1
00:00:00,320 --> 00:00:06,480
here we go hey welcome back horncon

2
00:00:03,120 --> 00:00:08,960
thanks for uh listening to us uh shoot

3
00:00:06,480 --> 00:00:10,559
uh the crap right now you have dr mark

4
00:00:08,960 --> 00:00:11,840
carney he's a technical specialist in

5
00:00:10,559 --> 00:00:14,399
research and he's going to give you a

6
00:00:11,840 --> 00:00:18,240
great talk on tears for quantum fierce

7
00:00:14,400 --> 00:00:19,920
uh see you guys after hello hello hello

8
00:00:18,240 --> 00:00:22,159
thank you very much so yes my name is

9
00:00:19,920 --> 00:00:24,640
mark carney bank of england

10
00:00:22,160 --> 00:00:25,599
uh i am very glad to be presenting a

11
00:00:24,640 --> 00:00:28,560
quick kind of uh

12
00:00:25,599 --> 00:00:29,519
overview of quantum and security so

13
00:00:28,560 --> 00:00:31,679
we're going to look at

14
00:00:29,519 --> 00:00:33,440
what quantum means and how it relates to

15
00:00:31,679 --> 00:00:34,160
security what do you need to know the

16
00:00:33,440 --> 00:00:35,839
idea is

17
00:00:34,160 --> 00:00:38,000
here not to give a condensed lecture

18
00:00:35,840 --> 00:00:41,040
series with just slides of mathematics

19
00:00:38,000 --> 00:00:42,719
there is one math slide sorry no

20
00:00:41,040 --> 00:00:44,480
actually no i'm not sorry

21
00:00:42,719 --> 00:00:45,840
you know just i'm sorry for my lack of

22
00:00:44,480 --> 00:00:47,680
apology there you go

23
00:00:45,840 --> 00:00:49,520
um what i'm gonna do is i'm gonna sort

24
00:00:47,680 --> 00:00:51,760
of present uh today uh this this

25
00:00:49,520 --> 00:00:54,000
particular talk now with added kittens

26
00:00:51,760 --> 00:00:55,199
um so we're gonna sort of go through

27
00:00:54,000 --> 00:00:56,480
what is quantum and then we're gonna

28
00:00:55,199 --> 00:00:58,399
look at the way in which it's gonna be

29
00:00:56,480 --> 00:01:00,480
affecting security and it's kind of like

30
00:00:58,399 --> 00:01:01,440
if you listen to quantum people too much

31
00:01:00,480 --> 00:01:04,399
they'll say things like

32
00:01:01,440 --> 00:01:05,840
the wind is coming it is already blowing

33
00:01:04,400 --> 00:01:06,799
and that's kind of like yeah that's not

34
00:01:05,840 --> 00:01:08,799
very helpful

35
00:01:06,799 --> 00:01:10,320
um what we need is actual facts and

36
00:01:08,799 --> 00:01:11,439
figures and a way of looking at this

37
00:01:10,320 --> 00:01:12,320
quite critically so that's what we're

38
00:01:11,439 --> 00:01:15,360
going to do today

39
00:01:12,320 --> 00:01:16,960
just facts resources links there's qr

40
00:01:15,360 --> 00:01:19,119
codes so if you have a mobile phone or a

41
00:01:16,960 --> 00:01:21,039
tablet get that ready because that'll be

42
00:01:19,119 --> 00:01:24,320
best to view some of the links so you

43
00:01:21,040 --> 00:01:27,119
know what is a quantum anyway

44
00:01:24,320 --> 00:01:27,919
if you ask google you'll hear one of a

45
00:01:27,119 --> 00:01:29,360
few things

46
00:01:27,920 --> 00:01:31,439
okay you'll find that quantum is

47
00:01:29,360 --> 00:01:34,159
dishwasher powder finish

48
00:01:31,439 --> 00:01:34,639
powerball quantum ultimate sounds very

49
00:01:34,159 --> 00:01:36,479
very

50
00:01:34,640 --> 00:01:38,159
sort of feeling so powerful you know the

51
00:01:36,479 --> 00:01:41,759
quantum ultimate

52
00:01:38,159 --> 00:01:42,720
um it's also beer the quantum state is a

53
00:01:41,759 --> 00:01:44,640
4.2

54
00:01:42,720 --> 00:01:46,960
section ipa i think they sell this at

55
00:01:44,640 --> 00:01:49,600
asda which is our version of walmart

56
00:01:46,960 --> 00:01:51,919
um then there's sure for men motion

57
00:01:49,600 --> 00:01:53,759
sense quantum dry

58
00:01:51,920 --> 00:01:56,000
for your man who's a photon in the

59
00:01:53,759 --> 00:01:58,159
streets and a boson in the sheets

60
00:01:56,000 --> 00:01:59,840
um and that might lead to needing a

61
00:01:58,159 --> 00:02:02,240
quantum buggy

62
00:01:59,840 --> 00:02:03,119
quantum pram yes genuinely there are

63
00:02:02,240 --> 00:02:07,119
prams

64
00:02:03,119 --> 00:02:10,399
marketed as quantum and my favorite

65
00:02:07,119 --> 00:02:13,840
is toilet roll bog roll

66
00:02:10,399 --> 00:02:15,760
lou roll toilet paper um

67
00:02:13,840 --> 00:02:17,120
yeah seriously quantum is everywhere

68
00:02:15,760 --> 00:02:19,200
it's a bit of a buzz word what does it

69
00:02:17,120 --> 00:02:20,879
actually mean so quantum means just this

70
00:02:19,200 --> 00:02:22,560
it's the smallest finite piece of

71
00:02:20,879 --> 00:02:24,879
something i did check

72
00:02:22,560 --> 00:02:26,959
quantum toilet roll they do not sell you

73
00:02:24,879 --> 00:02:28,720
the smallest finite piece of toilet roll

74
00:02:26,959 --> 00:02:30,800
don't ask how i checked that experiment

75
00:02:28,720 --> 00:02:32,239
is between me and me alone

76
00:02:30,800 --> 00:02:33,440
so quantum effects are what you're

77
00:02:32,239 --> 00:02:35,120
seeing when someone says we're going to

78
00:02:33,440 --> 00:02:36,560
use a quantum technology

79
00:02:35,120 --> 00:02:38,160
they're talking about quantum effects

80
00:02:36,560 --> 00:02:38,879
and there's generally two that they're

81
00:02:38,160 --> 00:02:42,640
talking about

82
00:02:38,879 --> 00:02:45,200
okay superposition and entanglement

83
00:02:42,640 --> 00:02:46,879
now superposition is when a particle is

84
00:02:45,200 --> 00:02:48,000
in multiple states simultaneously and

85
00:02:46,879 --> 00:02:49,120
i'm going to give you kind of a sense of

86
00:02:48,000 --> 00:02:51,519
what that really means

87
00:02:49,120 --> 00:02:52,239
okay and then entanglement has a lot of

88
00:02:51,519 --> 00:02:55,120
quantum

89
00:02:52,239 --> 00:02:56,400
woo about it but realistically two

90
00:02:55,120 --> 00:02:58,879
particles can

91
00:02:56,400 --> 00:03:00,480
in a sense become inseparable and

92
00:02:58,879 --> 00:03:01,920
actually the inner sense bits really

93
00:03:00,480 --> 00:03:04,000
important there there's a

94
00:03:01,920 --> 00:03:06,159
mathematical thing the the polynomials

95
00:03:04,000 --> 00:03:09,680
are inseparable for those who care

96
00:03:06,159 --> 00:03:10,799
um and they are linked in such a way

97
00:03:09,680 --> 00:03:12,480
that if you measure one

98
00:03:10,800 --> 00:03:13,920
you know some information about the

99
00:03:12,480 --> 00:03:15,440
other okay

100
00:03:13,920 --> 00:03:17,518
so that's all entanglement really means

101
00:03:15,440 --> 00:03:19,680
that's what superposition really means

102
00:03:17,519 --> 00:03:21,360
and they're kind of very very useful

103
00:03:19,680 --> 00:03:23,280
effects but they only really happen at

104
00:03:21,360 --> 00:03:25,519
the quantum level we just don't see them

105
00:03:23,280 --> 00:03:27,200
at the macro level where we live and

106
00:03:25,519 --> 00:03:30,640
breathe and talk to each other

107
00:03:27,200 --> 00:03:32,480
so we we i could go at length about what

108
00:03:30,640 --> 00:03:34,159
these are but i'm not going to i'm not

109
00:03:32,480 --> 00:03:34,560
really going to touch on entanglement at

110
00:03:34,159 --> 00:03:36,480
all

111
00:03:34,560 --> 00:03:38,000
superposition will become relevant but

112
00:03:36,480 --> 00:03:40,000
realistically when you hear

113
00:03:38,000 --> 00:03:41,440
quantum technology you're usually

114
00:03:40,000 --> 00:03:43,280
dealing with one of these two

115
00:03:41,440 --> 00:03:45,599
or possibly a similar technology which

116
00:03:43,280 --> 00:03:46,000
is kind of related to the like wave

117
00:03:45,599 --> 00:03:48,879
particle

118
00:03:46,000 --> 00:03:50,400
duality or something um so what i if

119
00:03:48,879 --> 00:03:52,798
there's only one slide you look at

120
00:03:50,400 --> 00:03:54,319
let it be this one okay uh there are

121
00:03:52,799 --> 00:03:54,959
four domains of effect that you're going

122
00:03:54,319 --> 00:03:56,798
to see

123
00:03:54,959 --> 00:03:57,920
for quantum when it comes to security

124
00:03:56,799 --> 00:03:59,360
the first thing you're going to see is

125
00:03:57,920 --> 00:04:01,679
quantum computing

126
00:03:59,360 --> 00:04:03,599
quantum computing does pose somewhat of

127
00:04:01,680 --> 00:04:05,920
a threat theoretically at least

128
00:04:03,599 --> 00:04:08,079
to classical cryptography then you're

129
00:04:05,920 --> 00:04:10,958
going to see post-quantum cryptography

130
00:04:08,080 --> 00:04:12,959
which is basically a cryptography that

131
00:04:10,959 --> 00:04:14,560
doesn't rely on the same hardness

132
00:04:12,959 --> 00:04:17,680
problem that regular cryptography

133
00:04:14,560 --> 00:04:19,199
has for the last 20 30 40 years

134
00:04:17,680 --> 00:04:21,040
and we're going to look why do we care

135
00:04:19,199 --> 00:04:22,720
why does nist care about it we're going

136
00:04:21,040 --> 00:04:23,440
to look at quantum key distribution

137
00:04:22,720 --> 00:04:26,560
which is where you

138
00:04:23,440 --> 00:04:28,080
use quantum effects to actually derive a

139
00:04:26,560 --> 00:04:29,360
key and do key exchange

140
00:04:28,080 --> 00:04:31,039
and lastly we're going to have a very

141
00:04:29,360 --> 00:04:32,320
quick look at some called some quantum

142
00:04:31,040 --> 00:04:34,560
algorithms

143
00:04:32,320 --> 00:04:36,080
to show how actually maybe we can use

144
00:04:34,560 --> 00:04:39,199
these like in our day-to-day

145
00:04:36,080 --> 00:04:41,120
uh work uh to gain benefit from quantum

146
00:04:39,199 --> 00:04:43,280
computers and their capabilities

147
00:04:41,120 --> 00:04:45,199
so um this is the first one if you want

148
00:04:43,280 --> 00:04:46,799
to have a look please scan that qr code

149
00:04:45,199 --> 00:04:47,680
i'll give people a second i will also

150
00:04:46,800 --> 00:04:49,520
put the link

151
00:04:47,680 --> 00:04:50,960
uh in the discord afterwards but

152
00:04:49,520 --> 00:04:51,840
someone's probably got a faster fingers

153
00:04:50,960 --> 00:04:54,080
than i have

154
00:04:51,840 --> 00:04:54,880
um so it's just a geogebra just to sort

155
00:04:54,080 --> 00:04:56,880
of demonstrate

156
00:04:54,880 --> 00:04:58,159
what's going on okay i'm going to assume

157
00:04:56,880 --> 00:05:01,360
you can see this if not i

158
00:04:58,160 --> 00:05:02,800
uh i guess quan con 2 will tell me um so

159
00:05:01,360 --> 00:05:06,160
this is how we usually think

160
00:05:02,800 --> 00:05:08,720
about uh uh bits

161
00:05:06,160 --> 00:05:09,440
okay you have either an up going towards

162
00:05:08,720 --> 00:05:11,520
a zero

163
00:05:09,440 --> 00:05:13,600
or you're in a one state and you're

164
00:05:11,520 --> 00:05:14,719
never in both you're always in one or

165
00:05:13,600 --> 00:05:16,800
the other okay

166
00:05:14,720 --> 00:05:18,639
so you have two states for one bit and

167
00:05:16,800 --> 00:05:19,199
then you have four states between two

168
00:05:18,639 --> 00:05:20,720
bits

169
00:05:19,199 --> 00:05:22,479
and then you have eight states between

170
00:05:20,720 --> 00:05:24,240
three bits and it just keeps multiplying

171
00:05:22,479 --> 00:05:27,680
up in powers of two from there

172
00:05:24,240 --> 00:05:30,240
right now that's absolutely fine um

173
00:05:27,680 --> 00:05:31,680
and we use that as inspiration uh when

174
00:05:30,240 --> 00:05:34,000
it comes to quantum bits

175
00:05:31,680 --> 00:05:35,919
so a qubit has a very similar structure

176
00:05:34,000 --> 00:05:38,160
now we change the notation you'll see

177
00:05:35,919 --> 00:05:40,000
things like this this is a zero with

178
00:05:38,160 --> 00:05:41,360
like a funny line and a right angle

179
00:05:40,000 --> 00:05:44,080
bracket here

180
00:05:41,360 --> 00:05:46,320
that's called a ket um the if it points

181
00:05:44,080 --> 00:05:49,599
the other way it's called a bra

182
00:05:46,320 --> 00:05:50,240
bra for those of you realizing yes

183
00:05:49,600 --> 00:05:52,080
that's why

184
00:05:50,240 --> 00:05:53,840
amazon called their quantum service

185
00:05:52,080 --> 00:05:56,159
bracket

186
00:05:53,840 --> 00:05:57,198
um so yeah if it's pointing down here

187
00:05:56,160 --> 00:06:00,000
it's going to be in the one

188
00:05:57,199 --> 00:06:02,000
state so the ket being state so there's

189
00:06:00,000 --> 00:06:04,639
a zero state and the one state

190
00:06:02,000 --> 00:06:05,280
and again we can be sort of either or

191
00:06:04,639 --> 00:06:07,280
okay

192
00:06:05,280 --> 00:06:09,679
or you know what we could be i don't

193
00:06:07,280 --> 00:06:11,679
know somewhere in a 3d plane

194
00:06:09,680 --> 00:06:12,960
okay now when you're in this 3d plane

195
00:06:11,680 --> 00:06:13,600
you can actually say well we're going to

196
00:06:12,960 --> 00:06:15,840
have

197
00:06:13,600 --> 00:06:17,360
um you know some other vector and some

198
00:06:15,840 --> 00:06:19,198
other points and actually there's a

199
00:06:17,360 --> 00:06:20,000
whole infinitive of points they line up

200
00:06:19,199 --> 00:06:22,080
on a sphere like

201
00:06:20,000 --> 00:06:23,680
this okay so when you're dealing with a

202
00:06:22,080 --> 00:06:24,400
qubit this is probably the best picture

203
00:06:23,680 --> 00:06:27,120
to have

204
00:06:24,400 --> 00:06:27,840
this point can be all the way up at zero

205
00:06:27,120 --> 00:06:31,440
okay

206
00:06:27,840 --> 00:06:33,919
or it can be all the way down in one

207
00:06:31,440 --> 00:06:35,840
but it can also be for example exactly

208
00:06:33,919 --> 00:06:38,318
halfway in between

209
00:06:35,840 --> 00:06:40,080
and this is superposition because it's

210
00:06:38,319 --> 00:06:40,880
neither in zero or one it's exactly

211
00:06:40,080 --> 00:06:42,400
halfway

212
00:06:40,880 --> 00:06:44,479
and that's superposition of these

213
00:06:42,400 --> 00:06:46,638
particular states okay

214
00:06:44,479 --> 00:06:48,000
so when you talk here about sort of the

215
00:06:46,639 --> 00:06:50,000
magic of qubits the magic

216
00:06:48,000 --> 00:06:51,599
of quantum this is the picture that i'd

217
00:06:50,000 --> 00:06:53,680
like you to have in your mind

218
00:06:51,599 --> 00:06:54,639
okay this is the kind of model that

219
00:06:53,680 --> 00:06:56,560
you're going to be using

220
00:06:54,639 --> 00:06:57,680
okay so you you don't just have zero or

221
00:06:56,560 --> 00:06:59,039
one you have

222
00:06:57,680 --> 00:07:01,039
a zero and a one and then you have this

223
00:06:59,039 --> 00:07:03,360
kind of sphere of possibilities

224
00:07:01,039 --> 00:07:04,479
uh that you can use on top of that okay

225
00:07:03,360 --> 00:07:06,319
now

226
00:07:04,479 --> 00:07:07,840
i have a play with that the judge bar is

227
00:07:06,319 --> 00:07:08,400
online i'll put the link in the discord

228
00:07:07,840 --> 00:07:11,440
if someone

229
00:07:08,400 --> 00:07:13,120
doesn't beat me to it um and there's

230
00:07:11,440 --> 00:07:14,479
there's the link again on the qr code if

231
00:07:13,120 --> 00:07:15,680
you want to be quick so now that we've

232
00:07:14,479 --> 00:07:16,000
gone through that that's kind of the

233
00:07:15,680 --> 00:07:17,840
overall

234
00:07:16,000 --> 00:07:19,520
model that's the fundamental kind of

235
00:07:17,840 --> 00:07:21,758
model that you deal with with a qubit

236
00:07:19,520 --> 00:07:23,120
okay um at least that's the one that

237
00:07:21,759 --> 00:07:25,120
most people talk about

238
00:07:23,120 --> 00:07:27,199
so we're going to have a very quick very

239
00:07:25,120 --> 00:07:29,039
fast quantum computation

240
00:07:27,199 --> 00:07:30,800
class zero we're not going to go into

241
00:07:29,039 --> 00:07:34,159
any details i'm just going to give you

242
00:07:30,800 --> 00:07:35,680
the shape for those who care of what the

243
00:07:34,160 --> 00:07:36,400
maths kind of looks like and it's very

244
00:07:35,680 --> 00:07:38,319
difficult to

245
00:07:36,400 --> 00:07:39,599
not have maths in a talk about quantum

246
00:07:38,319 --> 00:07:41,759
computing right

247
00:07:39,599 --> 00:07:43,039
because we haven't yet got the

248
00:07:41,759 --> 00:07:44,720
abstraction layers

249
00:07:43,039 --> 00:07:46,159
quantum computers are very close to

250
00:07:44,720 --> 00:07:47,599
zx80s okay

251
00:07:46,160 --> 00:07:49,599
you have to understand the very low

252
00:07:47,599 --> 00:07:50,479
level of what's going on in order to

253
00:07:49,599 --> 00:07:52,400
really program

254
00:07:50,479 --> 00:07:54,318
and make it make it do things so what's

255
00:07:52,400 --> 00:07:56,960
the model well a qubit

256
00:07:54,319 --> 00:07:58,240
a quantum bit is a computing unit that

257
00:07:56,960 --> 00:08:00,638
can be some combination

258
00:07:58,240 --> 00:08:01,360
of zero state and one state not just one

259
00:08:00,639 --> 00:08:03,520
or the other

260
00:08:01,360 --> 00:08:05,120
it could be somewhere in between as well

261
00:08:03,520 --> 00:08:06,560
mathematically it looks like this two

262
00:08:05,120 --> 00:08:08,639
lines of math that's all

263
00:08:06,560 --> 00:08:10,240
um so you get an alpha and beta these

264
00:08:08,639 --> 00:08:12,479
are complex numbers

265
00:08:10,240 --> 00:08:14,240
so when you saw this sphere earlier what

266
00:08:12,479 --> 00:08:15,440
two of those axes were real one was

267
00:08:14,240 --> 00:08:18,240
imaginary

268
00:08:15,440 --> 00:08:20,319
okay and you have a rule alpha squared

269
00:08:18,240 --> 00:08:21,840
plus beta squared must equal to one

270
00:08:20,319 --> 00:08:24,080
because otherwise it doesn't sit on the

271
00:08:21,840 --> 00:08:25,919
sphere basically um and then you can

272
00:08:24,080 --> 00:08:27,919
express the states as follows the zero

273
00:08:25,919 --> 00:08:29,039
state is the pointing up state the one

274
00:08:27,919 --> 00:08:31,359
zero state

275
00:08:29,039 --> 00:08:32,718
the one state is the pointing down state

276
00:08:31,360 --> 00:08:35,279
the zero one state

277
00:08:32,719 --> 00:08:36,560
and so some state psi is a combination

278
00:08:35,279 --> 00:08:39,279
of those two

279
00:08:36,559 --> 00:08:41,439
right so you can write rewrite this as a

280
00:08:39,279 --> 00:08:42,559
as a vector of alpha of the top and beta

281
00:08:41,440 --> 00:08:45,519
on the bottom

282
00:08:42,559 --> 00:08:47,359
so with that in mind the model is called

283
00:08:45,519 --> 00:08:48,720
the block sphere spelled this way for

284
00:08:47,360 --> 00:08:50,000
those who want to go and have a look

285
00:08:48,720 --> 00:08:51,279
but you don't really need to understand

286
00:08:50,000 --> 00:08:53,120
that all you need to see is that there's

287
00:08:51,279 --> 00:08:54,800
little two place vectors

288
00:08:53,120 --> 00:08:56,320
and then you need to know how to

289
00:08:54,800 --> 00:08:58,719
actually interact with them so

290
00:08:56,320 --> 00:09:00,000
classically a classical computer is a

291
00:08:58,720 --> 00:09:03,760
sequence of gates

292
00:09:00,000 --> 00:09:05,120
okay and or nand nor xor we're all very

293
00:09:03,760 --> 00:09:08,080
very familiar with those

294
00:09:05,120 --> 00:09:09,519
so logic gates form regular computers

295
00:09:08,080 --> 00:09:12,160
well quantum gates

296
00:09:09,519 --> 00:09:14,240
form quantum computers and quantum gates

297
00:09:12,160 --> 00:09:14,560
are two by two matrices to match with

298
00:09:14,240 --> 00:09:17,040
the

299
00:09:14,560 --> 00:09:18,399
two place vectors okay and then for

300
00:09:17,040 --> 00:09:20,800
every qubit you add

301
00:09:18,399 --> 00:09:21,519
uh you get the next power of two size

302
00:09:20,800 --> 00:09:24,560
matrix

303
00:09:21,519 --> 00:09:25,200
being uh in play and that's all that

304
00:09:24,560 --> 00:09:28,079
happens

305
00:09:25,200 --> 00:09:28,959
you string the various matrices together

306
00:09:28,080 --> 00:09:30,720
and what you get

307
00:09:28,959 --> 00:09:32,319
is what's called a quantum circuit and

308
00:09:30,720 --> 00:09:33,519
that's your quantum algorithm

309
00:09:32,320 --> 00:09:36,080
so you'll see these kinds of line

310
00:09:33,519 --> 00:09:38,240
diagrams and those are quantum circuits

311
00:09:36,080 --> 00:09:39,519
okay now a brief brief break it's fine

312
00:09:38,240 --> 00:09:41,440
that's fine these are your

313
00:09:39,519 --> 00:09:43,200
post math kittens to make sure we can

314
00:09:41,440 --> 00:09:45,680
get everyone through all right

315
00:09:43,200 --> 00:09:47,279
um little kid a little little one in the

316
00:09:45,680 --> 00:09:49,120
middle i can let naughty uncle

317
00:09:47,279 --> 00:09:51,920
schwerdinger near you

318
00:09:49,120 --> 00:09:52,880
so um what do these circuits look like

319
00:09:51,920 --> 00:09:55,599
well they look like this

320
00:09:52,880 --> 00:09:56,800
so this is a circuit it's got three uh

321
00:09:55,600 --> 00:09:58,399
qubits here

322
00:09:56,800 --> 00:10:00,479
and it's got a classical register at the

323
00:09:58,399 --> 00:10:02,880
bottom it's three bits wide but it

324
00:10:00,480 --> 00:10:04,480
doesn't really matter

325
00:10:02,880 --> 00:10:06,160
we assume that they are always in the

326
00:10:04,480 --> 00:10:08,399
zero reset state

327
00:10:06,160 --> 00:10:10,480
um until we sort of manipulate them with

328
00:10:08,399 --> 00:10:12,160
gates so the first gate is a paulie x

329
00:10:10,480 --> 00:10:13,920
that turns it from a zero to a one

330
00:10:12,160 --> 00:10:16,800
that's what that does

331
00:10:13,920 --> 00:10:18,560
the hadamard gate the h box is the

332
00:10:16,800 --> 00:10:19,839
superposition gate so you put it into

333
00:10:18,560 --> 00:10:21,279
one and then you put it into

334
00:10:19,839 --> 00:10:23,360
superposition from there

335
00:10:21,279 --> 00:10:25,839
okay and that's different from putting

336
00:10:23,360 --> 00:10:27,360
it into superposition from a zero

337
00:10:25,839 --> 00:10:29,360
and then you have interesting gates like

338
00:10:27,360 --> 00:10:32,640
this so this is the controlled not

339
00:10:29,360 --> 00:10:34,079
gate so this is a not if and only if the

340
00:10:32,640 --> 00:10:36,399
upper bit is a one

341
00:10:34,079 --> 00:10:37,199
so if this is a one then whatever goes

342
00:10:36,399 --> 00:10:40,640
in here is

343
00:10:37,200 --> 00:10:42,640
is not the output okay and then if this

344
00:10:40,640 --> 00:10:44,959
is a zero it's a straight through

345
00:10:42,640 --> 00:10:46,160
okay and that's quite straightforward

346
00:10:44,959 --> 00:10:48,319
why do we need this

347
00:10:46,160 --> 00:10:49,519
turns out a c naught controlled knot

348
00:10:48,320 --> 00:10:51,360
which is what this is

349
00:10:49,519 --> 00:10:53,279
um is actually an entangler and

350
00:10:51,360 --> 00:10:54,880
detangler kind of useful

351
00:10:53,279 --> 00:10:56,240
so you just build these out into a

352
00:10:54,880 --> 00:10:58,000
quantum algorithm you'll get something

353
00:10:56,240 --> 00:10:58,880
like this the little black boxes are

354
00:10:58,000 --> 00:11:01,360
measurements

355
00:10:58,880 --> 00:11:02,959
so you we measure qubit zero and we

356
00:11:01,360 --> 00:11:04,800
measure qubit one and then we measure

357
00:11:02,959 --> 00:11:05,439
qubit two and we put the answer into the

358
00:11:04,800 --> 00:11:08,160
register

359
00:11:05,440 --> 00:11:08,480
okay now this circuit has a very fancy

360
00:11:08,160 --> 00:11:11,680
name

361
00:11:08,480 --> 00:11:13,839
it's called the quantum teleporter

362
00:11:11,680 --> 00:11:15,920
very star trek so what this does is it

363
00:11:13,839 --> 00:11:18,800
just teleports the state

364
00:11:15,920 --> 00:11:19,360
of qubit one into the state of qubit two

365
00:11:18,800 --> 00:11:20,719
um

366
00:11:19,360 --> 00:11:22,880
even if you don't know what the state in

367
00:11:20,720 --> 00:11:25,519
q1 is so i said that the

368
00:11:22,880 --> 00:11:27,040
x sets it to one so you would expect

369
00:11:25,519 --> 00:11:28,880
then to see

370
00:11:27,040 --> 00:11:30,640
the one moved over into the most

371
00:11:28,880 --> 00:11:32,240
significant bit of the output

372
00:11:30,640 --> 00:11:33,839
and if you measure it that's what you

373
00:11:32,240 --> 00:11:36,160
see this is what it looks like

374
00:11:33,839 --> 00:11:38,399
it might look suspiciously like python

375
00:11:36,160 --> 00:11:40,560
that's because it is

376
00:11:38,399 --> 00:11:43,120
this is all run from a python notebook

377
00:11:40,560 --> 00:11:44,880
this is a qasm simulator

378
00:11:43,120 --> 00:11:46,959
how did i come up with these how did i

379
00:11:44,880 --> 00:11:49,600
learn what to do well i looked at the

380
00:11:46,959 --> 00:11:51,359
ibm quantum experience this is what it

381
00:11:49,600 --> 00:11:52,000
looks like in fact i can actually show

382
00:11:51,360 --> 00:11:54,240
you

383
00:11:52,000 --> 00:11:55,360
messing around with little circuits so

384
00:11:54,240 --> 00:11:57,519
it's drag and drop

385
00:11:55,360 --> 00:11:59,440
i want to add a qubit i press a button

386
00:11:57,519 --> 00:12:02,480
it's really straightforward i can just

387
00:11:59,440 --> 00:12:04,000
go through uh add some qubits oh i don't

388
00:12:02,480 --> 00:12:04,320
i like the way this one's arranged so i

389
00:12:04,000 --> 00:12:06,240
can

390
00:12:04,320 --> 00:12:07,360
rearrange it like that and if you know

391
00:12:06,240 --> 00:12:08,320
what you're doing this is quite

392
00:12:07,360 --> 00:12:09,839
straightforward

393
00:12:08,320 --> 00:12:10,959
and then i can just go okay i want a

394
00:12:09,839 --> 00:12:12,480
measurement there and i want a

395
00:12:10,959 --> 00:12:13,680
measurement there i can always check the

396
00:12:12,480 --> 00:12:15,760
phase here

397
00:12:13,680 --> 00:12:17,680
and i can look up here how many times i

398
00:12:15,760 --> 00:12:19,439
want to run it which computer do i want

399
00:12:17,680 --> 00:12:21,359
to use and just go run

400
00:12:19,440 --> 00:12:22,480
you'll also note here there's a thing

401
00:12:21,360 --> 00:12:25,120
called qasm

402
00:12:22,480 --> 00:12:25,600
there's a quantum assembly language okay

403
00:12:25,120 --> 00:12:28,959
it's

404
00:12:25,600 --> 00:12:30,800
not as complex as cisc from x86

405
00:12:28,959 --> 00:12:32,160
uh nowhere near in fact it's not really

406
00:12:30,800 --> 00:12:35,199
as complicated as

407
00:12:32,160 --> 00:12:37,040
as arm but it does exist and it lets you

408
00:12:35,200 --> 00:12:37,600
describe the circuits as sequences of

409
00:12:37,040 --> 00:12:40,079
gates

410
00:12:37,600 --> 00:12:41,519
okay so kind of the theory of quantum

411
00:12:40,079 --> 00:12:43,120
computing is growing

412
00:12:41,519 --> 00:12:45,040
and it's only going to kind of get more

413
00:12:43,120 --> 00:12:46,480
and more so like showing people and

414
00:12:45,040 --> 00:12:48,880
exposing people to this is

415
00:12:46,480 --> 00:12:50,880
i think entirely useful and appropriate

416
00:12:48,880 --> 00:12:53,120
especially with quantum computing

417
00:12:50,880 --> 00:12:54,720
not necessarily being a million miles

418
00:12:53,120 --> 00:12:55,760
away we're going to come to exactly what

419
00:12:54,720 --> 00:12:57,839
i mean by that

420
00:12:55,760 --> 00:12:59,360
in a minute so let's just look at some

421
00:12:57,839 --> 00:13:00,240
resources i'm not going to go on any

422
00:12:59,360 --> 00:13:02,639
further about this

423
00:13:00,240 --> 00:13:04,800
i'm going to show you um some resources

424
00:13:02,639 --> 00:13:08,320
so the first resource is kisked

425
00:13:04,800 --> 00:13:09,680
kiskit is a really easy to use very well

426
00:13:08,320 --> 00:13:13,120
documented

427
00:13:09,680 --> 00:13:15,279
uh quantum circuit uh library

428
00:13:13,120 --> 00:13:17,040
for python it plugs directly into the

429
00:13:15,279 --> 00:13:18,079
ibm quantum experience so you can use

430
00:13:17,040 --> 00:13:20,079
the ui

431
00:13:18,079 --> 00:13:21,760
or you can just use the api from a

432
00:13:20,079 --> 00:13:23,359
jupiter notebook and it

433
00:13:21,760 --> 00:13:25,279
just works it just connects up you give

434
00:13:23,360 --> 00:13:26,079
it the api token which you can sign up

435
00:13:25,279 --> 00:13:28,720
for free

436
00:13:26,079 --> 00:13:30,319
and you can just go and start running

437
00:13:28,720 --> 00:13:31,600
quantum circuits in fact if you get

438
00:13:30,320 --> 00:13:33,760
really bored in this talk

439
00:13:31,600 --> 00:13:36,079
you can have a play with that it's a you

440
00:13:33,760 --> 00:13:39,519
know it's kind of fun i think

441
00:13:36,079 --> 00:13:41,680
then again i'm that kind of person so

442
00:13:39,519 --> 00:13:42,800
once you've sort of made a few circuits

443
00:13:41,680 --> 00:13:43,279
and haven't had a bit of a play you

444
00:13:42,800 --> 00:13:44,399
might

445
00:13:43,279 --> 00:13:46,480
want to go oh i want some more

446
00:13:44,399 --> 00:13:47,600
algorithms well one resource is called

447
00:13:46,480 --> 00:13:49,440
quantiki

448
00:13:47,600 --> 00:13:50,720
which i'm not gonna lie sounds like a

449
00:13:49,440 --> 00:13:52,639
cocktail

450
00:13:50,720 --> 00:13:54,079
in fact actually maybe i should go like

451
00:13:52,639 --> 00:13:56,160
can we go into like a def con

452
00:13:54,079 --> 00:13:57,199
on a selfie with jack daniel holding a

453
00:13:56,160 --> 00:13:59,279
quan tikki

454
00:13:57,199 --> 00:14:00,800
we need to make that happen somehow i'll

455
00:13:59,279 --> 00:14:02,639
have a think about that anyway

456
00:14:00,800 --> 00:14:04,319
um you might go to kwan tiki first

457
00:14:02,639 --> 00:14:06,160
quantity is a little bit easier to read

458
00:14:04,320 --> 00:14:07,839
uh quantum algorithm zoo is more

459
00:14:06,160 --> 00:14:10,639
comprehensive i think but

460
00:14:07,839 --> 00:14:11,839
um it's kind of written by postgraduate

461
00:14:10,639 --> 00:14:13,760
physics students for

462
00:14:11,839 --> 00:14:15,199
postgraduate physics students so you

463
00:14:13,760 --> 00:14:16,000
might not understand everything

464
00:14:15,199 --> 00:14:17,680
everywhere

465
00:14:16,000 --> 00:14:19,199
because there's it's still mostly

466
00:14:17,680 --> 00:14:20,719
physicists doing this okay i'm a

467
00:14:19,199 --> 00:14:22,959
mathematician by training

468
00:14:20,720 --> 00:14:24,160
um that's what my phd is in uh so i'm

469
00:14:22,959 --> 00:14:26,239
looking at this and i'm like

470
00:14:24,160 --> 00:14:27,360
that physics that's physics okay because

471
00:14:26,240 --> 00:14:28,800
they do weird things like sort of

472
00:14:27,360 --> 00:14:30,399
they've got these equations and they go

473
00:14:28,800 --> 00:14:32,160
and this bit doesn't matter and i'm like

474
00:14:30,399 --> 00:14:33,680
doesn't it you can't just do that

475
00:14:32,160 --> 00:14:35,040
turns out you can and there's good

476
00:14:33,680 --> 00:14:35,760
reasons for it it's how they sort of

477
00:14:35,040 --> 00:14:37,519
rationalize

478
00:14:35,760 --> 00:14:38,959
uh the math that sort of makes the

479
00:14:37,519 --> 00:14:39,920
physics and the physical world make

480
00:14:38,959 --> 00:14:41,040
sense for us

481
00:14:39,920 --> 00:14:42,240
but these are these are some good

482
00:14:41,040 --> 00:14:43,040
resources if you want to actually look

483
00:14:42,240 --> 00:14:44,639
at

484
00:14:43,040 --> 00:14:46,880
some quantum computing resources there's

485
00:14:44,639 --> 00:14:48,480
loads out there the three that i'm

486
00:14:46,880 --> 00:14:50,160
going to mention here the ibm quantum

487
00:14:48,480 --> 00:14:53,279
experience i've already mentioned

488
00:14:50,160 --> 00:14:56,160
um likewise there's also d-wave leap

489
00:14:53,279 --> 00:14:57,279
um now d-wave's different d-o d-wave is

490
00:14:56,160 --> 00:14:59,120
a thing called a quantum

491
00:14:57,279 --> 00:15:01,120
annealing machine which is different

492
00:14:59,120 --> 00:15:02,800
from the ibm

493
00:15:01,120 --> 00:15:04,959
quantum experience quantum computer i

494
00:15:02,800 --> 00:15:05,599
wouldn't say how or why or what is

495
00:15:04,959 --> 00:15:08,479
different

496
00:15:05,600 --> 00:15:09,680
but you'll notice differences um and the

497
00:15:08,480 --> 00:15:11,680
last one is uh

498
00:15:09,680 --> 00:15:13,279
it's like azure quantum i mean i'd love

499
00:15:11,680 --> 00:15:15,359
to talk about them i've been trying to

500
00:15:13,279 --> 00:15:16,800
email them for six weeks if anyone knows

501
00:15:15,360 --> 00:15:19,519
anyone at azure quantum

502
00:15:16,800 --> 00:15:20,160
do let me know because it's like i'd

503
00:15:19,519 --> 00:15:21,839
love to just

504
00:15:20,160 --> 00:15:23,600
just get some more documentation ask

505
00:15:21,839 --> 00:15:26,959
some questions but they're like

506
00:15:23,600 --> 00:15:28,720
it's like trying to contact monks so

507
00:15:26,959 --> 00:15:30,079
let's get back to security which is why

508
00:15:28,720 --> 00:15:33,600
we're all here

509
00:15:30,079 --> 00:15:33,599
what breaks and how

510
00:15:33,680 --> 00:15:39,599
shaw's algorithm what does it do

511
00:15:36,880 --> 00:15:41,360
it takes a number and it finds its prime

512
00:15:39,600 --> 00:15:41,920
factors now you can write a program to

513
00:15:41,360 --> 00:15:44,800
do that

514
00:15:41,920 --> 00:15:46,560
in python it's a few lines right the

515
00:15:44,800 --> 00:15:47,120
difference is your python program will

516
00:15:46,560 --> 00:15:50,800
take

517
00:15:47,120 --> 00:15:52,560
exponentially long to run compared to

518
00:15:50,800 --> 00:15:54,399
uh you know sort of the act of actually

519
00:15:52,560 --> 00:15:56,079
just finding the number by multiplying

520
00:15:54,399 --> 00:15:59,199
two numbers together what do i mean

521
00:15:56,079 --> 00:16:02,160
take the numbers 2 and 60.

522
00:15:59,199 --> 00:16:03,439
multiply those together you get 120. but

523
00:16:02,160 --> 00:16:04,399
if you want to find out which two

524
00:16:03,440 --> 00:16:07,440
numbers i had

525
00:16:04,399 --> 00:16:10,480
to multiply together to get 120 you have

526
00:16:07,440 --> 00:16:12,639
seven pairs excluding 120 and one

527
00:16:10,480 --> 00:16:13,759
you've got seven pairs of 14 numbers to

528
00:16:12,639 --> 00:16:16,399
choose from and find

529
00:16:13,759 --> 00:16:18,079
right and that that's why it's a hard

530
00:16:16,399 --> 00:16:19,600
problem as the numbers get bigger and

531
00:16:18,079 --> 00:16:20,319
bigger and bigger and bigger the number

532
00:16:19,600 --> 00:16:22,079
of factors

533
00:16:20,320 --> 00:16:23,360
also gets bigger and bigger and bigger

534
00:16:22,079 --> 00:16:25,920
and bigger and bigger

535
00:16:23,360 --> 00:16:26,480
um so much so that we we we know that

536
00:16:25,920 --> 00:16:28,479
it's a very

537
00:16:26,480 --> 00:16:30,000
very hard problem mp hard problem to

538
00:16:28,480 --> 00:16:31,519
actually factor prime

539
00:16:30,000 --> 00:16:33,600
uh factor numbers and see if and find

540
00:16:31,519 --> 00:16:34,839
their primes uh the frames that make

541
00:16:33,600 --> 00:16:37,600
them up

542
00:16:34,839 --> 00:16:41,279
now quantum computers

543
00:16:37,600 --> 00:16:41,680
can factor numbers in sub-exponential

544
00:16:41,279 --> 00:16:44,480
time

545
00:16:41,680 --> 00:16:46,160
and that's why it's a thing that's why

546
00:16:44,480 --> 00:16:48,880
people are losing their heads

547
00:16:46,160 --> 00:16:51,040
okay it's bad because we rely on

548
00:16:48,880 --> 00:16:52,000
factoring to be hard if factoring is no

549
00:16:51,040 --> 00:16:55,759
longer hard

550
00:16:52,000 --> 00:16:58,880
rsa and ecc immediately start to have

551
00:16:55,759 --> 00:17:00,399
problems okay likewise grover's

552
00:16:58,880 --> 00:17:04,000
algorithm poses problems

553
00:17:00,399 --> 00:17:06,000
to basically any symmetric cipher

554
00:17:04,000 --> 00:17:07,199
because what it does is it lets you do a

555
00:17:06,000 --> 00:17:10,000
dictionary search or

556
00:17:07,199 --> 00:17:10,640
a database search that it has n many

557
00:17:10,000 --> 00:17:12,720
entries

558
00:17:10,640 --> 00:17:14,720
and lets you do it in square root of n

559
00:17:12,720 --> 00:17:15,760
many tries what does that really mean

560
00:17:14,720 --> 00:17:17,439
well

561
00:17:15,760 --> 00:17:19,039
square root is taking a thing to the

562
00:17:17,439 --> 00:17:19,839
exponent of one half if you don't

563
00:17:19,039 --> 00:17:23,679
believe me

564
00:17:19,839 --> 00:17:25,839
i mean have a go so if you take

565
00:17:23,679 --> 00:17:26,959
the key length and halve it that's the

566
00:17:25,839 --> 00:17:28,960
effective length now

567
00:17:26,959 --> 00:17:30,880
it's not actually quite that simple

568
00:17:28,960 --> 00:17:32,880
there's a lot of nuance and detail which

569
00:17:30,880 --> 00:17:34,640
i am skipping way over

570
00:17:32,880 --> 00:17:36,799
but for the purposes of genuine

571
00:17:34,640 --> 00:17:38,440
intuition take the key length harvard so

572
00:17:36,799 --> 00:17:41,679
as128 becomes

573
00:17:38,440 --> 00:17:44,640
aes64 or des

574
00:17:41,679 --> 00:17:47,520
as you might want to call it aas256

575
00:17:44,640 --> 00:17:49,840
becomes as128 now as128 is still very

576
00:17:47,520 --> 00:17:50,799
secure it's all got 128 bits of security

577
00:17:49,840 --> 00:17:54,080
inside it

578
00:17:50,799 --> 00:17:54,559
but you know it it you still have this

579
00:17:54,080 --> 00:17:57,280
problem

580
00:17:54,559 --> 00:17:58,840
okay doesn't sort of go away okay if you

581
00:17:57,280 --> 00:18:00,720
listen to crypto pros they just go

582
00:17:58,840 --> 00:18:03,280
quantum

583
00:18:00,720 --> 00:18:03,840
and that's it end of talk and it's like

584
00:18:03,280 --> 00:18:05,440
no

585
00:18:03,840 --> 00:18:06,879
there's more to it than that what can we

586
00:18:05,440 --> 00:18:08,080
do about it there's quite a few things

587
00:18:06,880 --> 00:18:10,160
we can do about it

588
00:18:08,080 --> 00:18:11,439
so shaw's algorithm is going to take

589
00:18:10,160 --> 00:18:13,919
roughly seven years

590
00:18:11,440 --> 00:18:15,360
to really sort of work to be workable on

591
00:18:13,919 --> 00:18:16,080
a quantum computer this is making

592
00:18:15,360 --> 00:18:17,918
assumptions

593
00:18:16,080 --> 00:18:19,439
the assumptions are that we're going to

594
00:18:17,919 --> 00:18:19,919
double the number of qubits in a single

595
00:18:19,440 --> 00:18:23,039
machine

596
00:18:19,919 --> 00:18:25,360
every year and we're going to have

597
00:18:23,039 --> 00:18:26,240
lots and lots and lots of stability so

598
00:18:25,360 --> 00:18:28,240
at the moment

599
00:18:26,240 --> 00:18:29,280
a qubit is stable for i think the

600
00:18:28,240 --> 00:18:32,480
longest is a

601
00:18:29,280 --> 00:18:35,600
millisecond no microsecond um

602
00:18:32,480 --> 00:18:37,280
however you would need orders of days

603
00:18:35,600 --> 00:18:38,080
and weeks potentially so there's a lot

604
00:18:37,280 --> 00:18:40,559
of engineering

605
00:18:38,080 --> 00:18:42,080
and magic and plumbing and you know sort

606
00:18:40,559 --> 00:18:43,760
of absolute wizardry that has to happen

607
00:18:42,080 --> 00:18:46,000
before these things come about

608
00:18:43,760 --> 00:18:47,520
seven years for sure's algorithm eight

609
00:18:46,000 --> 00:18:48,960
years based on estimates

610
00:18:47,520 --> 00:18:51,360
for grover's algorithm to become

611
00:18:48,960 --> 00:18:53,520
workable against say aes256

612
00:18:51,360 --> 00:18:55,120
right if you look at mit's most

613
00:18:53,520 --> 00:18:58,000
efficient algorithm which runs in a few

614
00:18:55,120 --> 00:19:00,320
hours still needs 20 million cubits

615
00:18:58,000 --> 00:19:01,679
right so like this is you know sort of

616
00:19:00,320 --> 00:19:04,799
seven to 15 years

617
00:19:01,679 --> 00:19:05,679
eight to 18 years potentially uh away

618
00:19:04,799 --> 00:19:08,720
and that's making

619
00:19:05,679 --> 00:19:09,280
grand assumptions okay now nist's

620
00:19:08,720 --> 00:19:10,880
project

621
00:19:09,280 --> 00:19:12,639
is going to come come live but there's

622
00:19:10,880 --> 00:19:14,400
things you can do in the short term

623
00:19:12,640 --> 00:19:16,240
first of all you can use longer keys if

624
00:19:14,400 --> 00:19:18,160
this is a genuine threat

625
00:19:16,240 --> 00:19:20,480
if you're actually caring about this

626
00:19:18,160 --> 00:19:24,160
then use longer keys

627
00:19:20,480 --> 00:19:28,320
all right um 496 bit for rsa

628
00:19:24,160 --> 00:19:29,280
use ecc with p384 um upgrade to as256 as

629
00:19:28,320 --> 00:19:32,480
much as possible

630
00:19:29,280 --> 00:19:33,360
and also use aas in gcm mode gal while

631
00:19:32,480 --> 00:19:35,440
counter mode

632
00:19:33,360 --> 00:19:37,039
is really kind of nifty it's got that

633
00:19:35,440 --> 00:19:37,679
lovely gawa feel which gives that nice

634
00:19:37,039 --> 00:19:39,520
auth tag

635
00:19:37,679 --> 00:19:41,200
it makes it more complicated doesn't

636
00:19:39,520 --> 00:19:42,000
make it quantum resist more quantum

637
00:19:41,200 --> 00:19:43,679
resistant in

638
00:19:42,000 --> 00:19:45,520
a sense of algorithms but it does make

639
00:19:43,679 --> 00:19:47,280
it like you've got to have more qubits

640
00:19:45,520 --> 00:19:49,520
you've got to do more work

641
00:19:47,280 --> 00:19:50,879
okay so just make it harder the way that

642
00:19:49,520 --> 00:19:51,520
you usually do when you want to make

643
00:19:50,880 --> 00:19:54,000
things

644
00:19:51,520 --> 00:19:55,840
sort of uh more secure you you start off

645
00:19:54,000 --> 00:19:57,679
by making things more difficult

646
00:19:55,840 --> 00:19:59,039
ultimately we're really going to rely on

647
00:19:57,679 --> 00:20:01,200
nist making their mind up

648
00:19:59,039 --> 00:20:02,559
at some point about post-quantum

649
00:20:01,200 --> 00:20:06,159
cryptography

650
00:20:02,559 --> 00:20:07,600
now um nist are on their way

651
00:20:06,159 --> 00:20:09,440
okay and we're gonna have a little look

652
00:20:07,600 --> 00:20:10,399
at what's going on with uh post quantum

653
00:20:09,440 --> 00:20:13,600
cryptography

654
00:20:10,400 --> 00:20:14,799
um they have currently sort of uh got it

655
00:20:13,600 --> 00:20:16,719
down to round three

656
00:20:14,799 --> 00:20:18,240
okay so let's just have a little look at

657
00:20:16,720 --> 00:20:20,240
what's going on so

658
00:20:18,240 --> 00:20:21,919
casual crypto with its algorithms that

659
00:20:20,240 --> 00:20:23,760
are vulnerable to quantum

660
00:20:21,919 --> 00:20:26,080
algorithms ah no no we don't want that

661
00:20:23,760 --> 00:20:26,640
we want post quantum cryptography my

662
00:20:26,080 --> 00:20:29,360
friend

663
00:20:26,640 --> 00:20:30,080
that is the way forward okay what does

664
00:20:29,360 --> 00:20:32,639
it mean

665
00:20:30,080 --> 00:20:33,760
it means this it's cryptography that

666
00:20:32,640 --> 00:20:36,159
doesn't rely

667
00:20:33,760 --> 00:20:38,960
on number factoring i think we know

668
00:20:36,159 --> 00:20:41,360
content computers are really good at

669
00:20:38,960 --> 00:20:43,360
it doesn't rely on that problem for its

670
00:20:41,360 --> 00:20:44,879
security that's all it really means okay

671
00:20:43,360 --> 00:20:46,639
so there's several different schemes out

672
00:20:44,880 --> 00:20:49,440
there the finalists are these

673
00:20:46,640 --> 00:20:51,600
so lattice-based crypto super singular

674
00:20:49,440 --> 00:20:53,200
isogeny is on elliptic curves

675
00:20:51,600 --> 00:20:55,039
that's going to get you a good score

676
00:20:53,200 --> 00:20:56,240
scrabble and there's some hashing

677
00:20:55,039 --> 00:20:58,640
algorithms as well

678
00:20:56,240 --> 00:21:00,559
so let's look at what's going on okay so

679
00:20:58,640 --> 00:21:01,360
this is a very bare-bones example if you

680
00:21:00,559 --> 00:21:02,960
want more

681
00:21:01,360 --> 00:21:04,479
uh link at the bottom there's a summer

682
00:21:02,960 --> 00:21:06,559
school slides which are really good

683
00:21:04,480 --> 00:21:10,080
these are taken from those slides okay

684
00:21:06,559 --> 00:21:11,600
verbatim not quite but never enough um

685
00:21:10,080 --> 00:21:13,360
where we're just gonna look at an old

686
00:21:11,600 --> 00:21:14,959
problem where we know the solution and

687
00:21:13,360 --> 00:21:16,158
then we can modulate it to make it that

688
00:21:14,960 --> 00:21:18,159
we don't have a solution

689
00:21:16,159 --> 00:21:19,440
and also we don't think content

690
00:21:18,159 --> 00:21:21,360
computers do either

691
00:21:19,440 --> 00:21:22,880
right so here's what happens you have an

692
00:21:21,360 --> 00:21:24,559
array of numbers

693
00:21:22,880 --> 00:21:26,480
and then you have a secret and you

694
00:21:24,559 --> 00:21:27,360
multiply that secret down and then take

695
00:21:26,480 --> 00:21:30,640
the answer

696
00:21:27,360 --> 00:21:32,158
mod 13. okay now if you did a linear

697
00:21:30,640 --> 00:21:33,600
algebra course at high school which by

698
00:21:32,159 --> 00:21:35,280
the way is all you really need

699
00:21:33,600 --> 00:21:35,918
mathematically to do a lot of quantum

700
00:21:35,280 --> 00:21:37,280
stuff

701
00:21:35,919 --> 00:21:38,640
then you'll know that this isn't secure

702
00:21:37,280 --> 00:21:39,760
there's a thing called gaussian

703
00:21:38,640 --> 00:21:42,080
elimination

704
00:21:39,760 --> 00:21:44,559
uh which is a very straightforward

705
00:21:42,080 --> 00:21:46,399
algorithm that actually lets you easily

706
00:21:44,559 --> 00:21:46,720
work out what the secret is okay if i

707
00:21:46,400 --> 00:21:49,120
know

708
00:21:46,720 --> 00:21:50,720
the random array and the output can very

709
00:21:49,120 --> 00:21:51,360
straightforward work out what's going on

710
00:21:50,720 --> 00:21:54,080
in the middle

711
00:21:51,360 --> 00:21:55,280
okay however i can make it more

712
00:21:54,080 --> 00:21:57,199
difficult if i do this

713
00:21:55,280 --> 00:21:58,960
if i add a little bit of noise there's

714
00:21:57,200 --> 00:22:00,640
two bits of noise the first bit is

715
00:21:58,960 --> 00:22:01,919
whether it's a zero or one

716
00:22:00,640 --> 00:22:04,000
and the second bit is whether it's

717
00:22:01,919 --> 00:22:05,679
adding or taking away and there's a

718
00:22:04,000 --> 00:22:07,679
confusion with two of the bits because

719
00:22:05,679 --> 00:22:08,240
if you add zero or take away zero then

720
00:22:07,679 --> 00:22:10,480
you know

721
00:22:08,240 --> 00:22:11,760
it's the same thing right so you've got

722
00:22:10,480 --> 00:22:13,200
this little bit of noise

723
00:22:11,760 --> 00:22:14,799
that doesn't get recorded it just gets

724
00:22:13,200 --> 00:22:16,480
used in this calculation

725
00:22:14,799 --> 00:22:18,000
so here you can see that all the numbers

726
00:22:16,480 --> 00:22:21,520
on the right in the output

727
00:22:18,000 --> 00:22:23,600
have kind of modulated just a little bit

728
00:22:21,520 --> 00:22:25,200
so the problem so this is called with a

729
00:22:23,600 --> 00:22:26,799
this is the with errors bit of learning

730
00:22:25,200 --> 00:22:29,280
with errors okay

731
00:22:26,799 --> 00:22:30,240
this random noise skews everything just

732
00:22:29,280 --> 00:22:32,158
a little bit

733
00:22:30,240 --> 00:22:33,760
but of course over very large arrays and

734
00:22:32,159 --> 00:22:34,960
very large numbers and lots and lots of

735
00:22:33,760 --> 00:22:37,600
iterations

736
00:22:34,960 --> 00:22:39,200
that error grows and grows and grows so

737
00:22:37,600 --> 00:22:40,959
actually the problem becomes this given

738
00:22:39,200 --> 00:22:42,559
the random array and the output

739
00:22:40,960 --> 00:22:44,799
which by the way is what forms your

740
00:22:42,559 --> 00:22:46,320
public key for learning with errors and

741
00:22:44,799 --> 00:22:47,760
a similar thing for ring learning with

742
00:22:46,320 --> 00:22:49,039
errors the difference being that you

743
00:22:47,760 --> 00:22:50,720
don't store the whole array

744
00:22:49,039 --> 00:22:53,280
the array is made up of rotational

745
00:22:50,720 --> 00:22:56,640
permutations of like various rows

746
00:22:53,280 --> 00:22:58,240
um given an array and given an output

747
00:22:56,640 --> 00:22:59,840
and facts you know about it like it's

748
00:22:58,240 --> 00:23:03,520
modulo 13

749
00:22:59,840 --> 00:23:05,280
find the red thing okay find the secret

750
00:23:03,520 --> 00:23:06,559
and that's actually really really hard

751
00:23:05,280 --> 00:23:08,559
we know it's hard because it's the same

752
00:23:06,559 --> 00:23:11,760
problem as this if i

753
00:23:08,559 --> 00:23:12,158
put a blue dot in the middle okay and

754
00:23:11,760 --> 00:23:14,640
then

755
00:23:12,159 --> 00:23:16,240
i have this lovely lattice array so i

756
00:23:14,640 --> 00:23:17,760
have all these lovely parallel

757
00:23:16,240 --> 00:23:20,080
vectors here and these lovely parallel

758
00:23:17,760 --> 00:23:21,679
vectors going this way then if i put a

759
00:23:20,080 --> 00:23:23,840
dot in the middle

760
00:23:21,679 --> 00:23:25,120
what's the shortest what's the nearest

761
00:23:23,840 --> 00:23:27,520
valid point

762
00:23:25,120 --> 00:23:29,678
to that dot now in two dimensions it

763
00:23:27,520 --> 00:23:31,520
doesn't look particularly difficult

764
00:23:29,679 --> 00:23:33,200
but as you put more and more and more

765
00:23:31,520 --> 00:23:34,720
dimensions on this as your arrays get

766
00:23:33,200 --> 00:23:35,200
bigger and bigger and bigger and bigger

767
00:23:34,720 --> 00:23:37,760
right

768
00:23:35,200 --> 00:23:39,760
that problem gets exponentially hard and

769
00:23:37,760 --> 00:23:42,158
the important point is this there is no

770
00:23:39,760 --> 00:23:43,279
known quantum algorithmic speed up for

771
00:23:42,159 --> 00:23:46,400
this problem

772
00:23:43,279 --> 00:23:46,720
okay and that my friends is what makes

773
00:23:46,400 --> 00:23:49,200
it

774
00:23:46,720 --> 00:23:50,559
post quantum okay we haven't used number

775
00:23:49,200 --> 00:23:52,080
factoring anywhere

776
00:23:50,559 --> 00:23:53,678
and okay you might say but how does it

777
00:23:52,080 --> 00:23:55,600
know if it's a one or a zero when the

778
00:23:53,679 --> 00:23:57,400
error term isn't is thrown away

779
00:23:55,600 --> 00:24:00,879
because what you get is you get like

780
00:23:57,400 --> 00:24:02,240
0.012 or you get 0.998

781
00:24:00,880 --> 00:24:04,799
so you get numbers that are very close

782
00:24:02,240 --> 00:24:06,799
to 1 and very close to zero and you go

783
00:24:04,799 --> 00:24:08,158
close enough close enough close enough

784
00:24:06,799 --> 00:24:10,000
close enough close enough

785
00:24:08,159 --> 00:24:11,679
and you get the right answer out the

786
00:24:10,000 --> 00:24:14,480
other end if you try and do it

787
00:24:11,679 --> 00:24:16,559
any other way you get nonsense which is

788
00:24:14,480 --> 00:24:19,760
kind of the idea of cryptography

789
00:24:16,559 --> 00:24:21,600
so we know it works but we don't know

790
00:24:19,760 --> 00:24:22,960
yet whether nist think that it works

791
00:24:21,600 --> 00:24:25,918
well enough for it to be used

792
00:24:22,960 --> 00:24:26,880
uh widespread so when will this be

793
00:24:25,919 --> 00:24:28,960
commonplace well

794
00:24:26,880 --> 00:24:30,480
currently the candidates just so they're

795
00:24:28,960 --> 00:24:31,679
here again by the way yes it is a

796
00:24:30,480 --> 00:24:34,400
hashing algorithm called

797
00:24:31,679 --> 00:24:36,240
crystals dilithium okay just in case you

798
00:24:34,400 --> 00:24:38,559
wonder that's the actual name

799
00:24:36,240 --> 00:24:39,360
um there's these which are set out so

800
00:24:38,559 --> 00:24:41,918
psych i think

801
00:24:39,360 --> 00:24:43,678
is from microsoft with aws well i know

802
00:24:41,919 --> 00:24:46,799
aws i think support it

803
00:24:43,679 --> 00:24:47,919
um ntru comes out of i can't remember

804
00:24:46,799 --> 00:24:49,360
where it comes out about how to double

805
00:24:47,919 --> 00:24:51,760
check that i'm sure someone in the chat

806
00:24:49,360 --> 00:24:52,879
somewhere will correct me um but like

807
00:24:51,760 --> 00:24:53,919
these things have all been submitted

808
00:24:52,880 --> 00:24:54,880
these are the main eight

809
00:24:53,919 --> 00:24:56,240
and then what you're going to have as

810
00:24:54,880 --> 00:24:58,559
well there's lots of backups so you i

811
00:24:56,240 --> 00:25:00,799
think you've got like 13 or 15 backups

812
00:24:58,559 --> 00:25:02,799
waiting in the wings because as they've

813
00:25:00,799 --> 00:25:04,720
they started with a huge list of them

814
00:25:02,799 --> 00:25:06,320
and gradually they've worked out this

815
00:25:04,720 --> 00:25:07,520
isn't as strong as we thought this is a

816
00:25:06,320 --> 00:25:09,279
problem over here

817
00:25:07,520 --> 00:25:10,960
someone finds a new attack against the

818
00:25:09,279 --> 00:25:12,159
crypto system and then they sort of take

819
00:25:10,960 --> 00:25:15,120
it out of the running

820
00:25:12,159 --> 00:25:16,960
so what's left should be pretty good now

821
00:25:15,120 --> 00:25:18,639
we're into round three we can expect

822
00:25:16,960 --> 00:25:20,159
more attacks to come about we can expect

823
00:25:18,640 --> 00:25:21,600
more writing to come about we don't even

824
00:25:20,159 --> 00:25:23,200
know it's going to be called though

825
00:25:21,600 --> 00:25:25,439
i don't even know if nist know what it's

826
00:25:23,200 --> 00:25:26,640
going to be called you had dez and aes

827
00:25:25,440 --> 00:25:28,640
and that was like we're going to choose

828
00:25:26,640 --> 00:25:30,000
one i don't think we're going to get one

829
00:25:28,640 --> 00:25:31,520
i think we're going to have in around

830
00:25:30,000 --> 00:25:33,600
two years we're going to have

831
00:25:31,520 --> 00:25:35,600
multiple ciphers with lots of notes and

832
00:25:33,600 --> 00:25:38,399
requirements and considerations

833
00:25:35,600 --> 00:25:38,799
and recommendations and estimations and

834
00:25:38,400 --> 00:25:41,679
all

835
00:25:38,799 --> 00:25:43,520
manner of different key lengths okay i i

836
00:25:41,679 --> 00:25:44,559
think it's a mistake to say that nist

837
00:25:43,520 --> 00:25:47,120
are gonna go

838
00:25:44,559 --> 00:25:48,720
here is the solution i think that

839
00:25:47,120 --> 00:25:51,439
genuinely it's gonna be

840
00:25:48,720 --> 00:25:53,120
so here's what we've ended up with and

841
00:25:51,440 --> 00:25:53,360
we're just gonna have to make sense of

842
00:25:53,120 --> 00:25:56,320
it

843
00:25:53,360 --> 00:25:56,719
somehow the way that we obviously have

844
00:25:56,320 --> 00:26:01,279
done

845
00:25:56,720 --> 00:26:01,279
in every other area of security

846
00:26:01,520 --> 00:26:05,200
again i make every apology for my lack

847
00:26:03,840 --> 00:26:11,840
of apology for the maths but here's a

848
00:26:05,200 --> 00:26:11,840
little math cut with a little mew

849
00:26:13,760 --> 00:26:17,919
just going to leave that there but i'm a

850
00:26:17,039 --> 00:26:20,400
mathematician

851
00:26:17,919 --> 00:26:21,039
okay the comedy is a bonus all right

852
00:26:20,400 --> 00:26:24,159
okay

853
00:26:21,039 --> 00:26:25,520
so quantum key distribution is

854
00:26:24,159 --> 00:26:26,960
the next kind of topic i'm not going to

855
00:26:25,520 --> 00:26:29,120
go too much in depth all i'm going to do

856
00:26:26,960 --> 00:26:31,120
is say that in 1984

857
00:26:29,120 --> 00:26:32,959
uh bennett and bressar worked out a way

858
00:26:31,120 --> 00:26:34,239
of using quantum effects to make a

859
00:26:32,960 --> 00:26:37,600
provably secure

860
00:26:34,240 --> 00:26:40,720
quantum key exchange mechanism okay so

861
00:26:37,600 --> 00:26:42,480
this became known as bb-84 okay

862
00:26:40,720 --> 00:26:44,640
and then there was a subsequent system

863
00:26:42,480 --> 00:26:46,240
called eckhart 91 or e-91

864
00:26:44,640 --> 00:26:48,000
and then after that there's been loads

865
00:26:46,240 --> 00:26:49,679
ever since all kind of modulations and

866
00:26:48,000 --> 00:26:50,559
changes bipide and all these other

867
00:26:49,679 --> 00:26:52,559
things

868
00:26:50,559 --> 00:26:54,320
um and these are in use there are

869
00:26:52,559 --> 00:26:56,399
companies that produce these systems in

870
00:26:54,320 --> 00:26:58,240
fact bt and toshiba announced just

871
00:26:56,400 --> 00:26:59,440
yesterday that they were building the

872
00:26:58,240 --> 00:27:01,600
uk's first

873
00:26:59,440 --> 00:27:03,679
quantum secure industrial network

874
00:27:01,600 --> 00:27:05,199
between key uk smart production

875
00:27:03,679 --> 00:27:06,720
facilities

876
00:27:05,200 --> 00:27:08,080
so what are they doing well here's a

877
00:27:06,720 --> 00:27:08,960
kind of a rough thing this paper by the

878
00:27:08,080 --> 00:27:11,600
way at the bottom

879
00:27:08,960 --> 00:27:13,200
is really good it's a little bit

880
00:27:11,600 --> 00:27:13,678
physically and messy at times but it's

881
00:27:13,200 --> 00:27:14,799
actually

882
00:27:13,679 --> 00:27:16,480
the bits that you that are quite

883
00:27:14,799 --> 00:27:18,080
readable are worth having a look at if

884
00:27:16,480 --> 00:27:19,600
this if you this tickles your fancy and

885
00:27:18,080 --> 00:27:21,840
you're a little bit curious

886
00:27:19,600 --> 00:27:22,639
so here's what happens alice chooses

887
00:27:21,840 --> 00:27:25,120
some basis

888
00:27:22,640 --> 00:27:26,159
left right up down diagonal or the other

889
00:27:25,120 --> 00:27:28,959
diagonal right

890
00:27:26,159 --> 00:27:30,480
and then polarizes a photon and then

891
00:27:28,960 --> 00:27:33,279
sends that down a fiber

892
00:27:30,480 --> 00:27:36,159
that goes around and it ends up in bob's

893
00:27:33,279 --> 00:27:38,559
inbox or bob's input port right

894
00:27:36,159 --> 00:27:40,799
bob then chooses either the up down left

895
00:27:38,559 --> 00:27:41,360
right basis or the 45 degree rotated

896
00:27:40,799 --> 00:27:42,879
basis

897
00:27:41,360 --> 00:27:44,479
and then does a measurement and then

898
00:27:42,880 --> 00:27:45,440
they swap information so alongside the

899
00:27:44,480 --> 00:27:47,120
fiber line

900
00:27:45,440 --> 00:27:49,279
you've also got a lot of quantum sorry a

901
00:27:47,120 --> 00:27:50,479
lot of classical channels as well okay

902
00:27:49,279 --> 00:27:51,919
where you're constantly swapping

903
00:27:50,480 --> 00:27:53,200
information i'm about to send a bit

904
00:27:51,919 --> 00:27:54,640
here's what i did here's all these

905
00:27:53,200 --> 00:27:56,159
things there's a lot of error correction

906
00:27:54,640 --> 00:27:58,159
that happens as well because quantum

907
00:27:56,159 --> 00:28:00,559
systems are very prone to error

908
00:27:58,159 --> 00:28:02,320
um and then once you go through

909
00:28:00,559 --> 00:28:03,600
everything you can see that uh

910
00:28:02,320 --> 00:28:06,240
you know where they match up in the

911
00:28:03,600 --> 00:28:08,879
bottom there um bob

912
00:28:06,240 --> 00:28:09,919
miss measures the wrong basis so there's

913
00:28:08,880 --> 00:28:11,840
a zero at the end

914
00:28:09,919 --> 00:28:13,039
then bob gets it right so there's a one

915
00:28:11,840 --> 00:28:15,039
then bob gets it wrong

916
00:28:13,039 --> 00:28:16,879
but here he measures a one instead and

917
00:28:15,039 --> 00:28:18,158
so you get this kind of output here

918
00:28:16,880 --> 00:28:20,480
and then once you've gone through it

919
00:28:18,159 --> 00:28:22,720
enough times

920
00:28:20,480 --> 00:28:23,919
um the output is the reason for that one

921
00:28:22,720 --> 00:28:26,080
by the way is um

922
00:28:23,919 --> 00:28:27,200
to be positional um once you go through

923
00:28:26,080 --> 00:28:28,960
enough times

924
00:28:27,200 --> 00:28:31,120
then you have enough information to be

925
00:28:28,960 --> 00:28:32,480
able to say ah okay this is the key and

926
00:28:31,120 --> 00:28:35,279
we know we have a secret

927
00:28:32,480 --> 00:28:38,720
okay i mentioned the error term if

928
00:28:35,279 --> 00:28:41,919
there's someone listening on the line

929
00:28:38,720 --> 00:28:43,760
then that error term will rise

930
00:28:41,919 --> 00:28:45,440
and that's how you know if there's an

931
00:28:43,760 --> 00:28:47,120
eve on the line it's called the beta

932
00:28:45,440 --> 00:28:48,720
term in the mathematics

933
00:28:47,120 --> 00:28:50,479
and that's how you measure it so you can

934
00:28:48,720 --> 00:28:51,760
actually work out oh we think there's an

935
00:28:50,480 --> 00:28:53,039
eve there we're going to go and do

936
00:28:51,760 --> 00:28:55,279
something else or we're going to start

937
00:28:53,039 --> 00:28:55,279
again

938
00:28:55,360 --> 00:28:58,719
so these algorithms are quite sort of

939
00:28:57,520 --> 00:28:59,918
quite quite nifty there's a lot of

940
00:28:58,720 --> 00:29:01,440
interesting stuff going on

941
00:28:59,919 --> 00:29:03,600
but if it's from the mid 80s you might

942
00:29:01,440 --> 00:29:04,799
say well why isn't it everywhere then

943
00:29:03,600 --> 00:29:07,760
well remember i said

944
00:29:04,799 --> 00:29:09,200
you have to have a fiber line between

945
00:29:07,760 --> 00:29:10,640
alice and bob

946
00:29:09,200 --> 00:29:12,240
now let me walk you through some facts

947
00:29:10,640 --> 00:29:13,120
about qkd which often aren't talked

948
00:29:12,240 --> 00:29:14,799
about

949
00:29:13,120 --> 00:29:16,639
the mathematics that makes regular

950
00:29:14,799 --> 00:29:18,399
crypto secure doesn't care about the

951
00:29:16,640 --> 00:29:20,320
channel it's communicated over

952
00:29:18,399 --> 00:29:22,479
whereas the mathematics that makes

953
00:29:20,320 --> 00:29:24,240
quantum systems secure

954
00:29:22,480 --> 00:29:26,000
they do care okay because it's part of

955
00:29:24,240 --> 00:29:28,080
the system so what you get out the

956
00:29:26,000 --> 00:29:29,840
end is that you you can't packet switch

957
00:29:28,080 --> 00:29:34,240
photons without measuring them

958
00:29:29,840 --> 00:29:36,320
all right now um i did a a a lot of

959
00:29:34,240 --> 00:29:38,080
another discussion in an article in

960
00:29:36,320 --> 00:29:39,279
forbes where with davey winder and a few

961
00:29:38,080 --> 00:29:41,678
others

962
00:29:39,279 --> 00:29:43,440
to talk about loft orbital and you know

963
00:29:41,679 --> 00:29:44,559
they're using space which is great you

964
00:29:43,440 --> 00:29:46,960
know sort of yeah space

965
00:29:44,559 --> 00:29:47,600
rockets brilliant um all the things

966
00:29:46,960 --> 00:29:51,039
quantum

967
00:29:47,600 --> 00:29:52,480
space oh my god um but ultimately like

968
00:29:51,039 --> 00:29:53,120
sort of there's problems like if it's

969
00:29:52,480 --> 00:29:56,159
cloudy

970
00:29:53,120 --> 00:29:58,639
you can't do key exchange

971
00:29:56,159 --> 00:29:59,360
you can't also you can't control

972
00:29:58,640 --> 00:30:02,399
atmospheric

973
00:29:59,360 --> 00:30:04,559
dispersion and diffusion and diffraction

974
00:30:02,399 --> 00:30:06,320
you can't that's just just happens no

975
00:30:04,559 --> 00:30:07,360
matter how narrow your laser beam is it

976
00:30:06,320 --> 00:30:09,200
happens

977
00:30:07,360 --> 00:30:10,879
so if you want to have a look there's a

978
00:30:09,200 --> 00:30:12,799
qr code to the link or you can just

979
00:30:10,880 --> 00:30:13,760
search for the meet the scrappy space

980
00:30:12,799 --> 00:30:16,320
startup

981
00:30:13,760 --> 00:30:18,399
love that really nice bit of a bit of

982
00:30:16,320 --> 00:30:20,639
headline manship there

983
00:30:18,399 --> 00:30:22,239
companies that make ukd systems include

984
00:30:20,640 --> 00:30:23,919
a company called id contigue

985
00:30:22,240 --> 00:30:26,480
uh based in switzerland but they've been

986
00:30:23,919 --> 00:30:28,799
doing it like 2001. they've got a lot of

987
00:30:26,480 --> 00:30:29,520
expertise around this magic with their

988
00:30:28,799 --> 00:30:31,200
queue

989
00:30:29,520 --> 00:30:32,639
um i've been around for a while as well

990
00:30:31,200 --> 00:30:33,520
and toshiba have been in this space for

991
00:30:32,640 --> 00:30:35,360
some time

992
00:30:33,520 --> 00:30:36,559
if you want to kind of scare a qkd

993
00:30:35,360 --> 00:30:38,320
vendor though

994
00:30:36,559 --> 00:30:40,639
ask them how their tech works with

995
00:30:38,320 --> 00:30:44,480
cloud-based infrastructure

996
00:30:40,640 --> 00:30:45,919
just ask i'll say no more

997
00:30:44,480 --> 00:30:47,360
now you might say i want to hack this

998
00:30:45,919 --> 00:30:49,760
wouldn't it be cool to do it to be a

999
00:30:47,360 --> 00:30:51,279
quantum hacker would be brilliant

1000
00:30:49,760 --> 00:30:53,679
it's you know absolutely fantastic

1001
00:30:51,279 --> 00:30:57,120
however you wouldn't be the first

1002
00:30:53,679 --> 00:30:59,200
back in 2011 uh lars leidersen

1003
00:30:57,120 --> 00:31:00,158
and company found that you could shine a

1004
00:30:59,200 --> 00:31:02,960
very bright

1005
00:31:00,159 --> 00:31:04,159
laser beam down the fiber and then work

1006
00:31:02,960 --> 00:31:06,880
out

1007
00:31:04,159 --> 00:31:08,320
by the reflection what the phase was uh

1008
00:31:06,880 --> 00:31:09,039
what the measurement phase that bob was

1009
00:31:08,320 --> 00:31:11,918
using was

1010
00:31:09,039 --> 00:31:12,720
so actually um you find that uh you know

1011
00:31:11,919 --> 00:31:14,320
you can actually

1012
00:31:12,720 --> 00:31:16,320
get all these little side channels and

1013
00:31:14,320 --> 00:31:16,960
hacks the original qkd system from the

1014
00:31:16,320 --> 00:31:19,200
80s

1015
00:31:16,960 --> 00:31:20,000
you didn't need to be eve you just

1016
00:31:19,200 --> 00:31:22,080
listened

1017
00:31:20,000 --> 00:31:23,760
for which optics were being switched in

1018
00:31:22,080 --> 00:31:24,480
and out because they all made different

1019
00:31:23,760 --> 00:31:26,399
sounds

1020
00:31:24,480 --> 00:31:28,159
so if you were in the lab you could

1021
00:31:26,399 --> 00:31:30,799
easily work out what the key was

1022
00:31:28,159 --> 00:31:31,760
you know there's stuff like injection

1023
00:31:30,799 --> 00:31:34,000
locking

1024
00:31:31,760 --> 00:31:35,600
and trojan states being put in through

1025
00:31:34,000 --> 00:31:37,360
various sort of laser modulation

1026
00:31:35,600 --> 00:31:39,279
patterns and laser frequencies

1027
00:31:37,360 --> 00:31:40,639
it's absolutely fascinating there was

1028
00:31:39,279 --> 00:31:41,840
these slides will be online afterwards

1029
00:31:40,640 --> 00:31:42,480
so don't worry about scribbling them

1030
00:31:41,840 --> 00:31:43,600
down now

1031
00:31:42,480 --> 00:31:45,840
but there's some really interesting

1032
00:31:43,600 --> 00:31:46,959
stuff out there and kind of the last

1033
00:31:45,840 --> 00:31:48,480
point in my final

1034
00:31:46,960 --> 00:31:50,320
few minutes i'm going to talk about like

1035
00:31:48,480 --> 00:31:51,679
sort of quantum algorithms okay

1036
00:31:50,320 --> 00:31:53,600
i've talked about quantum computers

1037
00:31:51,679 --> 00:31:56,000
surely there's more to it than just

1038
00:31:53,600 --> 00:31:57,840
you know breaking factoring big numbers

1039
00:31:56,000 --> 00:31:59,120
for tlas like sure there's more to life

1040
00:31:57,840 --> 00:32:01,039
than that and there is

1041
00:31:59,120 --> 00:32:02,479
quantum finance is a thing doing you

1042
00:32:01,039 --> 00:32:04,960
know sort of a stock market

1043
00:32:02,480 --> 00:32:07,120
uh modeling and pricing on a quantum

1044
00:32:04,960 --> 00:32:09,039
computer is an active area of study

1045
00:32:07,120 --> 00:32:10,799
okay quantum chemistry simulations i

1046
00:32:09,039 --> 00:32:13,440
mean you've got a quantum system

1047
00:32:10,799 --> 00:32:14,000
except you can control every part of it

1048
00:32:13,440 --> 00:32:16,320
so can you

1049
00:32:14,000 --> 00:32:17,440
simulate molecules for example and the

1050
00:32:16,320 --> 00:32:20,080
answer is yes like

1051
00:32:17,440 --> 00:32:21,679
google released a thing last uh last

1052
00:32:20,080 --> 00:32:23,039
week or month i think i think it was

1053
00:32:21,679 --> 00:32:25,279
where they said yes we're going to

1054
00:32:23,039 --> 00:32:26,158
actually uh you know state the claim

1055
00:32:25,279 --> 00:32:28,640
that we've we've

1056
00:32:26,159 --> 00:32:29,519
simulated um i think what was it salt i

1057
00:32:28,640 --> 00:32:31,440
can't remember

1058
00:32:29,519 --> 00:32:32,559
um but like so yeah these are really

1059
00:32:31,440 --> 00:32:35,279
growing areas of

1060
00:32:32,559 --> 00:32:36,158
of research and and innovation what

1061
00:32:35,279 --> 00:32:37,600
about quantum

1062
00:32:36,159 --> 00:32:39,679
random number generator because i've

1063
00:32:37,600 --> 00:32:41,039
mentioned that quantum is a really good

1064
00:32:39,679 --> 00:32:43,760
source of randomness

1065
00:32:41,039 --> 00:32:45,600
so why don't we actually go well can we

1066
00:32:43,760 --> 00:32:47,360
see the csprng

1067
00:32:45,600 --> 00:32:49,360
cryptographically secure studio random

1068
00:32:47,360 --> 00:32:52,158
number generator off quantum well

1069
00:32:49,360 --> 00:32:53,918
yes so what i'm going to release at the

1070
00:32:52,159 --> 00:32:55,279
end of in a couple of slides and show

1071
00:32:53,919 --> 00:32:57,919
you is some code that

1072
00:32:55,279 --> 00:32:59,039
actually does this okay um so this is

1073
00:32:57,919 --> 00:32:59,679
kind of the workflow that we're going to

1074
00:32:59,039 --> 00:33:02,799
follow

1075
00:32:59,679 --> 00:33:05,120
um we use local entropy

1076
00:33:02,799 --> 00:33:06,000
and then when we want more we just see

1077
00:33:05,120 --> 00:33:08,158
the entropy

1078
00:33:06,000 --> 00:33:09,600
uh in the from our entropy pool into the

1079
00:33:08,159 --> 00:33:12,000
random number generator

1080
00:33:09,600 --> 00:33:13,039
and then once we run low on bits we ask

1081
00:33:12,000 --> 00:33:15,360
ibm q

1082
00:33:13,039 --> 00:33:16,158
for some for a load of qubits and then

1083
00:33:15,360 --> 00:33:17,760
we get them

1084
00:33:16,159 --> 00:33:19,120
now the circuit we use is really simple

1085
00:33:17,760 --> 00:33:21,200
if you remember i said that we use the

1086
00:33:19,120 --> 00:33:23,360
hadamard gate to go to superposition

1087
00:33:21,200 --> 00:33:25,760
superposition means you have equal

1088
00:33:23,360 --> 00:33:29,120
probability of being a one or a zero

1089
00:33:25,760 --> 00:33:29,919
that's what it means right so if you do

1090
00:33:29,120 --> 00:33:31,439
a hadamard

1091
00:33:29,919 --> 00:33:33,519
rotation and you go into superposition

1092
00:33:31,440 --> 00:33:34,640
that's the same thing as a u2 rotation

1093
00:33:33,519 --> 00:33:36,159
which is why that's what you're seeing

1094
00:33:34,640 --> 00:33:38,080
in the circuit

1095
00:33:36,159 --> 00:33:39,360
quantum computers don't implement the

1096
00:33:38,080 --> 00:33:42,399
theoretical gates

1097
00:33:39,360 --> 00:33:42,879
they always transpile into what the

1098
00:33:42,399 --> 00:33:45,360
gates

1099
00:33:42,880 --> 00:33:46,960
are actually available on the system so

1100
00:33:45,360 --> 00:33:47,279
what i worked out was that actually i

1101
00:33:46,960 --> 00:33:50,480
could

1102
00:33:47,279 --> 00:33:53,120
on the larger machines the ibm q machine

1103
00:33:50,480 --> 00:33:57,039
in melbourne has 15 available qubits

1104
00:33:53,120 --> 00:34:00,199
um and you can run up to 8192

1105
00:33:57,039 --> 00:34:01,360
individual shots which means i can get

1106
00:34:00,200 --> 00:34:05,120
8192

1107
00:34:01,360 --> 00:34:09,040
multiplied by 15 bits of output okay

1108
00:34:05,120 --> 00:34:12,159
and then i can go right i've now got 122

1109
00:34:09,040 --> 00:34:13,918
880 potential random bits now the

1110
00:34:12,159 --> 00:34:15,359
problem is they're not secret to me

1111
00:34:13,918 --> 00:34:17,359
but they're very good high quality

1112
00:34:15,359 --> 00:34:18,560
random so if i blend them with the local

1113
00:34:17,359 --> 00:34:21,520
sources of randomness

1114
00:34:18,560 --> 00:34:21,839
i get something kind of cool so i've

1115
00:34:21,520 --> 00:34:23,839
just

1116
00:34:21,839 --> 00:34:25,679
made a little gist with a little kind of

1117
00:34:23,839 --> 00:34:28,239
python class to show you

1118
00:34:25,679 --> 00:34:29,839
um let me pull up some code like i can

1119
00:34:28,239 --> 00:34:33,040
demonstrate that it actually runs

1120
00:34:29,839 --> 00:34:34,000
it uh if i don't actually do that

1121
00:34:33,040 --> 00:34:36,480
actually run the code

1122
00:34:34,000 --> 00:34:38,079
it helps i suppose uh let me just shrink

1123
00:34:36,480 --> 00:34:39,280
this down so this is it just generally

1124
00:34:38,079 --> 00:34:40,800
generating things

1125
00:34:39,280 --> 00:34:42,560
um you'll see in a second when it gets

1126
00:34:40,800 --> 00:34:43,919
rolling it's gonna add a job

1127
00:34:42,560 --> 00:34:45,679
so it's gonna go oh look i've got all

1128
00:34:43,918 --> 00:34:48,560
these random bits oh i need to

1129
00:34:45,679 --> 00:34:50,000
run a background ibm cue process and if

1130
00:34:48,560 --> 00:34:52,239
i look at the results here

1131
00:34:50,000 --> 00:34:53,040
if i just refresh this page then you can

1132
00:34:52,239 --> 00:34:55,759
see it's

1133
00:34:53,040 --> 00:34:56,560
added uh this job here we go just a few

1134
00:34:55,760 --> 00:34:58,560
seconds ago

1135
00:34:56,560 --> 00:35:00,960
and it's going to run in 15 minutes so

1136
00:34:58,560 --> 00:35:03,920
long after i am off the stage

1137
00:35:00,960 --> 00:35:04,560
long halfway through uh chris's talk um

1138
00:35:03,920 --> 00:35:06,640
but like

1139
00:35:04,560 --> 00:35:07,759
yeah that will actually run and then if

1140
00:35:06,640 --> 00:35:09,680
i leave this running it will see

1141
00:35:07,760 --> 00:35:11,200
directly into it qr code here for those

1142
00:35:09,680 --> 00:35:12,960
who want to play along at home

1143
00:35:11,200 --> 00:35:14,560
what does the code look like it's just

1144
00:35:12,960 --> 00:35:18,400
python so i

1145
00:35:14,560 --> 00:35:22,160
coded in um the blum blum shub real name

1146
00:35:18,400 --> 00:35:24,400
the blum shrub uh csprng and i

1147
00:35:22,160 --> 00:35:25,839
coded it all in a few dirty hacks here

1148
00:35:24,400 --> 00:35:27,440
and there sorry this is not for

1149
00:35:25,839 --> 00:35:28,000
production if you put this in production

1150
00:35:27,440 --> 00:35:30,560
that's a

1151
00:35:28,000 --> 00:35:32,160
silly thing to do but i know that this

1152
00:35:30,560 --> 00:35:33,759
is a good random number generator so i

1153
00:35:32,160 --> 00:35:35,598
am just sort of implementing it kind of

1154
00:35:33,760 --> 00:35:36,960
haphazardly that i have not taken any

1155
00:35:35,599 --> 00:35:39,200
care or consideration

1156
00:35:36,960 --> 00:35:40,880
in this this is all proof of concept to

1157
00:35:39,200 --> 00:35:42,160
show that i can have a background

1158
00:35:40,880 --> 00:35:44,560
process that

1159
00:35:42,160 --> 00:35:46,240
deals with ibm q as you can see here so

1160
00:35:44,560 --> 00:35:49,440
i make the circuit

1161
00:35:46,240 --> 00:35:52,078
i run the circuit i get the bits

1162
00:35:49,440 --> 00:35:52,960
i do some hashing because ketchup three

1163
00:35:52,079 --> 00:35:55,920
is actually

1164
00:35:52,960 --> 00:35:56,960
really really useful um to be able to

1165
00:35:55,920 --> 00:35:58,640
say look

1166
00:35:56,960 --> 00:36:00,800
i take the hp pool which is my local

1167
00:35:58,640 --> 00:36:01,598
secrecy i take the really high quality

1168
00:36:00,800 --> 00:36:04,480
randomness

1169
00:36:01,599 --> 00:36:04,960
quality rounding risk courtesy of ibmq

1170
00:36:04,480 --> 00:36:06,800
and i

1171
00:36:04,960 --> 00:36:08,240
just hash the two together and i just do

1172
00:36:06,800 --> 00:36:08,800
that over and over again for all the

1173
00:36:08,240 --> 00:36:11,040
bits

1174
00:36:08,800 --> 00:36:12,400
i then do another one at the end just

1175
00:36:11,040 --> 00:36:13,920
because i can

1176
00:36:12,400 --> 00:36:15,839
and it adds a little bit more entropy to

1177
00:36:13,920 --> 00:36:17,440
the whole system and then it just keeps

1178
00:36:15,839 --> 00:36:19,279
going round and round all i'm doing here

1179
00:36:17,440 --> 00:36:21,200
is i'm just generating

1180
00:36:19,280 --> 00:36:22,480
billions of bits okay and you can see

1181
00:36:21,200 --> 00:36:24,879
the output here is uh

1182
00:36:22,480 --> 00:36:28,640
throwing out numbers every eight

1183
00:36:24,880 --> 00:36:30,880
thousand 8192 bits i think

1184
00:36:28,640 --> 00:36:32,799
and that's that's that's the code this

1185
00:36:30,880 --> 00:36:33,440
actually works i've had this running on

1186
00:36:32,800 --> 00:36:36,160
a system

1187
00:36:33,440 --> 00:36:36,960
for a couple of days and it ran just

1188
00:36:36,160 --> 00:36:40,000
fine

1189
00:36:36,960 --> 00:36:42,800
absolutely fine no problems no issues

1190
00:36:40,000 --> 00:36:44,480
it just kept going so like this is

1191
00:36:42,800 --> 00:36:46,160
perfectly workable the quantity the ibm

1192
00:36:44,480 --> 00:36:47,599
system works perfectly fine

1193
00:36:46,160 --> 00:36:49,440
if you want to have a look i put loads

1194
00:36:47,599 --> 00:36:51,040
of comments in to explain what i'm doing

1195
00:36:49,440 --> 00:36:52,000
and why and also why you shouldn't use

1196
00:36:51,040 --> 00:36:54,960
this in production

1197
00:36:52,000 --> 00:36:55,440
i cannot stress this enough but like

1198
00:36:54,960 --> 00:36:57,440
yeah

1199
00:36:55,440 --> 00:36:59,760
it works and it's there and if you want

1200
00:36:57,440 --> 00:37:01,680
to have a play there's a worked example

1201
00:36:59,760 --> 00:37:03,520
of how you can use quantum algorithms to

1202
00:37:01,680 --> 00:37:05,919
do something that's kind of nifty

1203
00:37:03,520 --> 00:37:07,040
so i'm gonna sort of put the fin slide

1204
00:37:05,920 --> 00:37:09,359
up and say

1205
00:37:07,040 --> 00:37:10,480
thank you very much i'm gonna end it

1206
00:37:09,359 --> 00:37:13,680
there and i'm gonna

1207
00:37:10,480 --> 00:37:17,119
entertain whatever questions you have

1208
00:37:13,680 --> 00:37:17,118
thank you very much for your time

1209
00:37:17,839 --> 00:37:22,640
all right thank you dr mark clan carney

1210
00:37:20,880 --> 00:37:23,839
uh was awesome uh you got a lot of good

1211
00:37:22,640 --> 00:37:25,598
comments here i'm gonna see if there's

1212
00:37:23,839 --> 00:37:28,078
any questions in them

1213
00:37:25,599 --> 00:37:28,960
uh besides more cow math from when you

1214
00:37:28,079 --> 00:37:32,160
know

1215
00:37:28,960 --> 00:37:33,839
um yeah that goes with the cowbell um

1216
00:37:32,160 --> 00:37:35,759
you know actually cow is the name of i

1217
00:37:33,839 --> 00:37:37,200
think cow is the name of a quantum key

1218
00:37:35,760 --> 00:37:37,839
distribution system i actually think

1219
00:37:37,200 --> 00:37:40,960
that's right

1220
00:37:37,839 --> 00:37:43,680
i'll check that i'll check that then i

1221
00:37:40,960 --> 00:37:46,800
you got one uh give the quantum computer

1222
00:37:43,680 --> 00:37:46,799
a limit to solve then

1223
00:37:47,119 --> 00:37:50,880
why is this right that was actually in

1224
00:37:49,440 --> 00:37:54,400
discord

1225
00:37:50,880 --> 00:37:55,839
i was in disco yeah i don't know uh

1226
00:37:54,400 --> 00:37:57,520
i don't know what part of your speech it

1227
00:37:55,839 --> 00:38:01,279
was at but

1228
00:37:57,520 --> 00:38:02,720
um i don't know i mean quantum computers

1229
00:38:01,280 --> 00:38:04,720
are really useful like

1230
00:38:02,720 --> 00:38:05,919
um i think that's at the moment they're

1231
00:38:04,720 --> 00:38:09,040
small like they are like

1232
00:38:05,920 --> 00:38:10,400
they're sub zx80s right so actually if

1233
00:38:09,040 --> 00:38:10,960
you're gonna start learning how things

1234
00:38:10,400 --> 00:38:12,480
are going

1235
00:38:10,960 --> 00:38:14,160
and learning what to do now is actually

1236
00:38:12,480 --> 00:38:15,440
a pretty good time because kiskit

1237
00:38:14,160 --> 00:38:17,440
documentation is

1238
00:38:15,440 --> 00:38:19,359
thorough if you don't remember your high

1239
00:38:17,440 --> 00:38:21,520
school 2x2 matrices

1240
00:38:19,359 --> 00:38:23,040
they've got a really nice page in line

1241
00:38:21,520 --> 00:38:23,839
for the page you're trying to read and

1242
00:38:23,040 --> 00:38:26,160
understand

1243
00:38:23,839 --> 00:38:28,160
it's all there so actually kiskit's got

1244
00:38:26,160 --> 00:38:30,720
a lot of really good documentation

1245
00:38:28,160 --> 00:38:32,000
um and i think that actually you can

1246
00:38:30,720 --> 00:38:33,598
have a lot of uh

1247
00:38:32,000 --> 00:38:34,720
uh fun with these things just like

1248
00:38:33,599 --> 00:38:37,680
playing around with things actually

1249
00:38:34,720 --> 00:38:40,560
having a 16 cubic well 15 cubic computer

1250
00:38:37,680 --> 00:38:42,960
from um from ibm available it's kind of

1251
00:38:40,560 --> 00:38:46,720
really handy you can do shaw's algorithm

1252
00:38:42,960 --> 00:38:50,640
to factor the number 17.

1253
00:38:46,720 --> 00:38:51,919
[Laughter]

1254
00:38:50,640 --> 00:38:52,960
i guess there's no other questions i'll

1255
00:38:51,920 --> 00:38:54,640
give them like a couple more cents for

1256
00:38:52,960 --> 00:38:56,079
anyone answering the questions or if

1257
00:38:54,640 --> 00:38:58,400
when has another comment he wants to

1258
00:38:56,079 --> 00:39:00,720
make i'm sure he'll i

1259
00:38:58,400 --> 00:39:03,200
i'm sure we just we just listened on the

1260
00:39:00,720 --> 00:39:07,919
wind and uh we will uh we will hear him

1261
00:39:03,200 --> 00:39:11,040
uh forthwith we got one uh can you share

1262
00:39:07,920 --> 00:39:13,599
quantum break timeline again

1263
00:39:11,040 --> 00:39:15,119
quantum break timeline yes let me find

1264
00:39:13,599 --> 00:39:18,560
that slide for you

1265
00:39:15,119 --> 00:39:20,480
so uh fundamentally uh

1266
00:39:18,560 --> 00:39:21,680
the assumptions i've made here if i get

1267
00:39:20,480 --> 00:39:23,280
rid of the q a

1268
00:39:21,680 --> 00:39:24,799
um the assumptions i've made here are

1269
00:39:23,280 --> 00:39:26,560
that we currently have about two to the

1270
00:39:24,800 --> 00:39:27,760
six qubits in one machine that's about

1271
00:39:26,560 --> 00:39:31,759
we got 53

1272
00:39:27,760 --> 00:39:33,520
so you know kind of nearby um

1273
00:39:31,760 --> 00:39:34,960
if we assume doubling every year so it's

1274
00:39:33,520 --> 00:39:36,480
gonna go from two to the six then next

1275
00:39:34,960 --> 00:39:38,000
year two to the seven then two to the

1276
00:39:36,480 --> 00:39:39,359
eight and two to the nine okay it's just

1277
00:39:38,000 --> 00:39:41,839
going to go up iteratively

1278
00:39:39,359 --> 00:39:43,759
so like a really really fast moore's law

1279
00:39:41,839 --> 00:39:47,279
because moore's law was 18 months

1280
00:39:43,760 --> 00:39:48,160
wasn't it so like we know the thresholds

1281
00:39:47,280 --> 00:39:50,880
are breaking

1282
00:39:48,160 --> 00:39:51,440
like rsa so we know how many qubits we

1283
00:39:50,880 --> 00:39:53,119
need

1284
00:39:51,440 --> 00:39:55,280
so like you just keep multiplying until

1285
00:39:53,119 --> 00:39:58,240
you've got enough okay and then you go

1286
00:39:55,280 --> 00:40:00,320
it's at least at least seven years it's

1287
00:39:58,240 --> 00:40:02,399
at least eight years but that's making

1288
00:40:00,320 --> 00:40:03,599
huge assumptions about engineering and

1289
00:40:02,400 --> 00:40:06,400
physics

1290
00:40:03,599 --> 00:40:07,599
realistically you're gonna be talking 10

1291
00:40:06,400 --> 00:40:10,160
to 15 years

1292
00:40:07,599 --> 00:40:11,680
but like you know like the iphone is 15

1293
00:40:10,160 --> 00:40:14,879
years old next year

1294
00:40:11,680 --> 00:40:16,078
okay like it's it's it it'll go quite

1295
00:40:14,880 --> 00:40:17,520
quickly i think

1296
00:40:16,079 --> 00:40:18,800
uh there's a lot of money being plumbed

1297
00:40:17,520 --> 00:40:19,520
into this there's a lot of people

1298
00:40:18,800 --> 00:40:21,920
looking at this

1299
00:40:19,520 --> 00:40:23,280
are these problems so i think there's a

1300
00:40:21,920 --> 00:40:24,720
high chance that we might see things

1301
00:40:23,280 --> 00:40:26,000
some things a little sooner some things

1302
00:40:24,720 --> 00:40:27,040
a little later we'll just have to see

1303
00:40:26,000 --> 00:40:28,400
how it goes but there's

1304
00:40:27,040 --> 00:40:30,720
there's the slide i think that people

1305
00:40:28,400 --> 00:40:32,720
are after all right great

1306
00:40:30,720 --> 00:40:33,759
i don't see any more questions in any

1307
00:40:32,720 --> 00:40:35,598
place

1308
00:40:33,760 --> 00:40:36,880
so i think you're all set thank you so

1309
00:40:35,599 --> 00:40:38,640
much mark

1310
00:40:36,880 --> 00:40:40,240
thank you very much corncon i've really

1311
00:40:38,640 --> 00:40:41,359
enjoyed it take care yeah hopefully see

1312
00:40:40,240 --> 00:40:44,479
you next year

1313
00:40:41,359 --> 00:40:46,880
absolutely all right

1314
00:40:44,480 --> 00:40:49,040
um next up at uh 3 30 we're gonna have

1315
00:40:46,880 --> 00:40:50,880
chris roberts he's a geek hacker and

1316
00:40:49,040 --> 00:40:52,000
he's at hillbilly hit squad he's gonna

1317
00:40:50,880 --> 00:40:55,280
talk to you about all your

1318
00:40:52,000 --> 00:40:55,920
votes are belong to i should be good

1319
00:40:55,280 --> 00:41:07,839
talk

1320
00:40:55,920 --> 00:41:07,839
talk to you soon bye guys

1321
00:41:13,119 --> 00:41:15,200
you