 00:20:04,140
Now, what does that mean?

399
00:20:04,140 --> 00:20:07,700
Well, the adversary's
algorithm, the step-by-step code

400
00:20:07,700 --> 00:20:10,580
that they write to try to hack
into your phone, is no different.

401
00:20:10,580 --> 00:20:13,670
And honestly, if your passcode
is eight characters long,

402
00:20:13,670 --> 00:20:18,500
but they're are 00000000, you're
no more secure fundamentally.

403
00:20:18,500 --> 00:20:22,010
You really want to be somewhere in
the sweet spot of that massive range

404
00:20:22,010 --> 00:20:25,200
of values, so that if the adversary
tries this brute force attack just

405
00:20:25,200 --> 00:20:28,590
running through all possibilities,
they will eventually reach

406
00:20:28,590 --> 00:20:31,320
your passcode just mathematically.

407
00:20:31,320 --> 00:20:32,400
It will be there.

408
00:20:32,400 --> 00:20:34,920
Hopefully, though-- well,
maybe not hopefully-- you

409
00:20:34,920 --> 00:20:37,380
and I and they will be
gone from this world

410
00:20:37,380 --> 00:20:39,840
because that much time will have passed.

411
00:20:39,840 --> 00:20:44,440
And if we do out the math here, this
number of seconds, for instance,

412
00:20:44,440 --> 00:20:48,940
is long past when we
will no longer be here.

413
00:20:48,940 --> 00:20:50,400
So that's the sort of measures.

414
00:20:50,400 --> 00:20:53,460
We don't sort of fundamentally change
the equation for the adversary.

415
00:20:53,460 --> 00:20:54,670
It's still the same risk.

416
00:20:54,670 --> 00:20:55,795
It's still the same attack.

417
00:20:55,795 --> 00:20:59,370
But you significantly drive down the
probability of success on their part.

418
00:20:59,370 --> 00:21:02,880
Or conceptually, you drive
up the cost to the adversary.

419
00:21:02,880 --> 00:21:05,850
And indeed, even in the
physical world, this is true.

420
00:21:05,850 --> 00:21:08,310
You just want your passcode
in the digital world

421
00:21:08,310 --> 00:21:11,220
really to be better than
someone else's because you

422
00:21:11,220 --> 00:21:12,985
want someone else's
passcode to be the one

423
00:21:12,985 --> 00:21:14,610
that the adversary does something with.

424
00:21:14,610 --> 00:21:18,000
Just like in the physical world,
even though it's a bit uncomfortable

425
00:21:18,000 --> 00:21:21,990
to consider, your house
doesn't need to be 100% secure.

426
00:21:21,990 --> 00:21:24,060
And indeed, it's
difficult to make it such.

427
00:21:24,060 --> 00:21:26,100
There's always going to
be a point of weakness.

428
00:21:26,100 --> 00:21:28,810
Maybe it's that window, the
door, or something like that.

429
00:21:28,810 --> 00:21:33,060
But if your home is more secure than the
next door home, just probabilistically,

430
00:21:33,060 --> 00:21:34,710
you are more secure.

431
00:21:34,710 --> 00:21:35,760
You're not secure.

432
00:21:35,760 --> 00:21:37,710
And indeed, any website
you see down the road

433
00:21:37,710 --> 00:21:41,400
that says, we are secure because
we do X, Y, or Z, that's nonsense.

434
00:21:41,400 --> 00:21:44,910
Security is really about
comparisons and evaluating things

435
00:21:44,910 --> 00:21:49,870
if quantitatively relative to some other
system, relative to some other code.

436
00:21:49,870 --> 00:21:51,520
So what's the takeaway here?

437
00:21:51,520 --> 00:21:55,320
Well, hopefully, a non-trivial number of
you will go home this weekend on Monday

438
00:21:55,320 --> 00:21:57,803
and change at least one passcode.

439
00:21:57,803 --> 00:21:59,470
But there's going to be a tradeoff here.

440
00:21:59,470 --> 00:22:01,260
We talk about this all the time in CS50.

441
00:22:01,260 --> 00:22:06,210
Any time we improve something, we pay
some price in time, in performance,

442
00:22:06,210 --> 00:22:07,360
in cost, somewhere else.

443
00:22:07,360 --> 00:22:10,470
So what's the downside then of
this advice that you should use

444
00:22:10,470 --> 00:22:13,710
minimally eight-character passcodes?

445
00:22:13,710 --> 00:22:16,350
Why might you want to
say nay and not do this?

446
00:22:16,350 --> 00:22:17,835
AUDIENCE: You have to remember it.

447
00:22:17,835 --> 00:22:18,390
DAVID J. MALAN: Say again?

448
00:22:18,390 --> 00:22:19,350
AUDIENCE: You have to remember it.

449
00:22:19,350 --> 00:22:21,308
DAVID J. MALAN: You have
to remember it, right?

450
00:22:21,308 --> 00:22:23,050
And so here, there is some sociology.

451
00:22:23,050 --> 00:22:24,570
There's some human behavior.

452
00:22:24,570 --> 00:22:25,980
Some of you might have
colleagues, if you're

453
00:22:25,980 --> 00:22:27,688
working in the real
world, at least, back

454
00:22:27,688 --> 00:22:30,510
in healthier times when you had
colleagues with desks in cubicles.

455
00:22:30,510 --> 00:22:33,218
And there's probably one person
in the office with a post-it note

456
00:22:33,218 --> 00:22:35,370
on their monitor with their passcode.

457
00:22:35,370 --> 00:22:41,160
It's a bit of a cybersecurity offense,
but it's also a real world side

458
00:22:41,160 --> 00:22:44,250
effect, maybe of corporate policies,
that aren't really calibrated

459
00:22:44,250 --> 00:22:45,640
for human behavior.

460
00:22:45,640 --> 00:22:47,610
So we'll see if there's
some other defenses.

461
00:22:47,610 --> 00:22:49,650
And indeed, let me propose
that we talk briefly

462
00:22:49,650 --> 00:22:52,530
about one that actually tends
to kick in automatically.

463
00:22:52,530 --> 00:22:54,840
Even if your passcode is
not as strong as we've just

464
00:22:54,840 --> 00:22:57,940
seen, one of these six
quadrillion possibilities, well,

465
00:22:57,940 --> 00:22:59,260
what could we do instead?

466
00:22:59,260 --> 00:23:01,590
Well, has anyone-- and
I'll zoom in on this here--

467
00:23:01,590 --> 00:23:05,940
accidentally locked themselves
out of their own phone before?

468
00:23:05,940 --> 00:23:07,380
When does that happen?

469
00:23:07,380 --> 00:23:08,828
Yeah, when you try the password--

470
00:23:08,828 --> 00:23:09,870
AUDIENCE: Too many times.

471
00:23:09,870 --> 00:23:11,020
DAVID J. MALAN: Yeah, so too many times.

472
00:23:11,020 --> 00:23:12,820
Maybe your finger is slightly off.

473
00:23:12,820 --> 00:23:17,010
Maybe you're slightly off, and you just
don't input the same passcode correctly

474
00:23:17,010 --> 00:23:18,803
after five times, 10 times.

475
00:23:18,803 --> 00:23:20,220
There's some reasonable threshold.

476
00:23:20,220 --> 00:23:21,580
And why does that happen?

477
00:23:21,580 --> 00:23:24,270
Well, Apple and Google
equivalently figure just

478
00:23:24,270 --> 00:23:27,060
probabilistically if after
10 guesses, you still

479
00:23:27,060 --> 00:23:29,932
haven't typed in the right
passcode, probably, you're not you.

480
00:23:29,932 --> 00:23:31,890
You're someone else who's
picked up your phone,

481
00:23:31,890 --> 00:23:33,660
so we're just going to go
ahead and lock you out.

482
00:23:33,660 --> 00:23:35,200
Now, what's the effect of this?

483
00:23:35,200 --> 00:23:37,860
Well, this means now that each
of those possible passcodes

484
00:23:37,860 --> 00:23:39,870
no longer takes roughly one second.

485
00:23:39,870 --> 00:23:41,890
Now it takes roughly one minute.

486
00:23:41,890 --> 00:23:43,360
So the attack is still the same.

487
00:23:43,360 --> 00:23:46,560
But if it's now one passcode
or 10 guesses per minute,

488
00:23:46,560 --> 00:23:50,680
we have significantly by a factor of
60 in this story slowed things down.

489
00:23:50,680 --> 00:23:52,680
And unfortunately, does
anyone know what happens

490
00:23:52,680 --> 00:23:54,967
if you screw up again after a minute?

491
00:23:54,967 --> 00:23:56,160
AUDIENCE: It goes longer.

492
00:23:56,160 --> 00:23:57,120
DAVID J. MALAN: Yeah, it goes longer.

493
00:23:57,120 --> 00:23:59,100
It's like five minutes
and then 10 minutes.

494
00:23:59,100 --> 00:24:01,477
And Google is kind of
obnoxious about it.

495
00:24:01,477 --> 00:24:03,060
They don't even give you a time frame.

496
00:24:03,060 --> 00:24:05,410
They just say, try again later.

497
00:24:05,410 --> 00:24:09,450
And so that keeps not only the
adversary out, but also potentially you.

498
00:24:09,450 --> 00:24:11,017
So therein lies that tradeoff.

499
00:24:11,017 --> 00:24:14,100
If you've forgotten your code, if--
nowadays, your finger is slightly wet,

500
00:24:14,100 --> 00:24:16,230
so the screen isn't
responding correctly.

501
00:24:16,230 --> 00:24:18,540
These could be usability downsides too.

502
00:24:18,540 --> 00:24:21,810
So security is really just
about finding the sweet spot

503
00:24:21,810 --> 00:24:24,180
among these various tradeoffs here.

504
00:24:24,180 --> 00:24:25,860
But there's other mechanisms too.

505
00:24:25,860 --> 00:24:30,030
And some of you might recognize this
screen from Gmail via which, of course,

506
00:24:30,030 --> 00:24:30,660
you log in.

507
00:24:30,660 --> 00:24:36,030
But after you log into Gmail or similar
websites, or apps, or systems at work

508
00:24:36,030 --> 00:24:39,780
nowadays, especially, you
might be presented with what's

509
00:24:39,780 --> 00:24:42,240
called two-factor authentication.

510
00:24:42,240 --> 00:24:46,110
And what is this in a
nutshell in layperson's terms?

511
00:24:46,110 --> 00:24:48,597
Many of you, if you do
anything digitally at work,

512
00:24:48,597 --> 00:24:49,680
might have to do this now.

513
00:24:49,680 --> 00:24:50,180
Yeah?

514
00:24:50,180 --> 00:24:51,860
AUDIENCE: It sends a text to your phone.

515
00:24:51,860 --> 00:24:52,860
DAVID J. MALAN: Exactly.

516
00:24:52,860 --> 00:24:56,040
You get texted at your phone,
an additional code that's

517
00:24:56,040 --> 00:24:57,180
not your same password.

518
00:24:57,180 --> 00:24:59,580
It's typically a numeric
code, maybe six digits long.

519
00:24:59,580 --> 00:25:01,510
It expires after a minute or 10 minutes.

520
00:25:01,510 --> 00:25:03,030
But why is this a good thing?

521
00:25:03,030 --> 00:25:06,795
Well, one, it's no longer just a
piece of information that you know

522
00:25:06,795 --> 00:25:08,760
or that you might have written down.

523
00:25:08,760 --> 00:25:12,180
It's information that changes
every time you try to log in.

524
00:25:12,180 --> 00:25:15,100
But more importantly, it's a
fundamentally second factor, which

525
00:25:15,100 --> 00:25:16,725
means it's not just something you know.

526
00:25:16,725 --> 00:25:18,180
Now it's something you have.

527
00:25:18,180 --> 00:25:20,832
So you, for instance, are
the only one theoretically

528
00:25:20,832 --> 00:25:22,290
that should be receiving that code.

529
00:25:22,290 --> 00:25:24,960
And so now the adversary, if they
want to get into your account,

530
00:25:24,960 --> 00:25:28,800
not only have to guess, or brute
force, or maybe read off of a post-it

531
00:25:28,800 --> 00:25:29,880
note your password.

532
00:25:29,880 --> 00:25:32,850
They also have to physically
have access now to that phone.

533
00:25:32,850 --> 00:25:35,940
So there's still a threat,
absolutely, but it's not everyone

534
00:25:35,940 --> 00:25:38,080
on the internet with
an internet connection.

535
00:25:38,080 --> 00:25:39,900
Now it's only the people in Starbucks.

536
00:25:39,900 --> 00:25:41,430
Now it's only the people at work.

537
00:25:41,430 --> 00:25:44,400
Now it's only the people in
your home who might have access

538
00:25:44,400 --> 00:25:45,400
to that second factor.

539
00:25:45,400 --> 00:25:48,690
So there too, it just raises the bar
to the adversary making it harder,

540
00:25:48,690 --> 00:25:52,390
more time consuming, more geographically
impossible for them to attack you.

541
00:25:52,390 --> 00:25:54,970
But what's the downside of
two-factor authentication,

542
00:25:54,970 --> 00:25:56,850
whether it's a device--
or even nowadays,

543
00:25:56,850 --> 00:25:59,232
it's in software, whether
it's on your keychain

544
00:25:59,232 --> 00:26:01,440
or on your phone where you're
prompted for this code.

545
00:26:01,440 --> 00:26:05,910
What's a downside that some of
us have probably experienced too?

546
00:26:05,910 --> 00:26:07,243
AUDIENCE: You forget your phone.

547
00:26:07,243 --> 00:26:09,535
DAVID J. MALAN: You forget
your cell phone, absolutely.

548
00:26:09,535 --> 00:26:11,910
Right, the factor that you
have, you don't have with you.

549
00:26:11,910 --> 00:26:14,040
Or maybe, you're in a basement
somewhere, don't have reception.

550
00:26:14,040 --> 00:26:14,790
You're on a plane.

551
00:26:14,790 --> 00:26:15,850
You can't get the code.

552
00:26:15,850 --> 00:26:17,850
And so there too are these tradeoffs.

553
00:26:17,850 --> 00:26:19,920
And even IT departments
need to keep that in mind

554
00:26:19,920 --> 00:26:21,462
because what does that mean for them?

555
00:26:21,462 --> 00:26:23,370
Well, if you don't have
your phone with you

556
00:26:23,370 --> 00:26:25,980
and you are in the habit of
calling IT to help you fix this,

557
00:26:25,980 --> 00:26:29,740
now there's a cost, a human cost,
maybe even a financial cost.

558
00:26:29,740 --> 00:26:34,022
And so IT policy nowadays is really
just about finding the right balance

559
00:26:34,022 --> 00:26:35,730
and where we want to
spend our resources,

560
00:26:35,730 --> 00:26:38,850
but at least raise the
bar to the adversary.

561
00:26:38,850 --> 00:26:41,280
But of course, there's other ways too.

562
00:26:41,280 --> 00:26:43,680
And this is going to be one
of our homework assignments

563
00:26:43,680 --> 00:26:44,880
if you will after today.

564
00:26:44,880 --> 00:26:46,920
There's this software
called password managers.

565
00:26:46,920 --> 00:26:50,280
And no need to buzz in on your phone,
but maybe with a physical hand.

566
00:26:50,280 --> 00:26:54,330
How many folks here
use a password manager?

567
00:26:54,330 --> 00:26:57,660
OK, let me ballpark this
at 10%, 20%, perhaps.

568
00:26:57,660 --> 00:27:00,960
So we've got 80% upside here and
a lesson learned potentially.

569
00:27:00,960 --> 00:27:03,450
So a password manager is
just a piece of software

570
00:27:03,450 --> 00:27:07,287
on your Mac, your PC, or your phone
nowadays that manages your passwords.

571
00:27:07,287 --> 00:27:08,370
Well, what does that mean?

572
00:27:08,370 --> 00:27:10,433
When you go to a website
for the first time

573
00:27:10,433 --> 00:27:13,600
or you download an app for the first
time and you have to create an account,

574
00:27:13,600 --> 00:27:16,918
you can still use your email
address, or David as your username,

575
00:27:16,918 --> 00:27:18,210
or whatever your name might be.

576
00:27:18,210 --> 00:27:20,250
So you don't have to
change that methodology.

577
00:27:20,250 --> 00:27:26,400
But instead of typing in 123456 as your
same password for that website or app

578
00:27:26,400 --> 00:27:30,360
as well as for every other, now you
use the password manager software

579
00:27:30,360 --> 00:27:33,570
to generate something
difficult to guess for you.

580
00:27:33,570 --> 00:27:37,350
That is you tell the password manager,
give me an eight-character random

581
00:27:37,350 --> 00:27:41,260
passcode, not 0000, but something
with punctuation, with numbers,

582
00:27:41,260 --> 00:27:41,850
with letters.

583
00:27:41,850 --> 00:27:44,670
And better yet, the password
manager, as the name suggests,

584
00:27:44,670 --> 00:27:46,857
remembers that password for you.

585
00:27:46,857 --> 00:27:48,690
And the next time you
go to another website,

586
00:27:48,690 --> 00:27:50,898
you do it again with a
completely different password,

587
00:27:50,898 --> 00:27:54,300
maybe same username, maybe two-factor
authentication, but different password,

588
00:27:54,300 --> 00:27:56,040
different password, different password.

589
00:27:56,040 --> 00:27:57,373
And it doesn't have to be eight.

590
00:27:57,373 --> 00:28:01,410
I mean, I'm in the habit of using a
dozen, two dozen characters in total.

591
00:28:01,410 --> 00:28:04,440
And at that point, I can't even
pronounce the number of possibilities

592
00:28:04,440 --> 00:28:07,560
because it goes well
beyond the quadrillions.

593
00:28:07,560 --> 00:28:10,290
So the probability that someone's
going to get into one of those

594
00:28:10,290 --> 00:28:13,170
accounts for me now is
very, very, very low.

595
00:28:13,170 --> 00:28:16,140
And they're going to take less
interest in me and maybe more interest

596
00:28:16,140 --> 00:28:18,660
in someone else that's not
using as good of a password.

597
00:28:18,660 --> 00:28:20,380
Now, what does this mean in real terms?

598
00:28:20,380 --> 00:28:23,250
Well, when you go to log
into that managed site,

599
00:28:23,250 --> 00:28:26,297
you don't manually type
your password anymore.

600
00:28:26,297 --> 00:28:28,380
In fact, you don't generally
even need to know it.

601
00:28:28,380 --> 00:28:32,820
Nowadays, I probably don't
90-plus, 99% of my passwords.

602
00:28:32,820 --> 00:28:35,190
I entrust them to this password manager.

603
00:28:35,190 --> 00:28:40,140
Now, of course, you'd like to think that
the password manager itself is secure.

604
00:28:40,140 --> 00:28:41,620
So what might that mean?

605
00:28:41,620 --> 00:28:43,770
Well, those of you who do
use a password manager,

606
00:28:43,770 --> 00:28:46,650
how do you access that software itself?

607
00:28:46,650 --> 00:28:49,615
What's protecting your
data in your understanding?

608
00:28:49,615 --> 00:28:50,490
AUDIENCE: Biometrics.

609
00:28:50,490 --> 00:28:53,370
DAVID J. MALAN: So maybe biometrics,
like your face ID, or maybe

610
00:28:53,370 --> 00:28:55,645
your fingerprint, or maybe
more simply, what else?

611
00:28:55,645 --> 00:28:56,520
AUDIENCE: A password.

612
00:28:56,520 --> 00:28:58,103
DAVID J. MALAN: Maybe just a password.

613
00:28:58,103 --> 00:29:02,790
And hopefully, that password that
primary password, that gatekeeper,

614
00:29:02,790 --> 00:29:05,190
is not itself 123456.

615
00:29:05,190 --> 00:29:08,520
Otherwise, it doesn't matter how
secure all of the others are.

616
00:29:08,520 --> 00:29:12,240
But if you're willing to put in
the effort and pick one pretty long

617
00:29:12,240 --> 00:29:15,390
somewhat random very unguessable
password that you just

618
00:29:15,390 --> 00:29:17,070
promise to commit to memory--

619
00:29:17,070 --> 00:29:19,140
and maybe for backup, you
literally print it out

620
00:29:19,140 --> 00:29:21,180
and put it in a safe
deposit box or a safe,

621
00:29:21,180 --> 00:29:23,490
or just hide it somewhere
physically that there's

622
00:29:23,490 --> 00:29:25,350
very low probability
someone's going to find

623
00:29:25,350 --> 00:29:27,630
the backup copy, that might be alone.

624
00:29:27,630 --> 00:29:31,980
But of course, the flip side is now
if you forget that primary password,

625
00:29:31,980 --> 00:29:34,590
you've now lost all of
the eggs in the basket.

626
00:29:34,590 --> 00:29:38,830
If someone gets that primary password,
now they have access to everything.

627
00:29:38,830 --> 00:29:40,350
So that's rather the tradeoff.

628
00:29:40,350 --> 00:29:45,030
But I dare say you're probably less
threatened, depending on your family,

629
00:29:45,030 --> 00:29:48,090
by the people immediately
around you than the billions

630
00:29:48,090 --> 00:29:51,420
of other people on the internet
that have access, potentially,

631
00:29:51,420 --> 00:29:52,648
to those same systems.

632
00:29:52,648 --> 00:29:53,940
So there, too, it's a tradeoff.

633
00:29:53,940 --> 00:29:55,732
But it's up to you to
decide whether or not

634
00:29:55,732 --> 00:29:57,880
to manage your passwords in this way.

635
00:29:57,880 --> 00:30:00,600
But if you were on that top 10
list, or even if you're not,

636
00:30:00,600 --> 00:30:04,110
but you can think of several accounts
that all have the same password,

637
00:30:04,110 --> 00:30:06,550
you're probably going to benefit
from something like this.

638
00:30:06,550 --> 00:30:11,310
And why is it bad, to be clear, to use
the same password on multiple sites

639
00:30:11,310 --> 00:30:15,780
in case that's never sort
of dawned in thought?

640
00:30:15,780 --> 00:30:18,480
Why is that a bad thing,
to reuse a password

641
00:30:18,480 --> 00:30:20,340
on different websites, different apps?

642
00:30:20,340 --> 00:30:21,270
Any intuition?

643
00:30:21,270 --> 00:30:22,050
Yeah, in the back?

644
00:30:22,050 --> 00:30:24,550
AUDIENCE: Once attacked,
it's easy to get to.

645
00:30:24,550 --> 00:30:25,550
DAVID J. MALAN: Exactly.

646
00:30:25,550 --> 00:30:27,170
Once it's attacked, you can--

647
00:30:27,170 --> 00:30:29,270
the adversary, presumably,
by transitivity,

648
00:30:29,270 --> 00:30:31,880
can see, oh, well, if
this user's username is

649
00:30:31,880 --> 00:30:35,030
malan@harvard.edu on this website,
and their password is foolishly

650
00:30:35,030 --> 00:30:38,840
123456 or even something
way more complicated,

651
00:30:38,840 --> 00:30:41,060
they can probably just
assume with high probability

652
00:30:41,060 --> 00:30:42,860
that if I'm being a
little reckless, let's

653
00:30:42,860 --> 00:30:47,210
try accessing malan@harvard.edu's
use other accounts, other apps using

654
00:30:47,210 --> 00:30:48,358
that exact same password.

655
00:30:48,358 --> 00:30:50,150
And so by transitivity,
essentially, you're

656
00:30:50,150 --> 00:30:52,530
putting your other accounts at risk.

657
00:30:52,530 --> 00:30:53,780
So what's maybe a takeaway?

658
00:30:53,780 --> 00:30:56,780
Minimally here, I would start
to reconsider your passcodes

659
00:30:56,780 --> 00:30:58,020
on your most important data.

660
00:30:58,020 --> 00:30:58,730
Maybe it's medical.

661
00:30:58,730 --> 00:30:59,600
Maybe it's financial.

662
00:30:59,600 --> 00:31:00,308
Maybe it's email.

663
00:31:00,308 --> 00:31:04,070
Anything remotely personal that you
really wouldn't want to have access.

664
00:31:04,070 --> 00:31:06,890
Do you necessarily need
the same level of security

665
00:31:06,890 --> 00:31:10,130
on e-commerce sites or sites
that you don't really care about

666
00:31:10,130 --> 00:31:13,080
or that you signed up for once
and after that, that's it?

667
00:31:13,080 --> 00:31:13,950
Probably not.

668
00:31:13,950 --> 00:31:15,960
So you can decide for
yourself, but again,

669
00:31:15,960 --> 00:31:17,570
software, like a password manager.

670
00:31:17,570 --> 00:31:19,570
And these are just some
of the possibilities out

671
00:31:19,570 --> 00:31:21,965
there are probably to be your friend.

672
00:31:21,965 --> 00:31:23,090
A couple of these are free.

673
00:31:23,090 --> 00:31:24,550
They come with Windows or Mac OS.

674
00:31:24,550 --> 00:31:25,550
A couple are commercial.

675
00:31:25,550 --> 00:31:29,070
Harvard has a site license for
students for one of these as well.

676
00:31:29,070 --> 00:31:30,770
So there are options out there.

677
00:31:30,770 --> 00:31:32,420
But what else do people use?

678
00:31:32,420 --> 00:31:35,060
What else can people use to
keep their systems secure?

679
00:31:35,060 --> 00:31:39,590
So most of us nowadays have probably
heard of encryption, this technique

680
00:31:39,590 --> 00:31:41,400
for just scrambling information.

681
00:31:41,400 --> 00:31:45,200
So when you want to send a message,
an email, or upload a photograph,

682
00:31:45,200 --> 00:31:49,085
or use your credit card, hopefully, it's
not just being sent out for all to see,

683
00:31:49,085 --> 00:31:50,960
but there's some kind
of scrambling going on.

684
00:31:50,960 --> 00:31:53,600
And some fancy mathematics
ensure that encryption

685
00:31:53,600 --> 00:31:57,200
ensures that only you, the sender,
and someone else, the receiver,

686
00:31:57,200 --> 00:32:01,580
can theoretically see what that credit
card number is, what that message is

687
00:32:01,580 --> 00:32:03,330
what that photograph is instead.

688
00:32:03,330 --> 00:32:06,200
So encryption is sort
of commonplace nowadays,

689
00:32:06,200 --> 00:32:09,555
both in websites, and apps, and
ATMs, and other such devices.

690
00:32:09,555 --> 00:32:10,430
But how does it work?

691
00:32:10,430 --> 00:32:13,238
Well, back in week two
of CS50, your child

692
00:32:13,238 --> 00:32:15,530
learned a little something
about encryption, otherwise,

693
00:32:15,530 --> 00:32:16,640
known as cryptography.

694
00:32:16,640 --> 00:32:18,770
And one of the algorithms
we talked about

695
00:32:18,770 --> 00:32:21,160
was quite simply something like this.

696
00:32:21,160 --> 00:32:25,952
This is what we might call, not only
CS50, but plain text, so very plain

697
00:32:25,952 --> 00:32:28,910
text that, in this case, is English
and obviously, everyone in the room

698
00:32:28,910 --> 00:32:29,670
can read it.

699
00:32:29,670 --> 00:32:32,600
But what if I wanted to send this
message out to someone in this room,

700
00:32:32,600 --> 00:32:35,720
or out on the internet, or maybe
equivalently back in the day,

701
00:32:35,720 --> 00:32:38,690
maybe write a message down on a
scrap of paper in grade school

702
00:32:38,690 --> 00:32:42,560
and pass a secret note, a secret love
note to someone in class with hopes

703
00:32:42,560 --> 00:32:44,990
that the teacher or any
other students in the class

704
00:32:44,990 --> 00:32:46,400
can't intercept it and read it?

705
00:32:46,400 --> 00:32:49,880
Well, you probably don't want to
say, this is CS50, or I love you,

706
00:32:49,880 --> 00:32:52,280
or anything remotely sensitive.

707
00:32:52,280 --> 00:32:54,290
But rather, maybe you
want to encrypt it.

708
00:32:54,290 --> 00:32:56,510
And let's change the T to a U.

709
00:32:56,510 --> 00:33:01,430
Maybe change the H to an I, the I
to a J, the S to a T, the I to a J,

710
00:33:01,430 --> 00:33:04,580
the S to a T again, the
C to a D, the S to a T.

711
00:33:04,580 --> 00:33:06,440
And we'll just leave
the numbers, alone even

712
00:33:06,440 --> 00:33:09,330
though I worry someone could probably
guess what this now does say,

713
00:33:09,330 --> 00:33:10,070
nonetheless.

714
00:33:10,070 --> 00:33:13,820
But what was the algorithm as
I rattled those changes off,

715
00:33:13,820 --> 00:33:17,715
whether a student from week
two or parents from week now?

716
00:33:17,715 --> 00:33:18,215
Yeah?

717
00:33:18,215 --> 00:33:19,730
AUDIENCE: It's a one-letter shift.

718
00:33:19,730 --> 00:33:21,397
DAVID J. MALAN: Just a one-letter shift.

719
00:33:21,397 --> 00:33:25,940
And this is more sophisticated called
a rotational cipher or a Caesar cipher

720
00:33:25,940 --> 00:33:27,320
after Caesar back in the day.

721
00:33:27,320 --> 00:33:28,760
It's relatively simplistic.

722
00:33:28,760 --> 00:33:31,040
But back in the day,
it's not so simplistic

723
00:33:31,040 --> 00:33:34,430
if you're the first person in the
world to ever use it or think of it.

724
00:33:34,430 --> 00:33:36,500
But nowadays, this is
not actually what we use.

725
00:33:36,500 --> 00:33:38,540
But it's similarly
mathematical in nature.

726
00:33:38,540 --> 00:33:41,660
It's not quite as simple as just
adding one or subtracting one

727
00:33:41,660 --> 00:33:44,810
to go from now what we call
ciphertext to plain text.

728
00:33:44,810 --> 00:33:46,502
But it's similarly math that's involved.

729
00:33:46,502 --> 00:33:48,710
And let me just stipulate
that the way the math works

730
00:33:48,710 --> 00:33:51,230
is that the sender and
the receiver just have

731
00:33:51,230 --> 00:33:53,390
to have in mind some kind of secret.

732
00:33:53,390 --> 00:33:56,360
And the secret in this case
would very trivially be one,

733
00:33:56,360 --> 00:33:59,270
but it could be a much bigger,
much more unguessable number,

734
00:33:59,270 --> 00:34:01,820
or maybe some other secret
we share, the presumption

735
00:34:01,820 --> 00:34:05,390
being that my classmates, my teacher
in that grade school classroom, if they

736
00:34:05,390 --> 00:34:07,760
don't know what that secret
is that number is, yeah,

737
00:34:07,760 --> 00:34:11,659
they could try to brute force it and
try all possible mathematics, plus one,

738
00:34:11,659 --> 00:34:12,770
plus two, plus three.

739
00:34:12,770 --> 00:34:14,120
But that's going to take them some time.

740
00:34:14,120 --> 00:34:15,630
And they probably don't care enough.

741
00:34:15,630 --> 00:34:18,750
And so my data might be,
therefore, relatively secure.

742
00:34:18,750 --> 00:34:21,150
But we use encryption
all the time nowadays.

743
00:34:21,150 --> 00:34:24,050
And so for instance, this
is at the start of most URLs

744
00:34:24,050 --> 00:34:26,179
nowadays, even if you
don't type it yourself.

745
00:34:26,179 --> 00:34:30,530
With that said, Safari and even Chrome
now are kind of simplifying, if not,

746
00:34:30,530 --> 00:34:33,590
dumbing down user interfaces
to just hide details

747
00:34:33,590 --> 00:34:37,370
that you and I, as normal
users, don't need to see 24/7.

748
00:34:37,370 --> 00:34:38,150
But it is there.

749
00:34:38,150 --> 00:34:41,449
And if in fact, on your phone
or laptop, you click on the URL,

750
00:34:41,449 --> 00:34:43,370
even if it's super
short initially, you'll

751
00:34:43,370 --> 00:34:46,040
probably see the whole
thing starting with this.

752
00:34:46,040 --> 00:34:47,570
And the S means secure.

753
00:34:47,570 --> 00:34:49,550
The S means that
encryption is being used.

754
00:34:49,550 --> 00:34:52,520
But there's other forms of this,
not just when you visit websites.

755
00:34:52,520 --> 00:34:55,159
There's this, end-to-end
encryption, which

756
00:34:55,159 --> 00:34:59,120
is being talked about more nowadays,
especially during COVID times with so

757
00:34:59,120 --> 00:35:02,900
many more of us on video and
talking about more sensitive things,

758
00:35:02,900 --> 00:35:05,390
telemedicine, talking to
doctors, things that you also

759
00:35:05,390 --> 00:35:09,530
wouldn't want to verbally or visually
get out into the wild just like text.

760
00:35:09,530 --> 00:35:13,700
What's different about end-to-end
encryption versus HTTPS

761
00:35:13,700 --> 00:35:18,410
and the type of encryption that most
of us use every day on websites alone?

762
00:35:18,410 --> 00:35:20,720
End-to-end encryption is
sort of a better feature

763
00:35:20,720 --> 00:35:25,150
that you want to increasingly seek when
using services like Zoom, or Microsoft

764
00:35:25,150 --> 00:35:29,170
Teams, or WhatsApp, or the like.

765
00:35:29,170 --> 00:35:31,270
Any instincts here?

766
00:35:31,270 --> 00:35:34,905
Yeah, over on the right.

767
00:35:34,905 --> 00:35:38,575
AUDIENCE: The encryption happens
in the source and destination.

768
00:35:38,575 --> 00:35:39,450
DAVID J. MALAN: Good.

769
00:35:39,450 --> 00:35:41,492
So the encryption, the
scrambling of information,

770
00:35:41,492 --> 00:35:44,790
happens in the source the sender,
and the destination, the receiver,

771
00:35:44,790 --> 00:35:47,140
without a so-called
middleman in between.

772
00:35:47,140 --> 00:35:49,530
And this is actually very
different from most contexts

773
00:35:49,530 --> 00:35:53,310
nowadays that use just HTTPS
because when you're using HTTPS

774
00:35:53,310 --> 00:35:57,190
to buy something on Amazon securely
with your credit card, well, of course,

775
00:35:57,190 --> 00:36:00,220
Amazon needs to be able to decrypt
the message at the end of the day.

776
00:36:00,220 --> 00:36:01,080
And so that's fine.

777
00:36:01,080 --> 00:36:05,820
But even when you're using services
like video conferencing or maybe text

778
00:36:05,820 --> 00:36:09,060
messaging nowadays-- well, if you're
using WhatsApp, that's owned by Meta.

779
00:36:09,060 --> 00:36:11,580
And if you're using Instagram,
that's owned by Meta.

780
00:36:11,580 --> 00:36:14,920
There's a lot of middlemen in
these apps that we're using.

781
00:36:14,920 --> 00:36:17,580
And if they were only
using encryption period

782
00:36:17,580 --> 00:36:21,660
or only using something like
HTTPS, yes, your connection

783
00:36:21,660 --> 00:36:24,630
from you to WhatsApp, and
in turn, to the recipient

784
00:36:24,630 --> 00:36:27,750
might very well be secure
on each end of that channel.

785
00:36:27,750 --> 00:36:31,680
But Meta in between the company
and any other company in between

786
00:36:31,680 --> 00:36:34,988
could theoretically, for better or
for worse, be looking at that data,

787
00:36:34,988 --> 00:36:37,030
whether it's to mine it
for advertising purposes,

788
00:36:37,030 --> 00:36:39,120
whether it's to snoop on
data that you're sending.

789
00:36:39,120 --> 00:36:43,140
That is not end-to-end encryption if
the middleman, a company, typically,

790
00:36:43,140 --> 00:36:45,270
has technically access to that data.

791
00:36:45,270 --> 00:36:49,080
Now, Zoom, and Microsoft Teams,
and WhatsApp, and iMessage,

792
00:36:49,080 --> 00:36:51,510
and other services with which
you're familiar increasingly

793
00:36:51,510 --> 00:36:53,670
are offering stronger
guarantees of encryption,

794
00:36:53,670 --> 00:36:58,500
whereby, it's indeed between parties
A and B and not the one in the middle.

795
00:36:58,500 --> 00:36:59,820
Now, there's downsides here.

796
00:36:59,820 --> 00:37:02,580
And you can actually see this
kind of functionality manifest

797
00:37:02,580 --> 00:37:03,750
in certain settings.

798
00:37:03,750 --> 00:37:07,620
For instance, besides iMessage,
which just does this for you

799
00:37:07,620 --> 00:37:10,890
on iPhones or Macs, besides
Zoom, you can actually

800
00:37:10,890 --> 00:37:13,210
fine tune these settings,
indeed, within Zoom itself.

801
00:37:13,210 --> 00:37:16,140
So here's a screenshot that I took
last night of just what the user

802
00:37:16,140 --> 00:37:19,770
interface looks like today to create
a new Zoom meeting with the latest

803
00:37:19,770 --> 00:37:21,060
version of Zoom software.

804
00:37:21,060 --> 00:37:24,870
And maybe unbeknownst to you, there
is a choice of buttons down here.

805
00:37:24,870 --> 00:37:28,870
And most likely, yours is, by
default, enhanced encryption,

806
00:37:28,870 --> 00:37:31,590
which is brilliant marketing speak
because it's just encryption.

807
00:37:31,590 --> 00:37:32,490
It's not enhanced.

808
00:37:32,490 --> 00:37:35,260
It actually ironically
means worse than this.

809
00:37:35,260 --> 00:37:37,873
But they want you using
it most likely, why?

810
00:37:37,873 --> 00:37:39,540
Well, it's a little easier to implement.

811
00:37:39,540 --> 00:37:41,970
It's a little less expensive
for them computationally.

812
00:37:41,970 --> 00:37:46,770
And to be fair, enhanced encryption
does scramble the data, but not in a way

813
00:37:46,770 --> 00:37:48,420
that Zoom can't see it.

814
00:37:48,420 --> 00:37:49,650
Zoom can, indeed, see it.

815
00:37:49,650 --> 00:37:51,600
But that's actually a
plus in some context

816
00:37:51,600 --> 00:37:54,090
because if you want
to do cloud recordings

817
00:37:54,090 --> 00:37:56,340
and you want a meeting
recorded not on your Mac or PC,

818
00:37:56,340 --> 00:37:57,810
but let Zoom deal with that.

819
00:37:57,810 --> 00:38:00,085
If you want automatic
transcription nowadays,

820
00:38:00,085 --> 00:38:02,460
so the words to appear, whether
it's English or something

821
00:38:02,460 --> 00:38:04,320
else on the screen,
well, you can't really

822
00:38:04,320 --> 00:38:06,810
lock Zoom or any other
middleman out of that

823
00:38:06,810 --> 00:38:08,910
because someone needs
to save it to the cloud.

824
00:38:08,910 --> 00:38:12,930
Someone needs to translate the voice
to those English or some other language

825
00:38:12,930 --> 00:38:13,480
words.

826
00:38:13,480 --> 00:38:15,660
So enhanced encryption
enables those features,

827
00:38:15,660 --> 00:38:18,630
but they also allow a bad
actor, malicious employees,

828
00:38:18,630 --> 00:38:21,690
someone who's just nosey at Zoom
or the equivalent middle man

829
00:38:21,690 --> 00:38:25,080
to just poke around your video
conference and hear what you've said

830
00:38:25,080 --> 00:38:29,770
or see what you've typed as well unless
you instead check this box as well.

831
00:38:29,770 --> 00:38:33,033
So increasingly look for mentions
of end-to-end encryption.

832
00:38:33,033 --> 00:38:35,700
Or give that some thought when
you choose a technology via which

833
00:38:35,700 --> 00:38:38,760
to communicate with someone,
whether it's within your family

834
00:38:38,760 --> 00:38:40,870
or without as well.

835
00:38:40,870 --> 00:38:45,150
Now, last, but not least, there's
other applications of encryption too.

836
00:38:45,150 --> 00:38:49,600
And this, too, might be a lesson
learned as well, full disk encryption.

837
00:38:49,600 --> 00:38:53,640
So a disk is where your data is stored
in your Mac, or PC, or even your phone.

838
00:38:53,640 --> 00:38:56,940
And full disk encryption just
means ideally that all of your data

839
00:38:56,940 --> 00:38:59,190
is encrypted that is somehow scrambled.

840
00:38:59,190 --> 00:39:02,100
Now, hopefully, your password
for your computer or phone

841
00:39:02,100 --> 00:39:04,952
is good enough so that
even though the device is

842
00:39:04,952 --> 00:39:07,410
encrypted with that password,
at least, you'll remember it.

843
00:39:07,410 --> 00:39:10,458
And your phone or your Mac or PC will
automatically decrypt it for you.

844
00:39:10,458 --> 00:39:13,500
Of course, you can't scramble the
information and hide it from ourselves.

845
00:39:13,500 --> 00:39:16,470
One of us, at least, for these
devices needs to have access.

846
00:39:16,470 --> 00:39:18,990
But full disk encryption
typically means that at least when

847
00:39:18,990 --> 00:39:22,110
you close the laptop lid or
power down for the night,

848
00:39:22,110 --> 00:39:24,690
that even if someone
else steals that device,

849
00:39:24,690 --> 00:39:28,620
opens the lid, unless
they have your passcode,

850
00:39:28,620 --> 00:39:31,470
they can't even plug in
fancy cables to the device

851
00:39:31,470 --> 00:39:34,320
and just rip the zeros
and ones off of the device

852
00:39:34,320 --> 00:39:35,823
and see what's actually there.

853
00:39:35,823 --> 00:39:37,740
Full disk encryption
means they could do that,

854
00:39:37,740 --> 00:39:40,890
but they would just see
seemingly random zeros and ones.

855
00:39:40,890 --> 00:39:42,270
Now, there's a downside here too.

856
00:39:42,270 --> 00:39:44,160
This might slow things down potentially.

857
00:39:44,160 --> 00:39:48,060
But it is a feature increasingly that's
offered and is absolutely something you

858
00:39:48,060 --> 00:39:51,510
should consider enabling in
general, especially if your laptop

859
00:39:51,510 --> 00:39:52,620
or phone travels with you.

860
00:39:52,620 --> 00:39:53,970
And certainly, your phone does.

861
00:39:53,970 --> 00:39:58,020
Or if you plan to donate, or
sell, or give away a device,

862
00:39:58,020 --> 00:40:00,180
you don't want to leave
all of the zeros and ones,

863
00:40:00,180 --> 00:40:02,650
the remnants of your own
sensitive data, passed on there.

864
00:40:02,650 --> 00:40:04,770
So Windows has a feature
called BitLocker.

865
00:40:04,770 --> 00:40:08,320
Mac OS has a feature called FileVault.
There's commercial options as well.

866
00:40:08,320 --> 00:40:12,210
But generally, we're at the point
now in 2022 where clicking a button

867
00:40:12,210 --> 00:40:14,940
is sufficient to enable these features.

868
00:40:14,940 --> 00:40:17,430
With that said, don't rush
into all of these decisions.

869
00:40:17,430 --> 00:40:19,440
I would make backups of your data.

870
00:40:19,440 --> 00:40:22,930
And don't maybe email CS50 if
something goes wrong with that process.

871
00:40:22,930 --> 00:40:24,570
But I would do your own due diligence.

872
00:40:24,570 --> 00:40:27,720
But this, too, would be
a menu of possibilities.

873
00:40:27,720 --> 00:40:32,070
And now, the bad side, the
downside, of what seems to be great,

874
00:40:32,070 --> 00:40:33,990
this notion of full disk encryption.

875
00:40:33,990 --> 00:40:36,990
Unfortunately, just as
we can encrypt our data

876
00:40:36,990 --> 00:40:39,280
to protect it from the
adversaries, so can

877
00:40:39,280 --> 00:40:44,140
the adversaries if they get into our
devices, encrypt our data and do what?

878
00:40:44,140 --> 00:40:46,420
Not tell us that secret key.

879
00:40:46,420 --> 00:40:49,190
And so this is generally applied
in the context of ransomware,

880
00:40:49,190 --> 00:40:51,100
which tragically, you
increasingly hear about

881
00:40:51,100 --> 00:40:53,992
in hospital systems, school
systems, municipalities

882
00:40:53,992 --> 00:40:55,450
where systems are getting attacked.

883
00:40:55,450 --> 00:40:57,730
And the data is not just
getting stolen because what

884
00:40:57,730 --> 00:41:01,570
is the adversary typically need with
local municipal or even hospital data?

885
00:41:01,570 --> 00:41:05,200
The value to the adversary
is encrypting all

886
00:41:05,200 --> 00:41:07,660
of the hospital, all of
the municipality's data

887
00:41:07,660 --> 00:41:11,298
preventing them from accessing it
if they have no backups or the like.

888
00:41:11,298 --> 00:41:13,090
And so ransomware is
literally about trying

889
00:41:13,090 --> 00:41:16,550
to convince someone to pay you money or
pay you Bitcoin or something like that

890
00:41:16,550 --> 00:41:18,490
to give you that secret key.

891
00:41:18,490 --> 00:41:21,910
And the key, in this case, is surely
more sophisticated than the number one.

892
00:41:21,910 --> 00:41:23,420
But it's really the same idea.

893
00:41:23,420 --> 00:41:26,380
So here too, yet again, a tradeoff
just as we sort of invent something

894
00:41:26,380 --> 00:41:30,350
for good, it can also be used
for evil in so to speak as well.

895
00:41:30,350 --> 00:41:32,890
But it's really the same
underlying principles,

896
00:41:32,890 --> 00:41:34,600
even though we keep
seeing it and hearing

897
00:41:34,600 --> 00:41:37,490
about it, in these different forms.

898
00:41:37,490 --> 00:41:40,360
And lastly, if only because
folks are generally familiar,

899
00:41:40,360 --> 00:41:43,750
but don't necessarily know what
it is that it's doing for them,

900
00:41:43,750 --> 00:41:46,810
browsers nowadays have what's
often called incognito mode

901
00:41:46,810 --> 00:41:49,750
or private mode, which has nothing
to do with encryption, but does

902
00:41:49,750 --> 00:41:52,720
have to do with cybersecurity,
or really, cyber privacy,

903
00:41:52,720 --> 00:41:55,450
keeping your data from prying eyes.

904
00:41:55,450 --> 00:41:57,742
Incognito mode, if you open
it in Chrome, for instance,

905
00:41:57,742 --> 00:41:59,200
looks a little something like this.

906
00:41:59,200 --> 00:42:02,080
And we use it in CS50 when introducing
students, as we did last week,

907
00:42:02,080 --> 00:42:04,780
to web programming because
it, in effect, lets

908
00:42:04,780 --> 00:42:07,690
you start with a clean slate like
a brand new browser that has never

909
00:42:07,690 --> 00:42:10,720
visited any websites before, which
is good for just diagnosing problems.

910
00:42:10,720 --> 00:42:14,512
But it's often commonly used if you want
to log into maybe your Gmail account

911
00:42:14,512 --> 00:42:17,470
on someone else's computer and you
don't want your password being saved

912
00:42:17,470 --> 00:42:19,262
or you want to visit
some website where you

913
00:42:19,262 --> 00:42:22,610
don't want the URL or the search terms
ending up in your autocomplete history.

914
00:42:22,610 --> 00:42:25,090
So there's multiple
uses for incognito mode.

915
00:42:25,090 --> 00:42:26,870
But what does it really do?

916
00:42:26,870 --> 00:42:30,310
Well, it doesn't stop your company,
it doesn't stop your university,

917
00:42:30,310 --> 00:42:33,040
your internet service provider,
be it, Comcast, Verizon,

918
00:42:33,040 --> 00:42:36,730
or the like from knowing what
websites you go to because--

919
00:42:36,730 --> 00:42:37,510
ask your students.

920
00:42:37,510 --> 00:42:39,920
A couple of weeks ago we talked
about-- actually, a week ago,

921
00:42:39,920 --> 00:42:41,620
we talked about how the internet works.

922
00:42:41,620 --> 00:42:44,000
And unfortunately, every
computer has an IP address,

923
00:42:44,000 --> 00:42:46,375
which is a unique identifier,
which goes out any time you

924
00:42:46,375 --> 00:42:48,462
go anywhere incognito mode or not.

925
00:42:48,462 --> 00:42:50,170
So this isn't really
covering your tracks

926
00:42:50,170 --> 00:42:54,130
outside of your office, or outside of
your home, or outside of your company.

927
00:42:54,130 --> 00:42:56,950
But it is, at least, throwing
away local information.

928
00:42:56,950 --> 00:43:00,580
And so we'll talk, in fact, in
CS50's week nine this coming Monday

929
00:43:00,580 --> 00:43:03,692
about cookies, which you
might generally know about

930
00:43:03,692 --> 00:43:04,900
and what are called sessions.

931
00:43:04,900 --> 00:43:07,900
And so long story short,
what incognito mode does

932
00:43:07,900 --> 00:43:12,225
is it throws away, when you close the
window, any locally stored information,

933
00:43:12,225 --> 00:43:15,100
so these things called cookies,
which are sort of virtual hand stamps

934
00:43:15,100 --> 00:43:16,892
that just remember what
you've logged in as

935
00:43:16,892 --> 00:43:19,030
or what's in your
shopping cart or the like.

936
00:43:19,030 --> 00:43:23,410
But it doesn't hide any information from
anyone outside of your own Mac or PC.

937
00:43:23,410 --> 00:43:25,715
It only prevents those
local prying eyes.

938
00:43:25,715 --> 00:43:28,090
So there, too, even though we
have tools that many of you

939
00:43:28,090 --> 00:43:30,310
are probably in the habit
of using or thinking

940
00:43:30,310 --> 00:43:34,360
you should use to be more private, be
more secure on the internet, what we do

941
00:43:34,360 --> 00:43:37,720
really in CS50, both
weeks past and future,

942
00:43:37,720 --> 00:43:40,220
is talk about how these
technologies work,

943
00:43:40,220 --> 00:43:44,260
so that ultimately, we have all the
more of an educated citizenry here

944
00:43:44,260 --> 00:43:46,800
among undergrads and
here as well as online,

945
00:43:46,800 --> 00:43:49,300
so that you can apply these
same lessons learned to problems

946
00:43:49,300 --> 00:43:50,900
you'll encounter in the future.

947
00:43:50,900 --> 00:43:52,600
So as promised, the homework.

948
00:43:52,600 --> 00:43:55,143
One, you should probably
use a password manager.

949
00:43:55,143 --> 00:43:57,310
It doesn't have to be one
of those ones on the list,

950
00:43:57,310 --> 00:44:00,640
but at least starting that conversation
maybe with someone who does,

951
00:44:00,640 --> 00:44:01,570
maybe the--

952
00:44:01,570 --> 00:44:04,748
it's often the students
in your family, perhaps,

953
00:44:04,748 --> 00:44:06,790
who can advise you on some
of these technologies.

954
00:44:06,790 --> 00:44:10,653
Consider using a password manager
too, using two-factor authentication,

955
00:44:10,653 --> 00:44:12,820
whether it's your phone or
some key fob or the like,

956
00:44:12,820 --> 00:44:16,030
but at least seeking out that feature
at least for accounts that you really

957
00:44:16,030 --> 00:44:18,760
care about, your email, social
media, financial, medical,

958
00:44:18,760 --> 00:44:21,010
anything where you'd be
embarrassed, at best,

959
00:44:21,010 --> 00:44:24,400
or really violated, at worst, if
that kind of information got out

960
00:44:24,400 --> 00:44:28,360
and then increasingly using not just
encryption, which you get automatically

961
00:44:28,360 --> 00:44:31,930
for most technologies today, but
increasingly choosing technologies

962
00:44:31,930 --> 00:44:35,980
that offer stronger guarantees that keep
those middlemen, those companies out

963
00:44:35,980 --> 00:44:39,040
of the way if only so that you
can trust with higher probability

964
00:44:39,040 --> 00:44:43,900
that only party B knows what
party A has said or sent.

965
00:44:43,900 --> 00:44:45,820
Now, this, of course,
was a whirlwind tour.

966
00:44:45,820 --> 00:44:48,190
There's so much more
that you can do online.

967
00:44:48,190 --> 00:44:52,150
Indeed, this course, CS50, can be taken
for free online via platforms like edX

968
00:44:52,150 --> 00:44:54,430
at edx.org/cs50.

969
00:44:54,430 --> 00:44:57,400
I thought it might be
appropriate to end on this note.

970
00:44:57,400 --> 00:45:01,300
If anyone would like
to conjecture before we

971
00:45:01,300 --> 00:45:07,050
start playing music and adjourn for
lunch, what our final message here is.

972
00:45:07,050 --> 00:45:12,270
If we reverse the plus one and maybe
start minus one here, minus one here--

973
00:45:12,270 --> 00:45:14,920
and indeed, thank you
so much for coming.

974
00:45:14,920 --> 00:45:18,240
This was CS50.

975
00:45:18,240 --> 00:45:21,290
[MUSIC PLAYING]

976
00:45:21,290 --> 00:45:55,000