1
00:00:00,120 --> 00:00:04,240
Speaker 1: Welcome to tech Stuff, a production from iHeartRadio. This season,

2
00:00:04,280 --> 00:00:07,880
Speaker 1: non Smart Talks with IBM, Malcolm Glabwell is back, and

3
00:00:07,920 --> 00:00:10,600
Speaker 1: this time he's taking the show on the road. Malcolm

4
00:00:10,640 --> 00:00:14,720
Speaker 1: is stepping outside the studio to explore how IBM clients

5
00:00:14,760 --> 00:00:18,759
Speaker 1: are using artificial intelligence to solve real world challenges and

6
00:00:18,840 --> 00:00:23,000
Speaker 1: transform the way they do business. From accelerating scientific breakthroughs

7
00:00:23,079 --> 00:00:27,520
Speaker 1: to reimagining education. It's a fresh look at innovation in action,

8
00:00:27,960 --> 00:00:31,720
Speaker 1: where big ideas meet cutting edge solutions. You'll hear from

9
00:00:31,720 --> 00:00:36,000
Speaker 1: industry leaders, creative thinkers, and of course Malcolm Glabwell himself

10
00:00:36,280 --> 00:00:39,879
Speaker 1: as he guides you through each story. New episodes of

11
00:00:39,920 --> 00:00:43,280
Speaker 1: Smart Talks with IBM drop every month on the iHeartRadio app,

12
00:00:43,440 --> 00:00:47,240
Speaker 1: Apple Podcasts, or wherever you get your podcasts. Learn more

13
00:00:47,280 --> 00:00:49,920
Speaker 1: at IBM dot com slash smart Talks.

14
00:00:53,120 --> 00:00:55,680
Speaker 2: Hello, this is Malcolm Gladwell and you're listening to Smart

15
00:00:55,680 --> 00:00:59,160
Speaker 2: Talks with IBM. Every year, Tech Week brings thousands of

16
00:00:59,200 --> 00:01:02,600
Speaker 2: people together to network and learn about what's emerging across

17
00:01:02,640 --> 00:01:07,000
Speaker 2: the technology ecosystem, and at this year's conference in San Francisco,

18
00:01:07,400 --> 00:01:10,080
Speaker 2: I had an amazing opportunity to sit down in front

19
00:01:10,120 --> 00:01:13,120
Speaker 2: of a live audience with Jay Gambetta. Jay has been

20
00:01:13,160 --> 00:01:16,480
Speaker 2: with IBM for years and was recently promoted to Director

21
00:01:16,520 --> 00:01:20,240
Speaker 2: of Research. In this job, Jay has an important mission

22
00:01:20,680 --> 00:01:24,040
Speaker 2: helping the company build the future of computing. In the

23
00:01:24,120 --> 00:01:26,720
Speaker 2: last episode of Smart Talks, I began to learn about

24
00:01:26,800 --> 00:01:31,640
Speaker 2: quantum computing from IBM Chairman and CEO Arvind Krishna. But

25
00:01:31,760 --> 00:01:35,160
Speaker 2: this conversation I had with Jay went even deeper and

26
00:01:35,360 --> 00:01:38,960
Speaker 2: convinced me that the development of quantum isn't just a fun,

27
00:01:39,360 --> 00:01:42,520
Speaker 2: exciting new paradigm of computing. It may be one of

28
00:01:42,520 --> 00:01:52,920
Speaker 2: the most important scientific achievements of my lifetime. Jay, Good morning, morning,

29
00:01:53,240 --> 00:01:56,320
Speaker 2: Welcome to Smart Talks with IBM. Thank you special live

30
00:01:56,360 --> 00:01:59,840
Speaker 2: recording here for tech Week and congratulations. How long have

31
00:01:59,880 --> 00:02:03,560
Speaker 2: you been Head of Research at IBM Since October one?

32
00:02:03,560 --> 00:02:07,480
Speaker 2: It's October tenth today, so nine days, nine days. Can

33
00:02:07,520 --> 00:02:10,160
Speaker 2: you just talk a little about the position. This is

34
00:02:10,440 --> 00:02:14,640
Speaker 2: one of the most important positions in research in the world.

35
00:02:15,200 --> 00:02:18,280
Speaker 3: IBM Research has been around for eighty years and it's

36
00:02:18,320 --> 00:02:23,359
Speaker 3: done some tremendous technology, a lot of inventions and fundamentals

37
00:02:23,360 --> 00:02:25,560
Speaker 3: for semiconductors, algorithms.

38
00:02:26,240 --> 00:02:26,560
Speaker 4: AI.

39
00:02:27,160 --> 00:02:29,200
Speaker 3: Yeah, I think if we look back to where a

40
00:02:29,280 --> 00:02:31,600
Speaker 3: lot of the innovation and the technology of the world

41
00:02:31,680 --> 00:02:35,000
Speaker 3: comes from. I think you can find Ibram's footprints on it,

42
00:02:35,080 --> 00:02:38,320
Speaker 3: and you can find IBM research. So yeah, I'm very

43
00:02:38,320 --> 00:02:41,200
Speaker 3: excited for the opportunity, but I'm also aware that there's

44
00:02:41,240 --> 00:02:44,280
Speaker 3: big shoes to fill, and I'm looking forward to how

45
00:02:44,360 --> 00:02:48,120
Speaker 3: we take IBM research forward. Obviously, I'm going to be

46
00:02:48,919 --> 00:02:50,880
Speaker 3: bringing a lot of the quantum side, which we're going

47
00:02:50,919 --> 00:02:54,840
Speaker 3: to talk about later. Beyond quantum, there's important work that

48
00:02:54,880 --> 00:02:58,600
Speaker 3: needs to happen in AI hybrid cloud, and I think

49
00:02:58,800 --> 00:03:01,200
Speaker 3: we're going to also enter to this new period of

50
00:03:01,440 --> 00:03:05,440
Speaker 3: mathematics where we get to use quantum machines and also

51
00:03:05,960 --> 00:03:09,800
Speaker 3: AI machines, and there's some really good, hard mathematical questions

52
00:03:09,840 --> 00:03:10,360
Speaker 3: to answer.

53
00:03:10,480 --> 00:03:12,080
Speaker 2: How many people do you have working for you?

54
00:03:12,480 --> 00:03:16,200
Speaker 3: I mean researchers in the three thousand researchers across many

55
00:03:16,200 --> 00:03:19,480
Speaker 3: different labs around the world. Our main lab is in Yorktown,

56
00:03:19,560 --> 00:03:21,640
Speaker 3: but then we have the lab actually out on the

57
00:03:21,720 --> 00:03:25,120
Speaker 3: West coast in Armadan or sbl now, and then we

58
00:03:25,200 --> 00:03:28,520
Speaker 3: have one in Zurich, Japan, and a few others around

59
00:03:28,560 --> 00:03:28,919
Speaker 3: the world.

60
00:03:29,560 --> 00:03:32,040
Speaker 2: Tell me a little bit before we get into quantum.

61
00:03:32,200 --> 00:03:35,480
Speaker 2: I'm just curious about your path. So you're Australian. Yep,

62
00:03:35,880 --> 00:03:39,280
Speaker 2: we were talking about earlier backstage. Your accent has become muted.

63
00:03:39,880 --> 00:03:42,560
Speaker 2: You should crank it up because it's.

64
00:03:42,520 --> 00:03:45,320
Speaker 3: Yeah, I'm slowly losing my Australian accent. I've been in

65
00:03:45,360 --> 00:03:48,360
Speaker 3: the US since two thousand and four, so accent, you know,

66
00:03:48,440 --> 00:03:51,200
Speaker 3: to sound very Australian. Yeah, but how do you practice it?

67
00:03:51,240 --> 00:03:54,280
Speaker 3: Maybe I got to go back to Australia. Here a

68
00:03:54,440 --> 00:03:57,520
Speaker 3: more Australians say gooday, how's it going? Things like that.

69
00:03:58,480 --> 00:04:01,120
Speaker 2: And you didn't grow up thinking you're going to be

70
00:04:01,160 --> 00:04:02,480
Speaker 2: a scientist one day. Now.

71
00:04:02,600 --> 00:04:06,600
Speaker 3: I grew up in a pretty normal life. My dreams

72
00:04:06,840 --> 00:04:09,520
Speaker 3: as a kid was building things, so I was either

73
00:04:09,560 --> 00:04:12,600
Speaker 3: going to be a carpenter or a mechanic. But I

74
00:04:12,640 --> 00:04:15,560
Speaker 3: had some great teachers that inspired me to go to university.

75
00:04:15,640 --> 00:04:18,080
Speaker 3: And I didn't even know, honestly what a scientist was.

76
00:04:18,400 --> 00:04:22,919
Speaker 3: And then I found myself at university doing science, particular physics,

77
00:04:23,080 --> 00:04:25,200
Speaker 3: and I ended up loving it. So you go from

78
00:04:25,240 --> 00:04:28,479
Speaker 3: there to what do you do your PhD? So I

79
00:04:28,480 --> 00:04:31,440
Speaker 3: did my undergrad in Australia. I did it actually in

80
00:04:31,680 --> 00:04:38,040
Speaker 3: laser science, so I think I watched some TV show

81
00:04:38,080 --> 00:04:40,960
Speaker 3: in lasers seemed interesting, so I wanted to learn about lasers,

82
00:04:41,520 --> 00:04:44,599
Speaker 3: and then I realized in trying to understand lasers, there

83
00:04:44,680 --> 00:04:48,000
Speaker 3: was this quantum mechanics, and so I was like, all right,

84
00:04:48,800 --> 00:04:51,279
Speaker 3: I want to actually understand this quantum mechanics. So I

85
00:04:51,279 --> 00:04:53,880
Speaker 3: did my equivalent of what you and the US school masters.

86
00:04:54,240 --> 00:04:56,159
Speaker 3: We call it honors in Australia, but we do a

87
00:04:56,200 --> 00:04:59,680
Speaker 3: research project. I said, I wanted to shoot lasers into

88
00:04:59,680 --> 00:05:03,440
Speaker 3: Adam and measure cross sections and I got really into

89
00:05:03,520 --> 00:05:06,080
Speaker 3: quantum physics. So then I decided, all right, I don't

90
00:05:06,160 --> 00:05:10,159
Speaker 3: understand this quantum physics. I want to do my PhD

91
00:05:10,279 --> 00:05:13,880
Speaker 3: in Interpretations of quantum mechanics. So I jumped in and said,

92
00:05:13,880 --> 00:05:16,640
Speaker 3: all right, what is this quantum mechanics? Why is everyone

93
00:05:16,800 --> 00:05:20,360
Speaker 3: arguing on these different interpretations. Then I finished my PhD

94
00:05:20,400 --> 00:05:24,000
Speaker 3: in Australia doing that. Then I moved over. At the

95
00:05:24,120 --> 00:05:29,239
Speaker 3: end of my PhD interpretations, it's more people arguing about

96
00:05:29,240 --> 00:05:33,360
Speaker 3: the equations whilst I think it's really important. I decided

97
00:05:33,440 --> 00:05:36,040
Speaker 3: if it's going to be like a collapse equation versus

98
00:05:36,040 --> 00:05:40,640
Speaker 3: many worlds, or a hidden variable model, or that just

99
00:05:40,720 --> 00:05:44,919
Speaker 3: quantum mechanics decoheres because we don't see supersitions in the

100
00:05:45,040 --> 00:05:49,279
Speaker 3: everyday world because it interacts with environment. The only way

101
00:05:49,320 --> 00:05:51,560
Speaker 3: to answer that question was to build a quantum computer.

102
00:05:52,279 --> 00:05:54,600
Speaker 3: And so then I decided at the end of my PhD,

103
00:05:54,680 --> 00:05:57,520
Speaker 3: I wanted to work out how to build a quantum computer.

104
00:05:58,080 --> 00:06:00,760
Speaker 3: And then I left there and I went to Yale.

105
00:06:01,200 --> 00:06:04,440
Speaker 3: And then at Yale, that's where I got into superconducting cubits,

106
00:06:04,800 --> 00:06:07,120
Speaker 3: which just a few days ago, one of the professors

107
00:06:07,120 --> 00:06:09,560
Speaker 3: there just won the Nobel Prize this year.

108
00:06:09,920 --> 00:06:13,400
Speaker 2: Oh wow, I'm very interested in tracing because your career

109
00:06:14,640 --> 00:06:17,800
Speaker 2: follows the arc of quantum computing in a certain way.

110
00:06:17,920 --> 00:06:20,960
Speaker 2: Right at the time when you asked the question, what

111
00:06:21,000 --> 00:06:22,640
Speaker 2: I really want to do is to figure out how

112
00:06:22,640 --> 00:06:25,560
Speaker 2: to build a quantum computer. Where are we in quantum

113
00:06:25,640 --> 00:06:26,839
Speaker 2: computing at that point?

114
00:06:27,000 --> 00:06:29,880
Speaker 3: Yeah, So that would have been nineteen ninety So there

115
00:06:29,920 --> 00:06:33,760
Speaker 3: was Shaw's algorithm came out, let's say ninety five. There

116
00:06:33,839 --> 00:06:36,800
Speaker 3: was a lot of theory. And then the reason I

117
00:06:36,839 --> 00:06:41,040
Speaker 3: went to Yale is because people had started to show

118
00:06:41,120 --> 00:06:45,800
Speaker 3: that they could see quantum effects in electrical circuits. So

119
00:06:45,920 --> 00:06:49,240
Speaker 3: these macroscopic objects they were starting to behave quantum mechanical

120
00:06:49,760 --> 00:06:52,800
Speaker 3: There was a really significant breakthrough in nineteen ninety nine

121
00:06:52,839 --> 00:06:57,000
Speaker 3: where Yazoo Nakamura in Japan showed that a qubit could

122
00:06:57,080 --> 00:07:00,960
Speaker 3: exist in these electrical circuits. I found out the group

123
00:07:01,000 --> 00:07:03,880
Speaker 3: at Yale were really trying to take these electrical circuits

124
00:07:04,240 --> 00:07:07,960
Speaker 3: and couple them together. And so it was like, if

125
00:07:08,000 --> 00:07:11,840
Speaker 3: I can build something using electrical circuits and they're big,

126
00:07:12,400 --> 00:07:14,720
Speaker 3: that that's the best way that you can decide to

127
00:07:14,800 --> 00:07:18,240
Speaker 3: test and understand whether quantum mechanics breaks down at a

128
00:07:18,240 --> 00:07:21,360
Speaker 3: macroscopic scale or not. Can we actually make them behave

129
00:07:21,360 --> 00:07:23,880
Speaker 3: as cubits? And I agree When I came to Yale,

130
00:07:23,920 --> 00:07:28,120
Speaker 3: the cubits were not very good. They were actually a

131
00:07:28,200 --> 00:07:32,960
Speaker 3: couple of nanoseconds. They were unstable. Electron would jump onto

132
00:07:33,040 --> 00:07:36,360
Speaker 3: the chip and then they would change all their configurations,

133
00:07:36,360 --> 00:07:39,480
Speaker 3: so you have to restart your experiment. And so for

134
00:07:39,520 --> 00:07:41,800
Speaker 3: the first time at Yale, it's kind of what the

135
00:07:41,880 --> 00:07:44,760
Speaker 3: challenge there was, how do we make a cubit? How

136
00:07:44,760 --> 00:07:47,560
Speaker 3: do we make a stable cubit? And that took about

137
00:07:47,680 --> 00:07:50,120
Speaker 3: five years, and that took us up to two thousand

138
00:07:50,160 --> 00:07:53,160
Speaker 3: and seven. And I think the rest of the world

139
00:07:53,360 --> 00:07:56,240
Speaker 3: looks and says quantums like just blowing up, but it's

140
00:07:56,280 --> 00:08:00,920
Speaker 3: actually been like almost phases theory showing that wet the algorithms,

141
00:08:01,240 --> 00:08:03,440
Speaker 3: how do we make a cubit? How do we couple

142
00:08:03,440 --> 00:08:06,240
Speaker 3: of the cubits together? And now we're in the scaling phase.

143
00:08:06,720 --> 00:08:11,200
Speaker 2: Describe for us because many people in this room, me included,

144
00:08:11,720 --> 00:08:15,000
Speaker 2: have only a kind of surface level understanding of what

145
00:08:15,040 --> 00:08:17,280
Speaker 2: we mean when we use that phrase. What is the

146
00:08:17,320 --> 00:08:21,040
Speaker 2: difference between classical computing and quantum computing? What does that

147
00:08:21,080 --> 00:08:21,560
Speaker 2: word mean?

148
00:08:21,960 --> 00:08:25,280
Speaker 3: Yeah, so you can go down the physics way and

149
00:08:25,360 --> 00:08:28,280
Speaker 3: talk about supersition and entanglement, which we can go in later,

150
00:08:28,360 --> 00:08:31,600
Speaker 3: but actually feel it's a bit of a distraction. So

151
00:08:31,920 --> 00:08:36,080
Speaker 3: when you think of classical computers, what they were is

152
00:08:36,120 --> 00:08:39,480
Speaker 3: there were machines that were very good at adding numbers together,

153
00:08:40,000 --> 00:08:43,880
Speaker 3: like simple addition, and they really showed that they could

154
00:08:44,400 --> 00:08:47,600
Speaker 3: add these numbers together really really fast. And now with

155
00:08:47,920 --> 00:08:51,440
Speaker 3: GPUs and other AI accelerators, we can add those numbers

156
00:08:51,440 --> 00:08:55,360
Speaker 3: together in parallel, and so the whole classical computing can

157
00:08:55,360 --> 00:08:59,160
Speaker 3: come down to just arithmetic, just adding numbers together. It

158
00:08:59,240 --> 00:09:03,520
Speaker 3: turns out that there's a math that is the quantum

159
00:09:03,559 --> 00:09:06,640
Speaker 3: mechanics shown to be true. It's more like a group

160
00:09:06,720 --> 00:09:10,280
Speaker 3: theory type structure, and the way quantum works is it

161
00:09:10,280 --> 00:09:12,880
Speaker 3: has a different math as are primitive, and if we

162
00:09:12,920 --> 00:09:15,800
Speaker 3: can exploit that new math and build a machine that

163
00:09:15,880 --> 00:09:19,120
Speaker 3: does it, it allows us to answer different questions. And

164
00:09:19,200 --> 00:09:22,640
Speaker 3: so think of it as a branching from classical compute

165
00:09:22,679 --> 00:09:25,440
Speaker 3: that is very good at adding just numbers together to

166
00:09:25,559 --> 00:09:28,880
Speaker 3: something that allows us to work with an algebra that

167
00:09:29,440 --> 00:09:33,120
Speaker 3: is much much harder to represent with addition. And that

168
00:09:33,200 --> 00:09:36,320
Speaker 3: algebra happens to be the same algebra that defines the

169
00:09:36,360 --> 00:09:39,920
Speaker 3: fundamental equations of nature, shirting this equation. So this is

170
00:09:39,960 --> 00:09:42,439
Speaker 3: why you say it computes the same way nature does,

171
00:09:42,760 --> 00:09:46,040
Speaker 3: but there are many other interesting problems. So the way

172
00:09:46,080 --> 00:09:48,560
Speaker 3: I explain it to people is think of it as

173
00:09:49,080 --> 00:09:54,319
Speaker 3: bringing a new primitive to computer science and allowing us

174
00:09:54,760 --> 00:09:57,440
Speaker 3: to work how to go with it. And I like

175
00:09:57,440 --> 00:10:01,360
Speaker 3: the analogy. Well, actually, maybe go back. So if you

176
00:10:01,400 --> 00:10:04,000
Speaker 3: went back in time, so we're one hundred years of quantum,

177
00:10:04,400 --> 00:10:06,480
Speaker 3: and you went back in time and you asked, what

178
00:10:06,640 --> 00:10:09,920
Speaker 3: is the foundation is a chemistry or physics? What would

179
00:10:09,920 --> 00:10:12,240
Speaker 3: have probably the scientists of one hundred years ago would

180
00:10:12,240 --> 00:10:14,520
Speaker 3: have said is they would have said, you know, chemistry

181
00:10:14,600 --> 00:10:17,679
Speaker 3: is about the small, physics is about planets and things

182
00:10:17,720 --> 00:10:23,120
Speaker 3: like this, and one hundred years ago when Heisenberg or Einstein,

183
00:10:23,200 --> 00:10:27,760
Speaker 3: all the greats, Schroding her himself invented quantum mechanics. It

184
00:10:27,800 --> 00:10:31,280
Speaker 3: was this concept that nature is discrete, not continuous. It

185
00:10:31,320 --> 00:10:34,880
Speaker 3: actually brought all the physical sciences together. And now quantum

186
00:10:34,920 --> 00:10:38,600
Speaker 3: mechanics is like it is the foundation of the science.

187
00:10:39,440 --> 00:10:42,480
Speaker 3: And so now what quantum computing is by that analogy

188
00:10:42,600 --> 00:10:45,520
Speaker 3: is computer science. The foundation of the math is coming

189
00:10:45,520 --> 00:10:48,959
Speaker 3: together with the physical science to allow us to compute

190
00:10:49,440 --> 00:10:52,360
Speaker 3: using math that if you were to try to represent

191
00:10:52,400 --> 00:10:55,840
Speaker 3: it with classical computers, it takes exponential time.

192
00:10:56,320 --> 00:11:00,120
Speaker 2: Yeah, and it was a classical computer an expence in

193
00:11:00,120 --> 00:11:02,360
Speaker 2: a way that someone is well informed as I am

194
00:11:02,400 --> 00:11:05,960
Speaker 2: can understand it. A customer computer. It works primarily on

195
00:11:06,040 --> 00:11:09,880
Speaker 2: problems that can be easily represented in numerical form in numbers. Yes,

196
00:11:10,200 --> 00:11:12,920
Speaker 2: quantum allows you to step outside to a class of

197
00:11:12,920 --> 00:11:17,840
Speaker 2: problems that don't necessarily have a simple numerical representation. Yeah.

198
00:11:17,880 --> 00:11:22,560
Speaker 3: And so imagine I got some medicine or or some

199
00:11:22,840 --> 00:11:25,640
Speaker 3: set of operation, but call it A, and I then

200
00:11:25,720 --> 00:11:29,319
Speaker 3: follow it by a different operation B. If A followed

201
00:11:29,320 --> 00:11:32,680
Speaker 3: by B gave a different answer than B first followed

202
00:11:32,679 --> 00:11:35,760
Speaker 3: by A. So in mathematics we call that commuting. But

203
00:11:35,960 --> 00:11:38,600
Speaker 3: like you can think of a correlation there one one

204
00:11:38,640 --> 00:11:41,720
Speaker 3: gives you a different outcome to the other. That means

205
00:11:41,760 --> 00:11:46,120
Speaker 3: there's an algebra behind it that Representing that algebra traditionally

206
00:11:46,120 --> 00:11:50,040
Speaker 3: on classical computers is really really hard, whereas that algebra,

207
00:11:50,400 --> 00:11:52,800
Speaker 3: if we can get creative, we can come up with

208
00:11:52,880 --> 00:11:55,720
Speaker 3: ways of representing that math. So we step as you say,

209
00:11:56,000 --> 00:11:59,720
Speaker 3: we step out aside of the simple math to a

210
00:11:59,720 --> 00:12:03,200
Speaker 3: new to allow us to calculate interesting problems.

211
00:12:03,640 --> 00:12:07,320
Speaker 2: So quite in a sense, compliments, it doesn't replace judicial.

212
00:12:08,520 --> 00:12:10,280
Speaker 3: I think this is one of the this is you're

213
00:12:10,440 --> 00:12:13,200
Speaker 3: exactly on is people think quantum is going to be

214
00:12:13,280 --> 00:12:16,360
Speaker 3: replacing classical If your problem is good at adding numbers together,

215
00:12:16,640 --> 00:12:20,080
Speaker 3: you should just keep using classical computers. I think the

216
00:12:20,120 --> 00:12:23,440
Speaker 3: future is going to be heterogeneous accelerators, and it will

217
00:12:23,480 --> 00:12:26,600
Speaker 3: definitely have quantum as one. But in some sense, the

218
00:12:26,640 --> 00:12:29,800
Speaker 3: next generation of superstars are going to be those applied

219
00:12:29,840 --> 00:12:33,760
Speaker 3: mathematicians that know, how do I write a problem using

220
00:12:33,800 --> 00:12:37,199
Speaker 3: the simple math of classical computers or the more complicated

221
00:12:37,280 --> 00:12:40,920
Speaker 3: math for quantum computers, and how do I actually iterate

222
00:12:41,000 --> 00:12:43,720
Speaker 3: between them? And things like this. This is where I

223
00:12:43,720 --> 00:12:46,520
Speaker 3: think the next generation of students are going to come

224
00:12:46,600 --> 00:12:48,520
Speaker 3: up with much more novel ideas. I can give you

225
00:12:48,559 --> 00:12:50,640
Speaker 3: examples of what we want to do on quantum, but

226
00:12:50,920 --> 00:12:56,320
Speaker 3: like you're giving them a fundamental, foundational new thing, and

227
00:12:56,400 --> 00:12:59,520
Speaker 3: so I'm optimistic they will do much better jobs than

228
00:12:59,800 --> 00:13:00,520
Speaker 3: my generation.

229
00:13:00,600 --> 00:13:03,120
Speaker 2: Well, yeah, we're to get to some of the albums

230
00:13:03,160 --> 00:13:05,600
Speaker 2: in a moment. But I wanted you to the most

231
00:13:05,720 --> 00:13:08,560
Speaker 2: kind of down that you said as a kid, you

232
00:13:08,679 --> 00:13:10,319
Speaker 2: thought you might want to be a mechanic because you'd

233
00:13:10,320 --> 00:13:14,040
Speaker 2: like to build things. Describe to me what it takes

234
00:13:14,080 --> 00:13:17,400
Speaker 2: to build a quantum computer, Like, what are you doing

235
00:13:17,480 --> 00:13:19,320
Speaker 2: that's different from building a classical computer.

236
00:13:19,600 --> 00:13:22,199
Speaker 3: Yeah, so maybe I'll give you analogy and then I'll

237
00:13:22,200 --> 00:13:25,720
Speaker 3: go in. So the way classical computers, we've got them

238
00:13:25,800 --> 00:13:30,920
Speaker 3: to get to smaller and smaller sizes like five seven animeters,

239
00:13:30,960 --> 00:13:35,920
Speaker 3: five animeters and things, is actually inventing material to kill

240
00:13:36,000 --> 00:13:40,600
Speaker 3: quantum effects. So you actually put dielectrics and other things

241
00:13:40,679 --> 00:13:43,640
Speaker 3: in there to kill the quantum tunneling effects, and you

242
00:13:43,720 --> 00:13:48,560
Speaker 3: want them to behave more classically in the quantum world,

243
00:13:48,679 --> 00:13:50,920
Speaker 3: you want to get rid of all the classical effects.

244
00:13:51,360 --> 00:13:53,760
Speaker 3: So you want to get rid of the ability of

245
00:13:53,800 --> 00:13:56,680
Speaker 3: the cubits to interact with the environment, and in the

246
00:13:56,920 --> 00:13:59,080
Speaker 3: in the sort of technical world, we call it this

247
00:13:59,360 --> 00:14:02,360
Speaker 3: quantum com The more ways you want to control the

248
00:14:02,440 --> 00:14:06,640
Speaker 3: quantum computer, you open it up to interacting with everything else,

249
00:14:06,840 --> 00:14:10,160
Speaker 3: like interacting with its environment. So the biggest challenge has

250
00:14:10,200 --> 00:14:14,760
Speaker 3: always been how do we give more control but don't

251
00:14:14,800 --> 00:14:18,240
Speaker 3: bring in other sources of noise. So I want to

252
00:14:18,280 --> 00:14:21,320
Speaker 3: be able to do gates on the cubit, but I

253
00:14:21,360 --> 00:14:24,720
Speaker 3: don't want it to decohere. I want to couple the cubits,

254
00:14:25,000 --> 00:14:27,320
Speaker 3: but I don't want them to couple to other things.

255
00:14:27,800 --> 00:14:32,080
Speaker 3: So the hardest challenge is the energy inside the cubits

256
00:14:32,240 --> 00:14:34,440
Speaker 3: is a nine gigahertz, and if your tames that by

257
00:14:34,600 --> 00:14:38,400
Speaker 3: HBO ten to the niggave thirty four with nine, you're

258
00:14:38,440 --> 00:14:41,600
Speaker 3: at a tender the negative twenty like three or something

259
00:14:41,680 --> 00:14:45,080
Speaker 3: in energy. That's a tiny amount of energy. So you're

260
00:14:45,120 --> 00:14:48,280
Speaker 3: trying to have a tiny, tiny amount of energy to control,

261
00:14:49,080 --> 00:14:52,120
Speaker 3: and you don't want that to interact with anything. So

262
00:14:52,160 --> 00:14:54,800
Speaker 3: you have to cool them down, you have to isolate them,

263
00:14:55,080 --> 00:14:58,320
Speaker 3: and you have to make the quantum effects dominate over

264
00:14:58,360 --> 00:14:59,440
Speaker 3: the classical effects.

265
00:15:00,200 --> 00:15:03,920
Speaker 2: So practically, if I'm trying to do that. Right now,

266
00:15:04,040 --> 00:15:05,080
Speaker 2: how big are these machines?

267
00:15:05,440 --> 00:15:07,840
Speaker 3: So the cubits themselves are not that big, So the

268
00:15:07,920 --> 00:15:12,560
Speaker 3: cubits themselves are like a few microns. But yeah, most

269
00:15:12,600 --> 00:15:14,960
Speaker 3: of the size so you can see some of our

270
00:15:15,200 --> 00:15:17,200
Speaker 3: I got a pleasure of showing you around to one

271
00:15:17,240 --> 00:15:19,400
Speaker 3: of the machines in Yorktown. You saw that they're like

272
00:15:20,000 --> 00:15:22,880
Speaker 3: twenty foot by twenty foot in size. Most of that

273
00:15:23,480 --> 00:15:27,760
Speaker 3: is all that equipment to isolate the cubit chip, which

274
00:15:27,840 --> 00:15:30,440
Speaker 3: is only a few millimeters when you put it together

275
00:15:30,840 --> 00:15:34,320
Speaker 3: from the rest of the environment. We will, as we

276
00:15:34,400 --> 00:15:38,000
Speaker 3: get better at that, miniaturize all the isolation. But that's

277
00:15:38,200 --> 00:15:41,760
Speaker 3: cooling it down to a few milli calvin, so about

278
00:15:41,760 --> 00:15:45,520
Speaker 3: a thousand times colder than outer space. It's isolating the

279
00:15:45,720 --> 00:15:49,280
Speaker 3: noise on any electrical signal so that no noise from

280
00:15:49,320 --> 00:15:52,680
Speaker 3: the outside world gets into the system. And so that's

281
00:15:52,680 --> 00:15:56,120
Speaker 3: a lot of isolators, filters, and things like that that

282
00:15:56,160 --> 00:15:58,960
Speaker 3: we've had to invent to allow us to make the

283
00:15:59,040 --> 00:16:00,720
Speaker 3: quantum properties of the chip go.

284
00:16:01,160 --> 00:16:03,840
Speaker 2: It's like the Princess and the peak, mounds and mounds

285
00:16:03,880 --> 00:16:07,480
Speaker 2: and mounds of mattresses trying to isolate the impact of

286
00:16:07,560 --> 00:16:10,200
Speaker 2: this little thing and that maybe that's the best way

287
00:16:10,240 --> 00:16:10,840
Speaker 2: to describe it.

288
00:16:10,880 --> 00:16:14,040
Speaker 3: Yeah, and you've got to keep it really really prestige.

289
00:16:14,320 --> 00:16:16,200
Speaker 2: But that when you show me. So in the in

290
00:16:16,280 --> 00:16:20,000
Speaker 2: the lobby of the Watson Research Center in New Yorktown,

291
00:16:20,200 --> 00:16:23,080
Speaker 2: which by the way, is just the coolest building. It's

292
00:16:23,080 --> 00:16:27,880
Speaker 2: like a it's like a modernist it's awesome master piece. Anyway,

293
00:16:28,080 --> 00:16:30,120
Speaker 2: in the lobby there is there are these is it

294
00:16:30,280 --> 00:16:31,480
Speaker 2: two machines.

295
00:16:31,680 --> 00:16:35,120
Speaker 3: It's it's inside a container that has three machines.

296
00:16:35,160 --> 00:16:38,360
Speaker 2: Three machines, So what can you can you tell me

297
00:16:38,400 --> 00:16:40,720
Speaker 2: what would one of those machines cost to build right now?

298
00:16:41,360 --> 00:16:45,520
Speaker 3: So typically we put them together in a way where

299
00:16:45,600 --> 00:16:48,720
Speaker 3: we upgrade them because we want to as I as

300
00:16:48,720 --> 00:16:51,000
Speaker 3: I was talking about before, one of the things we

301
00:16:51,040 --> 00:16:54,640
Speaker 3: want to do is always get algorithms done on our machines.

302
00:16:55,320 --> 00:16:58,320
Speaker 3: And I've got a roadmap of building bigger and bigger machines.

303
00:16:58,840 --> 00:17:02,320
Speaker 3: So usually one of those quantum processors today is out

304
00:17:02,360 --> 00:17:06,439
Speaker 3: of date in six months. So we want to build

305
00:17:06,480 --> 00:17:10,480
Speaker 3: this future of computing that leverages quantum computing where every

306
00:17:10,520 --> 00:17:16,199
Speaker 3: six months we've outdated a quantum processor. Eventually, hopefully we

307
00:17:16,240 --> 00:17:19,280
Speaker 3: get to a point where it's like stable and it

308
00:17:19,280 --> 00:17:22,640
Speaker 3: can be many years operating. But we want to get

309
00:17:22,800 --> 00:17:25,959
Speaker 3: as large a quantum computer in the hands of people

310
00:17:26,000 --> 00:17:28,159
Speaker 3: to explore the math as possible to come up with

311
00:17:28,200 --> 00:17:31,080
Speaker 3: those new algorithms. So we've had a philosophy of having

312
00:17:31,119 --> 00:17:34,919
Speaker 3: them open, working with universities and things like that. So

313
00:17:34,920 --> 00:17:37,119
Speaker 3: to answer a question of costs, yes, there's cost in

314
00:17:37,200 --> 00:17:40,320
Speaker 3: building the system, but we are operating in them much

315
00:17:40,359 --> 00:17:43,320
Speaker 3: more in a service model where people pay to use

316
00:17:43,359 --> 00:17:47,159
Speaker 3: the machine because we have to continuously calibrate it and

317
00:17:47,200 --> 00:17:52,360
Speaker 3: operate it and so depending on various different things. Professors,

318
00:17:52,400 --> 00:17:54,920
Speaker 3: we have a credits program where they get free access.

319
00:17:55,560 --> 00:17:59,080
Speaker 3: Some universities and enterprises they can buy premium access and

320
00:17:59,119 --> 00:18:02,399
Speaker 3: get more access. So think of not like a cost

321
00:18:02,480 --> 00:18:05,480
Speaker 3: of it, because it's almost like a continuum. I want

322
00:18:05,520 --> 00:18:08,480
Speaker 3: to make sure that the best quantum processors that I

323
00:18:08,520 --> 00:18:12,160
Speaker 3: can build get in the hands of students and professors

324
00:18:12,560 --> 00:18:15,440
Speaker 3: and interested enterprises that want to explore these machines as

325
00:18:15,480 --> 00:18:20,200
Speaker 3: fast as possible. And typically every six months we upgrade it. Yeah,

326
00:18:20,400 --> 00:18:24,360
Speaker 3: you don't start over, you upgrade. We upgrade various different pieces,

327
00:18:24,400 --> 00:18:29,240
Speaker 3: the processor, the electronics. Some upgrades are just simply replaced

328
00:18:29,280 --> 00:18:33,359
Speaker 3: the processor. But as an example, I think many people

329
00:18:33,359 --> 00:18:35,920
Speaker 3: have probably seen photos of quantum computers and you see

330
00:18:35,960 --> 00:18:39,440
Speaker 3: this scary thing with all these wires hanging down, as

331
00:18:39,560 --> 00:18:42,040
Speaker 3: I've referred to as the chandelier, and it's got all

332
00:18:42,040 --> 00:18:45,320
Speaker 3: these wires with loops and things like that. They're called

333
00:18:45,520 --> 00:18:48,320
Speaker 3: co x cables. When we first put the quantum computer

334
00:18:48,400 --> 00:18:51,400
Speaker 3: on the cloud in twenty sixteen, you could probably only

335
00:18:51,440 --> 00:18:55,920
Speaker 3: fit about fifty cubits inside one cryostat. We've had to

336
00:18:56,000 --> 00:18:58,920
Speaker 3: upgrade all those cables so that we can fit around

337
00:18:59,000 --> 00:19:03,000
Speaker 3: one thousand to get to three thousand, and that's about minaturizing.

338
00:19:03,480 --> 00:19:06,120
Speaker 3: So to answer your question, an upgrade, it depends. It

339
00:19:06,160 --> 00:19:08,760
Speaker 3: can be just the processor or it can be the

340
00:19:08,840 --> 00:19:12,600
Speaker 3: complete insides. And we're actually in our third generation of

341
00:19:12,640 --> 00:19:16,320
Speaker 3: our electronics to control the systems to make them faster,

342
00:19:16,680 --> 00:19:20,639
Speaker 3: less noise. Internally. We've got exciting results of going to

343
00:19:20,720 --> 00:19:24,479
Speaker 3: something like cold cryocemos. So you can bring down the

344
00:19:24,520 --> 00:19:28,440
Speaker 3: cost in terms of energy of running these quantum computers

345
00:19:28,720 --> 00:19:32,440
Speaker 3: almost to negligible, and you could imagine future quantum computers.

346
00:19:33,080 --> 00:19:35,639
Speaker 3: I'm not going to require much energy to run, so

347
00:19:35,960 --> 00:19:39,800
Speaker 3: unlike classical compute that requires lots of energy. The biggest

348
00:19:39,840 --> 00:19:42,440
Speaker 3: machines that we envision is only in the few megawatts,

349
00:19:42,960 --> 00:19:46,000
Speaker 3: but we have to upgrade to future controls that use

350
00:19:46,080 --> 00:19:50,920
Speaker 3: less energy. So it depends it's my long answer, short

351
00:19:50,960 --> 00:19:53,840
Speaker 3: answer to how it upgrades, and it depends on what

352
00:19:53,920 --> 00:19:54,199
Speaker 3: it is.

353
00:19:54,480 --> 00:19:57,320
Speaker 2: The only observation that I felt I was capable of

354
00:19:57,320 --> 00:20:01,119
Speaker 2: making when you showed me the quantum machine is it's gorgeous.

355
00:20:01,880 --> 00:20:02,560
Speaker 2: I look a art.

356
00:20:02,920 --> 00:20:06,399
Speaker 3: I've always believed that, and I think that there's an

357
00:20:06,400 --> 00:20:09,520
Speaker 3: IBM saying good design is good business. But we've always

358
00:20:09,840 --> 00:20:14,240
Speaker 3: taken pride in making sure what we build. I don't know,

359
00:20:14,359 --> 00:20:17,840
Speaker 3: I feel if you're going to build something that is new,

360
00:20:18,400 --> 00:20:21,560
Speaker 3: that can change, you should take the time to make

361
00:20:21,600 --> 00:20:23,040
Speaker 3: sure it looks and feels good.

362
00:20:23,200 --> 00:20:26,520
Speaker 2: Will you donated to MoMA when you're through with that particular?

363
00:20:27,320 --> 00:20:31,479
Speaker 3: Actually, I think we just put an old version of

364
00:20:31,480 --> 00:20:36,159
Speaker 3: one of our insights with the United Airlines and the AAPS,

365
00:20:36,160 --> 00:20:39,080
Speaker 3: which is the American Physical Society and the University of Chicago.

366
00:20:39,400 --> 00:20:42,240
Speaker 3: There's a replica right now. If you fly into one

367
00:20:42,280 --> 00:20:45,280
Speaker 3: of the terminals in Chicago, you can walk and see one.

368
00:20:45,600 --> 00:20:47,920
Speaker 2: Oh really, yeah, well the most advanced thing at a

369
00:20:48,040 --> 00:20:49,120
Speaker 2: air I'm sure.

370
00:20:49,119 --> 00:20:52,200
Speaker 3: Probably, but yeah, I hopefully. I think, yeah, we're open

371
00:20:52,240 --> 00:20:55,800
Speaker 3: to that. But yeah, I appreciate that you love the design.

372
00:20:55,960 --> 00:20:59,280
Speaker 2: It was beautiful. So I last week I interviewed for

373
00:20:59,320 --> 00:21:05,080
Speaker 2: another episode Smotox, your CEO, Ivin Krishne, and when we

374
00:21:05,160 --> 00:21:08,480
Speaker 2: got to the quantum question. I mean, he's always alliant

375
00:21:08,600 --> 00:21:13,800
Speaker 2: and brilliant, and but quantum, he's like lit up. I

376
00:21:13,800 --> 00:21:17,639
Speaker 2: mean right in thinking that IBM is much more invested

377
00:21:18,280 --> 00:21:20,800
Speaker 2: in quantum than anybody else. Is that a fair statement? Oh?

378
00:21:20,880 --> 00:21:21,840
Speaker 3: Yeah, most definitely.

379
00:21:22,040 --> 00:21:25,320
Speaker 2: Why Why did IBM choose to kind of make this

380
00:21:25,400 --> 00:21:26,200
Speaker 2: such a priority.

381
00:21:26,560 --> 00:21:29,639
Speaker 3: So when I took to the history of the physics side,

382
00:21:30,440 --> 00:21:33,400
Speaker 3: there's this interesting thing in the history of computing. So

383
00:21:33,560 --> 00:21:37,600
Speaker 3: we build computer like classical computers today using bits and

384
00:21:37,680 --> 00:21:40,640
Speaker 3: ce moss, and they consume energy. Do you know that

385
00:21:40,680 --> 00:21:43,720
Speaker 3: there is a way in classical where you can actually

386
00:21:43,920 --> 00:21:48,200
Speaker 3: compute without using energy. It's called reversal computing. Turns out

387
00:21:48,240 --> 00:21:52,040
Speaker 3: to be a terrible idea, it's not practical to build.

388
00:21:52,440 --> 00:21:57,320
Speaker 3: But IBM investigated that with Ralph Laura and Charlie Bennett

389
00:21:57,480 --> 00:22:00,920
Speaker 3: early on, and they proved the concept that we're versable computing.

390
00:22:01,400 --> 00:22:04,879
Speaker 3: The first use of quantum information theory. One of the

391
00:22:04,920 --> 00:22:08,440
Speaker 3: first actually was from IBM. When I did my PhD,

392
00:22:09,119 --> 00:22:12,480
Speaker 3: I remember actually picking up this paper on quantum teleportation

393
00:22:13,080 --> 00:22:15,480
Speaker 3: and seeing IBM written there, and at the time I

394
00:22:15,480 --> 00:22:18,040
Speaker 3: remember thinking that they make PCs. Well, what the hell

395
00:22:18,080 --> 00:22:23,240
Speaker 3: are they doing this foundational paper on quantum teleportation? Why

396
00:22:23,240 --> 00:22:25,800
Speaker 3: are they doing it? So to answer your question, actually,

397
00:22:25,840 --> 00:22:30,360
Speaker 3: IBM was the first in quantum information science because it's

398
00:22:30,400 --> 00:22:34,280
Speaker 3: the fundamental of computation. Can we actually come up with

399
00:22:34,359 --> 00:22:38,639
Speaker 3: compute that we can go beyond the classical So way

400
00:22:38,720 --> 00:22:42,679
Speaker 3: before anyone was talking about it, they were doing fundamental theory.

401
00:22:43,359 --> 00:22:46,119
Speaker 3: And then as we've built it, we've always When I

402
00:22:46,160 --> 00:22:49,479
Speaker 3: first came there, the experimental team was small. In twenty eleven,

403
00:22:50,000 --> 00:22:53,919
Speaker 3: we've had a small team that we're focusing on single

404
00:22:54,000 --> 00:22:58,320
Speaker 3: cubits coupling in them. I think in twenty twelve was

405
00:22:58,320 --> 00:23:01,879
Speaker 3: the first time we showed really good two Cuba gates

406
00:23:02,680 --> 00:23:05,919
Speaker 3: and no one was talking about quantum computing that And

407
00:23:05,960 --> 00:23:09,720
Speaker 3: then I remember in about twenty sixteen I said to

408
00:23:10,040 --> 00:23:13,520
Speaker 3: actually Arvin was the director of research, then can we

409
00:23:13,680 --> 00:23:17,000
Speaker 3: actually put our quantum computer on the cloud? Well that's

410
00:23:17,000 --> 00:23:21,719
Speaker 3: probably twenty fifteen, and it was always supporting that. So

411
00:23:21,800 --> 00:23:24,240
Speaker 3: as we've done more and more we've been able to

412
00:23:24,280 --> 00:23:28,160
Speaker 3: do it. It's had this program going now, I agree

413
00:23:28,560 --> 00:23:31,800
Speaker 3: is very visible, like because we're in this scaling phase

414
00:23:32,200 --> 00:23:35,800
Speaker 3: and so we're invested to keep scaling it and to

415
00:23:35,840 --> 00:23:40,280
Speaker 3: get why is At IBM research, what we always do

416
00:23:40,520 --> 00:23:43,520
Speaker 3: is answer what is the future of computing? Whether it's

417
00:23:43,520 --> 00:23:47,480
Speaker 3: coming up with new algorithms, coming up with better AI,

418
00:23:47,800 --> 00:23:50,840
Speaker 3: coming up with quantum, or coming up with just how

419
00:23:50,880 --> 00:23:54,240
Speaker 3: do different accelerators go together. It's our DNA to answer

420
00:23:54,240 --> 00:23:55,919
Speaker 3: the question of what is the future.

421
00:23:56,080 --> 00:23:58,399
Speaker 2: Need a perfect problem for IBM because you kind of

422
00:23:58,440 --> 00:24:04,280
Speaker 2: need to have a legacy of buildings, building actual physical machines.

423
00:24:04,880 --> 00:24:08,800
Speaker 3: Yeah, it's why I came to IBM. I wanted the experience,

424
00:24:09,520 --> 00:24:14,040
Speaker 3: the culture of building hard things that others have not

425
00:24:14,160 --> 00:24:14,760
Speaker 3: done before.

426
00:24:16,119 --> 00:24:18,480
Speaker 2: Where do you imagine we are in the timeline of

427
00:24:18,480 --> 00:24:22,680
Speaker 2: this technology? It will come a point when it will mature.

428
00:24:23,760 --> 00:24:26,479
Speaker 2: My cell phone is a mature technology at this point.

429
00:24:26,680 --> 00:24:29,080
Speaker 2: How far are we from that point with condom?

430
00:24:29,520 --> 00:24:32,120
Speaker 3: So I think there's various aspects of it. So we

431
00:24:32,160 --> 00:24:35,600
Speaker 3: set in twenty and seventy we set our goal that

432
00:24:35,680 --> 00:24:38,400
Speaker 3: in twenty twenty three we would be able to build

433
00:24:38,440 --> 00:24:42,920
Speaker 3: a machine that was beyond classical computers to simulate it,

434
00:24:43,560 --> 00:24:46,479
Speaker 3: and we achieved that in twenty twenty three. So to

435
00:24:47,520 --> 00:24:49,439
Speaker 3: run a bigger we call it a quantum circuit. The

436
00:24:49,480 --> 00:24:51,560
Speaker 3: details of it din't matter, but to run a quantum

437
00:24:51,600 --> 00:24:55,760
Speaker 3: workload that if you were to simulate that workload how

438
00:24:55,800 --> 00:24:58,840
Speaker 3: a quantum computer operates on a classical computer, you couldn't

439
00:24:58,840 --> 00:25:01,560
Speaker 3: do it. So we set that does our first and

440
00:25:01,600 --> 00:25:04,720
Speaker 3: now I've made it publicly that by twenty twenty nine

441
00:25:04,960 --> 00:25:08,240
Speaker 3: we'll build the first fault tolerant corantum computer. That is,

442
00:25:08,280 --> 00:25:12,920
Speaker 3: one that can completely handle the noise to the level

443
00:25:13,000 --> 00:25:16,680
Speaker 3: to allow you to run a very very large, large problem.

444
00:25:16,800 --> 00:25:19,160
Speaker 2: So an example of a large problem.

445
00:25:18,880 --> 00:25:22,480
Speaker 3: Yeah, a large quantum problem. So for around a couple

446
00:25:22,480 --> 00:25:26,040
Speaker 3: of one hundred cubits and one hundred million operations, you're

447
00:25:26,080 --> 00:25:31,639
Speaker 3: talking still interesting science problems like simulating a molecule, or

448
00:25:32,080 --> 00:25:39,480
Speaker 3: calculating a small optimization problem, or calculating say some part

449
00:25:39,760 --> 00:25:42,560
Speaker 3: of a matrix update in some type of differential. So

450
00:25:42,560 --> 00:25:45,560
Speaker 3: it'll still be scientific, but it'll be at the point

451
00:25:45,640 --> 00:25:51,439
Speaker 3: where it's beyond, well beyond any classical approximate method. And

452
00:25:51,480 --> 00:25:54,440
Speaker 3: then I think that's twenty twenty nine that's twenty twenty nine,

453
00:25:54,640 --> 00:25:55,920
Speaker 3: so we're four.

454
00:25:55,800 --> 00:25:58,560
Speaker 2: Years away from something that can start to handle.

455
00:25:58,880 --> 00:26:02,800
Speaker 3: Interesting problem, serious problems. I do believe the scientists will

456
00:26:02,840 --> 00:26:06,320
Speaker 3: find interesting heuristic problems before that, and so over the

457
00:26:06,320 --> 00:26:08,800
Speaker 3: next four years, you're going to continue to see more

458
00:26:08,840 --> 00:26:13,400
Speaker 3: and more let's call them heuristic not provable quantum problems

459
00:26:13,400 --> 00:26:16,440
Speaker 3: that run on quantum computers that come out. We're seeing

460
00:26:16,560 --> 00:26:19,320
Speaker 3: more and more come from many of our partners and ourselves.

461
00:26:19,600 --> 00:26:22,400
Speaker 3: Heuristic problems have value, but they have to be tested,

462
00:26:22,480 --> 00:26:24,200
Speaker 3: they have to stand up over time. You have to

463
00:26:24,280 --> 00:26:26,679
Speaker 3: run them many, many times, you have to try different ones,

464
00:26:27,160 --> 00:26:30,200
Speaker 3: and many times heuristic can lead to formal problems. So

465
00:26:30,280 --> 00:26:32,359
Speaker 3: you're going to see because we're beyond now the point

466
00:26:32,400 --> 00:26:36,760
Speaker 3: that you can simulate these quantum computers with any classical computer.

467
00:26:36,840 --> 00:26:40,320
Speaker 3: They're kind of like a scientific tool. So they're exploring

468
00:26:40,400 --> 00:26:40,960
Speaker 3: the heuristic.

469
00:26:41,200 --> 00:26:42,800
Speaker 2: What do you have to get done between now and

470
00:26:42,840 --> 00:26:44,160
Speaker 2: twenty twenty nine to get there?

471
00:26:44,600 --> 00:26:46,760
Speaker 3: So we had to reinvent how we wanted to do

472
00:26:46,880 --> 00:26:50,840
Speaker 3: error correction. So we have to demonstrate modules and if

473
00:26:50,880 --> 00:26:54,560
Speaker 3: we can demonstrate these error corrected module and our goal

474
00:26:54,680 --> 00:26:57,320
Speaker 3: is actually it's called Crookobarro I name all our chips

475
00:26:57,359 --> 00:26:59,920
Speaker 3: after birds, so it's called Kookobara. It is named after

476
00:27:00,040 --> 00:27:02,720
Speaker 3: an Australian verte. I think I still say kooka burrow

477
00:27:02,840 --> 00:27:07,080
Speaker 3: the way Australians do. We need to then show that

478
00:27:07,119 --> 00:27:09,320
Speaker 3: we can make a single module and then we want

479
00:27:09,359 --> 00:27:11,760
Speaker 3: to connect two of those modules together, and I call

480
00:27:11,840 --> 00:27:15,600
Speaker 3: that one cockatoo, which is another Australian vert. And then

481
00:27:15,640 --> 00:27:18,720
Speaker 3: if we can do that, so that's twenty six and

482
00:27:18,880 --> 00:27:21,760
Speaker 3: twenty seven, and then we want to scale the scale

483
00:27:21,800 --> 00:27:24,640
Speaker 3: those modules, and that we call starling and we want

484
00:27:24,640 --> 00:27:27,760
Speaker 3: to scale that in twenty twenty nine. So get a module,

485
00:27:28,119 --> 00:27:31,679
Speaker 3: join two modules together and scale and so each module

486
00:27:31,720 --> 00:27:33,400
Speaker 3: is going to be around one thousand cubits.

487
00:27:34,440 --> 00:27:36,960
Speaker 2: The challenge to getting there is it finding the right

488
00:27:37,200 --> 00:27:41,000
Speaker 2: material or how would you describe what? That's the beauty

489
00:27:41,119 --> 00:27:41,440
Speaker 2: to be done.

490
00:27:41,440 --> 00:27:45,399
Speaker 3: That's the beauty of it is if we would have

491
00:27:45,400 --> 00:27:48,480
Speaker 3: been here two years ago, I couldn't tell you how

492
00:27:48,520 --> 00:27:51,439
Speaker 3: it would be done. So we had a huge breakthrough.

493
00:27:51,680 --> 00:27:54,360
Speaker 3: We came up with a new code, a new quantumeric

494
00:27:54,440 --> 00:27:58,639
Speaker 3: Russian code, and that code. The biggest im part of

495
00:27:58,680 --> 00:28:01,600
Speaker 3: that code that is the most important is it's modular

496
00:28:01,640 --> 00:28:06,280
Speaker 3: in nature. So previous codes without getting too technical, they

497
00:28:06,280 --> 00:28:09,120
Speaker 3: were very monolithic and you had to build a very

498
00:28:09,119 --> 00:28:11,399
Speaker 3: big device, and I wouldn't have known we would have

499
00:28:11,520 --> 00:28:15,400
Speaker 3: to invent tools like new Simos tools to do that.

500
00:28:16,160 --> 00:28:19,080
Speaker 3: So we came up with this new code. We started

501
00:28:19,160 --> 00:28:21,879
Speaker 3: on twenty nineteen, we published in twenty twenty four. We

502
00:28:21,960 --> 00:28:23,879
Speaker 3: kind of had most of things worked out in twenty

503
00:28:23,920 --> 00:28:26,960
Speaker 3: twenty three. That's why we got confident to release the thing.

504
00:28:27,359 --> 00:28:29,520
Speaker 3: So the biggest breakthrough we had is coming up with

505
00:28:29,560 --> 00:28:32,320
Speaker 3: a code that's modular in nature, and think of that

506
00:28:32,359 --> 00:28:35,159
Speaker 3: as a like a blueprint. And so now we have

507
00:28:35,320 --> 00:28:40,320
Speaker 3: the blueprint, and now we're doing engineering tasks to implement

508
00:28:40,400 --> 00:28:41,960
Speaker 3: every part of that blueprint.

509
00:28:42,200 --> 00:28:45,280
Speaker 2: And so the minute you had that breakthrough, then you

510
00:28:45,400 --> 00:28:48,200
Speaker 2: began to have confidence at something exactly these goals could

511
00:28:48,200 --> 00:28:49,000
Speaker 2: be met.

512
00:28:48,800 --> 00:28:52,520
Speaker 3: And then you can't. And then anyone that's done engineering

513
00:28:52,760 --> 00:28:54,560
Speaker 3: will know what I'm talking about when I say this

514
00:28:54,680 --> 00:28:58,880
Speaker 3: is cycles are learning. It takes so long from test

515
00:28:58,920 --> 00:29:03,000
Speaker 3: idea to build two tests. In hardware, the cycles of

516
00:29:03,120 --> 00:29:05,440
Speaker 3: learning are much much lower than software, Like you can

517
00:29:05,480 --> 00:29:08,720
Speaker 3: be really really faster in the software. So then we've

518
00:29:08,800 --> 00:29:12,479
Speaker 3: planned out our iterations over the next few years, and

519
00:29:12,560 --> 00:29:16,760
Speaker 3: so we have to successfully demonstrate them. I may slip,

520
00:29:16,840 --> 00:29:21,600
Speaker 3: because sometimes you may estimate your time wrong, but we

521
00:29:21,720 --> 00:29:24,160
Speaker 3: now have exactly what we want to do for the

522
00:29:24,200 --> 00:29:24,920
Speaker 3: next four years.

523
00:29:25,040 --> 00:29:26,600
Speaker 2: I want to go back to that breakthrough for a moment.

524
00:29:26,760 --> 00:29:29,440
Speaker 2: What does the word breaks we mean in that context, Like,

525
00:29:29,720 --> 00:29:31,800
Speaker 2: it's not that you get a call in the morning

526
00:29:32,240 --> 00:29:34,720
Speaker 2: from somebody who says, I did it. Do you see

527
00:29:34,760 --> 00:29:36,920
Speaker 2: it coming? Or is it a surprise when they get there.

528
00:29:37,120 --> 00:29:40,760
Speaker 3: So the way this one worked is Sogo Brave, who's

529
00:29:41,280 --> 00:29:44,240
Speaker 3: an algorithm person at IBM, one of the smartest and

530
00:29:44,360 --> 00:29:45,160
Speaker 3: quantum information.

531
00:29:45,520 --> 00:29:49,440
Speaker 2: Don't mention his name to everyone value you'll come for him.

532
00:29:49,520 --> 00:29:52,320
Speaker 3: Everyone in quantum already knows his name. I don't think

533
00:29:52,360 --> 00:29:56,280
Speaker 3: there's an idea that has not originated from him in quantum.

534
00:29:58,360 --> 00:30:01,040
Speaker 3: So we're looking at other codes and we'll go all right,

535
00:30:02,080 --> 00:30:05,840
Speaker 3: we've got to get serious about these codes. And others

536
00:30:05,960 --> 00:30:08,560
Speaker 3: were starting to propose to bring these and then we

537
00:30:08,600 --> 00:30:13,320
Speaker 3: call them LDBC codes from the classical space into the quantum.

538
00:30:13,840 --> 00:30:16,560
Speaker 3: And I asked him, we need to get ahead of

539
00:30:16,600 --> 00:30:19,280
Speaker 3: this and understand what they're doing it. He's like the

540
00:30:19,320 --> 00:30:23,080
Speaker 3: most modest perfuse late, Jay, let me learn about them

541
00:30:23,160 --> 00:30:26,280
Speaker 3: and I'll generate a report for us and we'll read

542
00:30:26,280 --> 00:30:29,320
Speaker 3: through it. And then I said, great, Then I don't know.

543
00:30:29,400 --> 00:30:31,760
Speaker 3: Six months later, he comes back with one hundred page

544
00:30:31,800 --> 00:30:35,760
Speaker 3: report on everyone. Everyone had done an LTPC codes. I'm like, awesome.

545
00:30:35,800 --> 00:30:39,280
Speaker 3: So I started then to read from them. And then

546
00:30:39,400 --> 00:30:42,440
Speaker 3: we said, all right, how do we under the assumptions

547
00:30:42,440 --> 00:30:45,360
Speaker 3: of the hardware we can build? Can we get an

548
00:30:45,520 --> 00:30:51,280
Speaker 3: LTPC code knowing what we can build? And that's a

549
00:30:51,280 --> 00:30:54,160
Speaker 3: great question, and so we put a small team together

550
00:30:54,560 --> 00:30:58,280
Speaker 3: to investigate and honestly took two to three years, and

551
00:30:58,680 --> 00:31:02,840
Speaker 3: we iterated, and we used the constraints, so we had

552
00:31:02,880 --> 00:31:05,720
Speaker 3: the sort of theory, and then we had the constraints

553
00:31:05,720 --> 00:31:08,440
Speaker 3: of what we could build. And we iterated for a

554
00:31:08,480 --> 00:31:10,920
Speaker 3: few years, and then at the end of that we

555
00:31:11,000 --> 00:31:13,880
Speaker 3: came out with a solution that, yes, it is possible

556
00:31:13,960 --> 00:31:17,080
Speaker 3: to meet all the constraints of the hardware and build

557
00:31:17,120 --> 00:31:18,360
Speaker 3: a code that will work.

558
00:31:18,960 --> 00:31:22,720
Speaker 2: I'm just curious about So you had this task, this

559
00:31:22,880 --> 00:31:26,320
Speaker 2: problem you want to solve, and when you set out

560
00:31:26,320 --> 00:31:28,520
Speaker 2: on the task of trying to solve the problem, what's

561
00:31:28,560 --> 00:31:31,400
Speaker 2: your certainty level that you'll get a solution.

562
00:31:31,840 --> 00:31:35,160
Speaker 3: Well, that's the beauty of science. For things where you

563
00:31:35,320 --> 00:31:39,160
Speaker 3: kind of have a few ideas. My philosophy is try

564
00:31:39,200 --> 00:31:41,760
Speaker 3: a few for the ones that need to be in

565
00:31:41,840 --> 00:31:46,760
Speaker 3: that like wow moment. It's honestly, you've got to set

566
00:31:46,760 --> 00:31:50,120
Speaker 3: the ambition really, really high, but know when to stop.

567
00:31:51,040 --> 00:31:53,080
Speaker 3: It was a great team that went together to get

568
00:31:53,120 --> 00:31:56,080
Speaker 3: that breakthrough, and we knew that we needed to come

569
00:31:56,160 --> 00:31:59,959
Speaker 3: up with a code that met the requirements of the experiment.

570
00:32:00,760 --> 00:32:04,480
Speaker 3: And I think what was different before then is the

571
00:32:04,480 --> 00:32:09,200
Speaker 3: theorists that were doing error correction codes didn't necessarily know

572
00:32:09,440 --> 00:32:12,960
Speaker 3: the constraints of experiments, so it was like really more

573
00:32:13,000 --> 00:32:15,480
Speaker 3: pen and paper. So this became one all right, given

574
00:32:15,520 --> 00:32:18,240
Speaker 3: these sets of constraints, is it possible?

575
00:32:18,720 --> 00:32:21,600
Speaker 2: When that's questions about this? Sorry, And I love these

576
00:32:21,680 --> 00:32:24,760
Speaker 2: kind of moments when things become clear. At the time

577
00:32:24,840 --> 00:32:28,240
Speaker 2: the problem was solved, were you aware of the implications

578
00:32:28,240 --> 00:32:31,960
Speaker 2: of the solution or did that takes you knew exactly.

579
00:32:31,600 --> 00:32:35,959
Speaker 3: What we set out exactly like either we were going

580
00:32:36,040 --> 00:32:38,080
Speaker 3: to have to work out how to cool down a

581
00:32:38,200 --> 00:32:41,720
Speaker 3: very large piece of silicon, which would require a lot

582
00:32:41,760 --> 00:32:44,880
Speaker 3: of engineering and building tools beyond what anyone has ever

583
00:32:44,920 --> 00:32:49,600
Speaker 3: built in the silicon semoss industry. To implement the known

584
00:32:49,680 --> 00:32:52,959
Speaker 3: codes or we had to come up with a different one,

585
00:32:53,200 --> 00:32:55,760
Speaker 3: and once I knew that we had one that I

586
00:32:55,960 --> 00:33:00,800
Speaker 3: didn't need to reinvent any tools to build. The implications

587
00:33:00,800 --> 00:33:01,600
Speaker 3: are clear how.

588
00:33:01,600 --> 00:33:04,680
Speaker 2: Much time elapsed between the time you heard the problem

589
00:33:04,760 --> 00:33:07,960
Speaker 2: was solved and the time you told Arvin Krishna, the CEO,

590
00:33:08,080 --> 00:33:09,320
Speaker 2: the problem was solved.

591
00:33:10,320 --> 00:33:12,840
Speaker 3: I'm sure the next time I spoke to him, I update,

592
00:33:12,840 --> 00:33:15,760
Speaker 3: but I don't remember. The beauty of Avin is he

593
00:33:15,840 --> 00:33:18,400
Speaker 3: trusts the scientists will do it, and so he doesn't

594
00:33:18,400 --> 00:33:21,040
Speaker 3: really check on us. We update him when when it

595
00:33:21,160 --> 00:33:24,400
Speaker 3: is and he he empowers us to do really hard problems.

596
00:33:24,600 --> 00:33:28,320
Speaker 2: Yeah, so let's talk about uses. I mean they're really

597
00:33:28,440 --> 00:33:32,280
Speaker 2: like cool, big shiny machine. I think you'll get paid

598
00:33:32,320 --> 00:33:35,560
Speaker 2: twenty twenty nine. But there's all kinds of really interesting

599
00:33:35,560 --> 00:33:37,000
Speaker 2: problems you're already working on.

600
00:33:37,400 --> 00:33:42,640
Speaker 3: Yes, this is like another interesting area is I can

601
00:33:42,800 --> 00:33:46,360
Speaker 3: prove in pen and paper algorithms that we want to

602
00:33:46,440 --> 00:33:48,680
Speaker 3: run that. Like, it's not that we don't know what

603
00:33:48,720 --> 00:33:51,680
Speaker 3: to do with a quantum computer. There are hundreds of algorithms.

604
00:33:51,680 --> 00:33:54,240
Speaker 3: So you can go to I think it's called quantumzoo

605
00:33:54,280 --> 00:33:57,280
Speaker 3: dot com and you can see many many algorithms people

606
00:33:57,320 --> 00:33:58,880
Speaker 3: are coming up with more of more of them that

607
00:33:59,000 --> 00:34:02,440
Speaker 3: they prove by pen and paper. But imagine now we

608
00:34:02,560 --> 00:34:07,160
Speaker 3: have a machine that you can't simulate. How do you

609
00:34:07,280 --> 00:34:12,799
Speaker 3: actually discover algorithms in a scientific way? How do you

610
00:34:12,920 --> 00:34:16,680
Speaker 3: look and discover algorithms using a quantum computer. We're in

611
00:34:16,719 --> 00:34:21,120
Speaker 3: this exciting period right now, and so even though I

612
00:34:21,120 --> 00:34:23,600
Speaker 3: can prove these ones that we can run in the future,

613
00:34:24,120 --> 00:34:27,960
Speaker 3: there's a big white space between what the machines we

614
00:34:28,040 --> 00:34:30,040
Speaker 3: have and we're going to build and continue to do

615
00:34:30,480 --> 00:34:35,080
Speaker 3: and those ones that want the provable ones. And I'm

616
00:34:35,120 --> 00:34:40,440
Speaker 3: an optimistic person by nature. I think getting those machines

617
00:34:40,480 --> 00:34:42,879
Speaker 3: in the hands of students to explore and look at

618
00:34:42,880 --> 00:34:47,080
Speaker 3: heuristic algorithms. So looking at the equivalent of doing numerical

619
00:34:47,120 --> 00:34:52,919
Speaker 3: algorithms on computers, which there's many histories of numerical algorithms

620
00:34:53,000 --> 00:34:57,240
Speaker 3: being discovered on classical computers before we had formal proofs

621
00:34:57,560 --> 00:35:00,239
Speaker 3: that we rely on today people would even are you

622
00:35:00,400 --> 00:35:04,040
Speaker 3: the way AI works was driven numerically, even though we

623
00:35:04,080 --> 00:35:08,480
Speaker 3: have input into it. There are ones in optimization driven numerically.

624
00:35:09,000 --> 00:35:12,839
Speaker 3: We are entering that phase. So the computer scientists now

625
00:35:13,440 --> 00:35:17,120
Speaker 3: need to go play with these primitives. Our prediction is

626
00:35:17,840 --> 00:35:20,560
Speaker 3: over the next couple of years, we're going to see

627
00:35:20,960 --> 00:35:25,759
Speaker 3: valuable numerical equivalent algorithms emerge. And where the scientists are

628
00:35:25,800 --> 00:35:29,680
Speaker 3: going is in four categories. One is simulating nature, so

629
00:35:29,800 --> 00:35:34,279
Speaker 3: looking at either hay Enerji physics, chemistry, light problems. As

630
00:35:34,280 --> 00:35:38,000
Speaker 3: an example, with our partners in Japan, they took one

631
00:35:38,040 --> 00:35:42,040
Speaker 3: of our quantum computers and for Gackle, a very large

632
00:35:42,760 --> 00:35:46,960
Speaker 3: classical supercomputer, and they ran a problem where quantum was

633
00:35:47,080 --> 00:35:50,280
Speaker 3: just a sub routine of the problem that was running

634
00:35:50,320 --> 00:35:52,080
Speaker 3: on all of for Gackle, and they were able to

635
00:35:52,120 --> 00:35:55,000
Speaker 3: look at an interesting molecule, a molecule that if you

636
00:35:55,000 --> 00:35:56,759
Speaker 3: would go by pan and paper you would have said,

637
00:35:56,800 --> 00:35:58,440
Speaker 3: it's going to take me a very long time to

638
00:35:58,520 --> 00:36:01,160
Speaker 3: run that. They were able to on that quite accurately,

639
00:36:01,239 --> 00:36:04,799
Speaker 3: heuristically and already get results that are comparable with the

640
00:36:04,800 --> 00:36:08,040
Speaker 3: best classical methods. So they are extremely excited because they

641
00:36:08,040 --> 00:36:10,320
Speaker 3: want to push that further, and they're sort of showing

642
00:36:10,360 --> 00:36:13,480
Speaker 3: that you can take a classical supercomputer with quantum as

643
00:36:13,480 --> 00:36:16,200
Speaker 3: a subroutine and start to push the level.

644
00:36:16,160 --> 00:36:19,200
Speaker 2: They were This was trying to solve a medical problem.

645
00:36:19,239 --> 00:36:23,560
Speaker 3: Is this one is a like most people don't realize,

646
00:36:23,640 --> 00:36:26,520
Speaker 3: like iron sulfur, just something as simple as iron and sulfur,

647
00:36:26,880 --> 00:36:31,480
Speaker 3: we can't solve that exactly, Like iron sulfur, molecules are

648
00:36:31,520 --> 00:36:35,839
Speaker 3: too hard. So really small small molecules are really really hard,

649
00:36:35,920 --> 00:36:38,759
Speaker 3: too hard for classical computers to solve. People think we

650
00:36:38,800 --> 00:36:41,000
Speaker 3: can solve a lot of things. It actually turns out

651
00:36:41,040 --> 00:36:42,240
Speaker 3: we can't solve very much.

652
00:36:42,440 --> 00:36:45,319
Speaker 2: You say solve ins instance, you know precisely how that

653
00:36:45,360 --> 00:36:47,480
Speaker 2: molecule works and is constructed.

654
00:36:47,480 --> 00:36:51,560
Speaker 3: No, precisely what the energy levels of that molecule is

655
00:36:51,640 --> 00:36:54,400
Speaker 3: and how they come together, and then be able to

656
00:36:54,400 --> 00:36:56,719
Speaker 3: do that on a classical computer and compare it to

657
00:36:56,719 --> 00:36:57,640
Speaker 3: a quantum.

658
00:36:57,320 --> 00:37:00,520
Speaker 2: It would be really really useful to know that specific.

659
00:37:00,360 --> 00:37:03,759
Speaker 3: Because if you can have energy levels, then you can

660
00:37:03,880 --> 00:37:07,640
Speaker 3: estimate reaction rates. If you can estimate reaction rates, you

661
00:37:07,640 --> 00:37:11,440
Speaker 3: can see how different types of chemicals will react. That

662
00:37:11,480 --> 00:37:14,680
Speaker 3: can then lead to better informing eventually how to build

663
00:37:14,719 --> 00:37:17,719
Speaker 3: materials or even drug design. I just want to be

664
00:37:17,719 --> 00:37:19,759
Speaker 3: careful and not say, oh, we're going to solve drug

665
00:37:19,800 --> 00:37:24,520
Speaker 3: design or that, because there's many scientific steps to make

666
00:37:24,560 --> 00:37:27,600
Speaker 3: that so and so what quantum gives you as a

667
00:37:27,600 --> 00:37:30,879
Speaker 3: different tool to give you more accuracy and then lead

668
00:37:30,960 --> 00:37:32,719
Speaker 3: to making the different methods work.

669
00:37:33,320 --> 00:37:38,160
Speaker 2: You can subcontract out aspects of a problem quantum right now,

670
00:37:38,200 --> 00:37:41,440
Speaker 2: and that just gets you further along and you would.

671
00:37:41,200 --> 00:37:44,840
Speaker 3: Have been so at the moment. Even this result still

672
00:37:44,960 --> 00:37:49,120
Speaker 3: does not beat the best approximate classical method. It's comparable.

673
00:37:49,520 --> 00:37:53,680
Speaker 3: So the art of chemistry for the last hundred years

674
00:37:53,960 --> 00:37:57,680
Speaker 3: has been about approximating. So what we've done is we

675
00:37:57,800 --> 00:38:01,320
Speaker 3: have got very very good are coming up with ways

676
00:38:01,360 --> 00:38:05,279
Speaker 3: of approximating nature. And a lot of the things that

677
00:38:05,320 --> 00:38:09,080
Speaker 3: we do and we exploit and we use to estimate approximations.

678
00:38:09,120 --> 00:38:11,360
Speaker 3: They don't a stimulate nature of the way nature is.

679
00:38:11,400 --> 00:38:15,680
Speaker 3: They approximate it. And there's I could list many different

680
00:38:15,760 --> 00:38:20,560
Speaker 3: acronyms of different methods that go into approximating nature. What

681
00:38:20,680 --> 00:38:24,560
Speaker 3: quantum gives us is to eventually get beyond that approximation

682
00:38:25,160 --> 00:38:27,800
Speaker 3: and do it the way nature works. And so we

683
00:38:28,280 --> 00:38:31,680
Speaker 3: aren't beating those approximation methods. And this is why I think,

684
00:38:31,719 --> 00:38:33,239
Speaker 3: this is why it's still in the science. But they're

685
00:38:33,239 --> 00:38:36,960
Speaker 3: getting comparable. Getting comparable with a new tool where the

686
00:38:37,000 --> 00:38:41,640
Speaker 3: previous tool is a dead end makes scientists very excited. Yeah,

687
00:38:41,800 --> 00:38:44,080
Speaker 3: that nuance is where it is, and so that's in

688
00:38:44,160 --> 00:38:48,600
Speaker 3: machine learning, sorry Hamiltonian. Then there's examples in differential equations,

689
00:38:49,080 --> 00:38:51,960
Speaker 3: So can I actually come up with differential equations and

690
00:38:52,000 --> 00:38:54,520
Speaker 3: solve them? And if I can solve them, you could

691
00:38:54,520 --> 00:38:57,920
Speaker 3: look at things like an avious Stokes equation goes into weather.

692
00:38:58,440 --> 00:39:01,960
Speaker 3: There's financial differential equations that you can better predict. So

693
00:39:02,040 --> 00:39:05,640
Speaker 3: differential equations. There's many different examples there. And then I

694
00:39:05,640 --> 00:39:08,759
Speaker 3: would say that two others are optimization, and then there's

695
00:39:08,840 --> 00:39:12,359
Speaker 3: quantum versions of machine learning that are very exciting as well.

696
00:39:13,040 --> 00:39:16,120
Speaker 2: Cleveland Clinic one of the organizations that you guys have

697
00:39:16,200 --> 00:39:18,960
Speaker 2: worked with. Why would the Cleveland Clinic be calling you up?

698
00:39:19,320 --> 00:39:22,200
Speaker 3: Because that problem that they want to look at. So

699
00:39:23,080 --> 00:39:26,360
Speaker 3: they've also done similar problem to the recent lab. So

700
00:39:26,400 --> 00:39:29,640
Speaker 3: they've taken that method now and they've looked at molecules

701
00:39:29,640 --> 00:39:34,160
Speaker 3: that matter for drug design. So they're fundamentally looking at

702
00:39:34,200 --> 00:39:39,040
Speaker 3: those molecules that matter for eventually replacing some of the steps.

703
00:39:39,360 --> 00:39:43,080
Speaker 3: So they're investing to see how reliable it can be done.

704
00:39:43,120 --> 00:39:45,840
Speaker 3: And so there's a scientist there that's done many iterations

705
00:39:45,880 --> 00:39:49,000
Speaker 3: now using the techniques that were done first with the

706
00:39:49,040 --> 00:39:53,000
Speaker 3: team in Japan. They've now replicated that for new molecules

707
00:39:53,480 --> 00:39:58,040
Speaker 3: that are essential primitives for eventually designing drugs and things

708
00:39:58,080 --> 00:39:59,720
Speaker 3: that may matter for medical.

709
00:40:00,239 --> 00:40:05,319
Speaker 2: And also there's some finance firms yep, HBC, Vango, yep,

710
00:40:05,800 --> 00:40:06,920
Speaker 2: and their interest is.

711
00:40:06,880 --> 00:40:10,640
Speaker 3: What so that was the differential equation and optimization. So

712
00:40:11,160 --> 00:40:15,360
Speaker 3: if you are doing very large calculations like risk portfolio,

713
00:40:15,960 --> 00:40:18,400
Speaker 3: or if you want to model the Black Shaws equation

714
00:40:18,560 --> 00:40:20,719
Speaker 3: or things like this that are fundamental for them to

715
00:40:20,960 --> 00:40:23,960
Speaker 3: make better predictions, come up with better trades and things

716
00:40:24,040 --> 00:40:28,359
Speaker 3: like this. That is a very hard computational task. And

717
00:40:28,440 --> 00:40:32,160
Speaker 3: so rather than quantum replacing that whole problem, can quantum

718
00:40:32,239 --> 00:40:36,160
Speaker 3: be a subroutine in there? And what HSBC showed is

719
00:40:36,200 --> 00:40:38,680
Speaker 3: they showed they could take their real data, they could

720
00:40:38,680 --> 00:40:42,200
Speaker 3: take their real classical method and they just replaced a

721
00:40:42,280 --> 00:40:45,120
Speaker 3: tiny part of it. They replaced a tiny part of

722
00:40:45,160 --> 00:40:48,319
Speaker 3: it with a quantum subroutine that allowed them to come

723
00:40:48,360 --> 00:40:51,719
Speaker 3: up with better predictions of the weights that then when

724
00:40:51,760 --> 00:40:54,800
Speaker 3: they were to compare TRIALA versus Trial B, it was

725
00:40:54,880 --> 00:40:59,160
Speaker 3: thirty four percent better at predicting algorithmic tron And that's

726
00:40:59,160 --> 00:41:00,000
Speaker 3: a big deal for them.

727
00:41:00,520 --> 00:41:01,120
Speaker 2: It's huge.

728
00:41:01,239 --> 00:41:04,480
Speaker 3: Yes, Now, do they need to do more trials? Do

729
00:41:04,560 --> 00:41:07,120
Speaker 3: they need to see is this a heuristic algorithm? Do

730
00:41:07,200 --> 00:41:10,200
Speaker 3: we need to be careful? Is there other classical algorithms

731
00:41:10,239 --> 00:41:12,480
Speaker 3: that go into these are great questions that are now

732
00:41:13,000 --> 00:41:17,319
Speaker 3: being investigated. So think of this period of heuristic algorithms

733
00:41:17,880 --> 00:41:21,600
Speaker 3: is really a period of scientific discovery using these machines,

734
00:41:22,600 --> 00:41:25,520
Speaker 3: knowing that we want to continue and build the ones

735
00:41:25,600 --> 00:41:29,120
Speaker 3: which have determinist their algorithms that can run.

736
00:41:29,960 --> 00:41:33,239
Speaker 2: Do the people who would profit the most by starting

737
00:41:33,280 --> 00:41:38,600
Speaker 2: to run quantum experiments realize that they would profit so

738
00:41:38,680 --> 00:41:42,040
Speaker 2: much from running quantum experience And does the world know this.

739
00:41:42,680 --> 00:41:46,480
Speaker 2: You've given us a couple of specific examples, but generally speaking,

740
00:41:46,520 --> 00:41:48,640
Speaker 2: there must be a very large universe of people who

741
00:41:48,680 --> 00:41:51,720
Speaker 2: could gain from at least starting to play in the space.

742
00:41:52,320 --> 00:41:56,759
Speaker 3: So the enterprises that use computation as key for their

743
00:41:56,800 --> 00:42:01,279
Speaker 3: survival understand the limits of classical computation and they're very

744
00:42:01,320 --> 00:42:06,160
Speaker 3: interested to get started. The universities are very interested. Could

745
00:42:06,239 --> 00:42:09,840
Speaker 3: we get more students doing more algorithms? One hundred percent?

746
00:42:10,800 --> 00:42:14,000
Speaker 3: Some of the limitations on the rate of algorithm discovery

747
00:42:14,040 --> 00:42:17,080
Speaker 3: is because people are thinking through the classical way of

748
00:42:17,120 --> 00:42:20,200
Speaker 3: writing algorithms. My belief is yes, So this is why

749
00:42:20,280 --> 00:42:22,439
Speaker 3: we want to get more and more students and things,

750
00:42:22,440 --> 00:42:25,480
Speaker 3: because it's just starting. But I would say in general,

751
00:42:25,560 --> 00:42:28,879
Speaker 3: most people are aware of it. Could we get more,

752
00:42:29,000 --> 00:42:30,200
Speaker 3: could we accelerate it?

753
00:42:30,440 --> 00:42:30,760
Speaker 2: Yes?

754
00:42:30,880 --> 00:42:32,960
Speaker 3: Do we need to make better hardware, do we need

755
00:42:33,040 --> 00:42:35,560
Speaker 3: to come up with better libraries, yes? Do we need

756
00:42:35,600 --> 00:42:39,040
Speaker 3: better software yes, But it's all happening over the next

757
00:42:39,040 --> 00:42:39,680
Speaker 3: few years.

758
00:42:40,040 --> 00:42:42,239
Speaker 2: Is it hard to get someone who's spent their entire

759
00:42:42,280 --> 00:42:45,520
Speaker 2: life thinking in terms of solving problems to classical means

760
00:42:45,719 --> 00:42:48,040
Speaker 2: to make the transition to this new paradigm.

761
00:42:48,320 --> 00:42:51,160
Speaker 3: There's a lot of examples when you approach something with

762
00:42:51,239 --> 00:42:54,640
Speaker 3: the classical intuition, it's not the right way to do

763
00:42:54,680 --> 00:42:57,879
Speaker 3: it when you approach it through the quantum. But if

764
00:42:57,880 --> 00:43:02,160
Speaker 3: people are being taught to understand the fundamentals of the math,

765
00:43:02,680 --> 00:43:06,279
Speaker 3: then a lot of the techniques carry across. I don't

766
00:43:06,320 --> 00:43:10,200
Speaker 3: recommend people need to learn about entanglement or supersition because

767
00:43:11,000 --> 00:43:15,240
Speaker 3: whilst the physicists will argue like spooky action a distance

768
00:43:15,280 --> 00:43:18,439
Speaker 3: and all these type of things, entanglement is the power. Yes,

769
00:43:18,560 --> 00:43:21,759
Speaker 3: that's how physicists are labeled. How quantum is different. But

770
00:43:21,880 --> 00:43:25,160
Speaker 3: I would say, do we need some physicists really worrying

771
00:43:25,400 --> 00:43:26,040
Speaker 3: thinking about that?

772
00:43:26,200 --> 00:43:26,520
Speaker 1: Yes?

773
00:43:26,600 --> 00:43:30,239
Speaker 3: But We need more applied mathematicians that are realizing they

774
00:43:30,239 --> 00:43:32,640
Speaker 3: can use this as a as a different way of

775
00:43:32,680 --> 00:43:33,640
Speaker 3: looking at the problems.

776
00:43:33,800 --> 00:43:36,560
Speaker 2: Yeah, I when I asked you one question, No, we're

777
00:43:36,560 --> 00:43:39,920
Speaker 2: describing a a It's more than a new technology. We're

778
00:43:39,960 --> 00:43:43,120
Speaker 2: talking about a new paradigm. It's a way of thinking

779
00:43:43,160 --> 00:43:47,400
Speaker 2: about problems. Can you compare this to kind of previous

780
00:43:47,840 --> 00:43:51,879
Speaker 2: technological paradigms. If I'm thinking at the last couple hundred years,

781
00:43:51,880 --> 00:43:55,480
Speaker 2: what does this rank in terms of a new field

782
00:43:55,560 --> 00:43:56,440
Speaker 2: that we've opened up.

783
00:43:57,040 --> 00:43:59,160
Speaker 3: It's a hard question to answer, but I often say

784
00:43:59,320 --> 00:44:02,560
Speaker 3: the history of computing, this will be the first time

785
00:44:03,200 --> 00:44:07,840
Speaker 3: that computation has branched between classical and quantum. I like

786
00:44:08,200 --> 00:44:11,759
Speaker 3: thinking reading a lot in the past. One of the

787
00:44:11,760 --> 00:44:15,239
Speaker 3: things that I think was a way we changed as

788
00:44:15,239 --> 00:44:19,520
Speaker 3: a society was the invention of zero. Before zero, math

789
00:44:19,760 --> 00:44:24,320
Speaker 3: was limited. Realizing that numbers have a number as zero

790
00:44:24,840 --> 00:44:27,680
Speaker 3: allowed us to develop a whole set of new mathematics

791
00:44:28,120 --> 00:44:32,280
Speaker 3: that then went on and defined like everything from waves

792
00:44:32,320 --> 00:44:36,120
Speaker 3: to calculus to all of that. Yes, we can describe

793
00:44:36,160 --> 00:44:38,560
Speaker 3: it with that same math, but when we describe it

794
00:44:38,600 --> 00:44:41,839
Speaker 3: with that math, it gets exponentially big and gets impractical

795
00:44:41,880 --> 00:44:44,920
Speaker 3: to do. Now we can actually work on it. I

796
00:44:44,920 --> 00:44:47,480
Speaker 3: would say, if I had to give you a quick answer,

797
00:44:47,600 --> 00:44:50,440
Speaker 3: maybe going all the way back to when we were

798
00:44:51,120 --> 00:44:51,960
Speaker 3: accepted zero.

799
00:44:52,280 --> 00:44:53,840
Speaker 2: I thought you were going to say, like the airplane,

800
00:44:53,960 --> 00:44:56,840
Speaker 2: but in fact, yeah, you went several orders of magnitude

801
00:44:56,880 --> 00:44:57,239
Speaker 2: beyond that.

802
00:44:57,520 --> 00:45:00,839
Speaker 3: Yes, but I think it's so fundamental.

803
00:45:01,080 --> 00:45:04,480
Speaker 2: This is absolutely fascinating. Thank you so much for chatting

804
00:45:04,560 --> 00:45:05,040
Speaker 2: with me about it.

805
00:45:05,120 --> 00:45:05,799
Speaker 3: Thank you. Fret time.

806
00:45:08,239 --> 00:45:11,400
Speaker 2: Hey listeners. So normally we end this episode here, but

807
00:45:11,480 --> 00:45:14,759
Speaker 2: the Tech Week attendees asked Jay some really great questions,

808
00:45:15,200 --> 00:45:18,000
Speaker 2: questions I wish I'd asked, so we wanted to include

809
00:45:18,000 --> 00:45:19,640
Speaker 2: those here. Enjoy.

810
00:45:21,440 --> 00:45:23,719
Speaker 5: Hi, J, thank you so much for the great presentation.

811
00:45:24,200 --> 00:45:27,800
Speaker 5: My name is Trixie Apiado. I work for Willis Towers Watson,

812
00:45:27,840 --> 00:45:31,799
Speaker 5: an insurance broker. I help seisos identify and quantify their

813
00:45:31,840 --> 00:45:35,840
Speaker 5: cyber risk so they can prepare for threats before they happen.

814
00:45:36,280 --> 00:45:38,960
Speaker 5: And so quantum threats keep me up at night. You

815
00:45:39,080 --> 00:45:43,080
Speaker 5: mentioned so many good problems that quantum can solve. It

816
00:45:43,120 --> 00:45:47,799
Speaker 5: can also break encryptions in our classical computer systems. So

817
00:45:48,920 --> 00:45:53,040
Speaker 5: what safeguards or policies do you implement in your teams

818
00:45:53,400 --> 00:45:57,880
Speaker 5: to build quantum capabilities responsibly and what can we do

819
00:45:58,640 --> 00:46:02,040
Speaker 5: for people in this room, US builders and users to

820
00:46:02,120 --> 00:46:06,560
Speaker 5: secure our data in systems before quantum computers become more

821
00:46:06,640 --> 00:46:08,920
Speaker 5: energy efficient, cheaper, and more available.

822
00:46:09,800 --> 00:46:13,080
Speaker 3: So it's a great question. So yes, one of the

823
00:46:13,120 --> 00:46:17,480
Speaker 3: algorithms for quantum computing is to break our traditional encryption.

824
00:46:18,200 --> 00:46:22,080
Speaker 3: So at IBM Research we were aware of this from

825
00:46:22,160 --> 00:46:26,600
Speaker 3: day one. We've come up with algorithms that we believe

826
00:46:26,800 --> 00:46:29,719
Speaker 3: and have very strong evidence will not be broken by

827
00:46:29,760 --> 00:46:33,680
Speaker 3: a quantum or classical computer, and has selected them. So

828
00:46:34,760 --> 00:46:39,120
Speaker 3: first the scientific technical question, security is saved. There are

829
00:46:39,200 --> 00:46:44,040
Speaker 3: algorithms that exist that we can implement that neither a

830
00:46:44,160 --> 00:46:48,280
Speaker 3: quantum or classical computer can break. So the technical answer

831
00:46:48,400 --> 00:46:52,400
Speaker 3: is we're all okay. The more complicated answer is a

832
00:46:52,440 --> 00:46:57,480
Speaker 3: social and society answer. Encryption was built in classical computing

833
00:46:57,600 --> 00:47:01,320
Speaker 3: in a way that was never thought of being grade it.

834
00:47:01,320 --> 00:47:05,360
Speaker 3: It's mixed everywhere. Some of it is downstream, some of

835
00:47:05,400 --> 00:47:08,000
Speaker 3: it is like software that you may use, Some of

836
00:47:08,080 --> 00:47:12,279
Speaker 3: it is software that you've developed. And I get that

837
00:47:12,400 --> 00:47:14,279
Speaker 3: if you've got a product and you want to have

838
00:47:14,360 --> 00:47:17,319
Speaker 3: it secure for the next ten years, you probably want

839
00:47:17,320 --> 00:47:19,839
Speaker 3: to think about how you're going to upgrade it, or

840
00:47:19,960 --> 00:47:23,160
Speaker 3: if you have data that needs to be secure for

841
00:47:23,360 --> 00:47:26,640
Speaker 3: the next ten years, it needs to upgrade to new encryption.

842
00:47:27,280 --> 00:47:30,320
Speaker 3: So the real challenge is more of a social business

843
00:47:30,360 --> 00:47:34,840
Speaker 3: problem of how do we actually transition from old encryption

844
00:47:34,960 --> 00:47:38,080
Speaker 3: to new encryption knowing this is going to happen. So

845
00:47:38,200 --> 00:47:40,719
Speaker 3: we at IBM have been very proactive on this. We've

846
00:47:40,760 --> 00:47:44,160
Speaker 3: developed tools where we can determine where encryption is used,

847
00:47:44,560 --> 00:47:48,560
Speaker 3: We've developed tools which can show you how to replace it,

848
00:47:48,719 --> 00:47:52,960
Speaker 3: and we early on have made sure the Mainframe when

849
00:47:53,000 --> 00:47:55,560
Speaker 3: we made these algorithms. So I think it was Z

850
00:47:55,719 --> 00:47:58,560
Speaker 3: sixteen that was the first version of the Mainframe to

851
00:47:58,760 --> 00:48:03,759
Speaker 3: have these quantum safe algorithms implemented. So my answer to

852
00:48:03,800 --> 00:48:07,480
Speaker 3: your question is, yes, there's a real problem, but it's

853
00:48:07,520 --> 00:48:10,520
Speaker 3: not a technical problem. It's a social and business problem.

854
00:48:10,520 --> 00:48:14,000
Speaker 3: And I'm not minimizing that. I understand that that is

855
00:48:14,040 --> 00:48:17,000
Speaker 3: a lot of work you need to start now. You

856
00:48:17,080 --> 00:48:19,160
Speaker 3: need to come up and do a you need to

857
00:48:19,200 --> 00:48:22,239
Speaker 3: make it part of your IT transformation. You need to

858
00:48:22,680 --> 00:48:27,200
Speaker 3: get onto it. And I realize, I realize it's not

859
00:48:27,239 --> 00:48:29,960
Speaker 3: going to take zero time because it's not an easy

860
00:48:30,000 --> 00:48:33,279
Speaker 3: problem to do. So the short answer is one we

861
00:48:33,360 --> 00:48:36,440
Speaker 3: developed algorithms that we can't, and we're developing tools to

862
00:48:36,440 --> 00:48:37,600
Speaker 3: help you in that transformation.

863
00:48:38,080 --> 00:48:38,839
Speaker 5: Thank you so much.

864
00:48:40,040 --> 00:48:41,520
Speaker 2: Thank you. My name is Emma.

865
00:48:41,640 --> 00:48:45,239
Speaker 6: I'm a product manager at Expedia, working on software side

866
00:48:45,239 --> 00:48:48,440
Speaker 6: of things. My question is around the non technical roles

867
00:48:48,560 --> 00:48:52,359
Speaker 6: outside of the researchers, the mathematicians, the builders. How can

868
00:48:52,400 --> 00:48:55,680
Speaker 6: the rest of us, whether it be policymakers, those in

869
00:48:55,680 --> 00:48:59,040
Speaker 6: the legal fields, those thinking about what use cases quantum

870
00:48:59,120 --> 00:49:01,440
Speaker 6: can solve for in a few what should we be

871
00:49:01,480 --> 00:49:03,839
Speaker 6: thinking about and how can we prepare for that.

872
00:49:04,160 --> 00:49:06,279
Speaker 3: It's a good question. I think this is part of

873
00:49:06,360 --> 00:49:10,239
Speaker 3: the requirement of the scientists to being able to articulate

874
00:49:10,920 --> 00:49:13,520
Speaker 3: where they are. We need a forum for those type

875
00:49:13,560 --> 00:49:16,600
Speaker 3: of discussions. I think a lot of this can fit

876
00:49:16,719 --> 00:49:19,959
Speaker 3: within the forums that we already have for classical and AI,

877
00:49:20,640 --> 00:49:22,600
Speaker 3: and I think we need to just be asking how

878
00:49:22,640 --> 00:49:25,720
Speaker 3: do we actually bring them into them Because I don't

879
00:49:25,920 --> 00:49:29,120
Speaker 3: think of quantum as a replacement of compute. I think

880
00:49:29,120 --> 00:49:32,120
Speaker 3: of it as an accelerator that expands what is possible,

881
00:49:32,800 --> 00:49:35,560
Speaker 3: and I think we can ask those questions in those forums.

882
00:49:35,960 --> 00:49:38,440
Speaker 3: Are we doing enough now? I think I agree with you. No,

883
00:49:38,719 --> 00:49:40,000
Speaker 3: I don't know the answer to it.

884
00:49:40,880 --> 00:49:44,160
Speaker 6: I think it's a really interesting perspective because those existing

885
00:49:44,239 --> 00:49:47,640
Speaker 6: forums do start to bring in those other fields as well,

886
00:49:47,920 --> 00:49:50,320
Speaker 6: so it could warrant the same sort of discussion.

887
00:49:50,480 --> 00:49:54,720
Speaker 3: And yeah, acts, and I understand those forums. Right now,

888
00:49:56,239 --> 00:49:59,799
Speaker 3: AI is probably dominating and it should be like we

889
00:50:00,120 --> 00:50:04,400
Speaker 3: going through a period of time where AI is impacting society.

890
00:50:04,960 --> 00:50:08,120
Speaker 3: The technology is impacting society in big ways. So I

891
00:50:08,200 --> 00:50:11,800
Speaker 3: totally understand that most of their focus should be on AI,

892
00:50:12,040 --> 00:50:14,640
Speaker 3: but we should start to ask where is quantum in

893
00:50:15,080 --> 00:50:15,759
Speaker 3: that as well?

894
00:50:16,840 --> 00:50:20,560
Speaker 4: Hi, I'm Gobi and I'm a graduating PhD student at

895
00:50:20,640 --> 00:50:24,600
Speaker 4: Northwestern and also a member of south Park Commons, which

896
00:50:24,640 --> 00:50:28,200
Speaker 4: is a fund here. You mentioned earlier that some problems

897
00:50:28,239 --> 00:50:31,200
Speaker 4: are best solved by classical versus some problems are best

898
00:50:31,239 --> 00:50:33,759
Speaker 4: solved by quantum. When we're thinking about this, if we're

899
00:50:33,800 --> 00:50:36,200
Speaker 4: not experts in quantum, but we're thinking about this from

900
00:50:36,200 --> 00:50:38,719
Speaker 4: an AI perspective, could you just clarify when we think

901
00:50:38,719 --> 00:50:43,080
Speaker 4: about quantum, what is deterministic and what is not deterministic.

902
00:50:43,600 --> 00:50:46,120
Speaker 3: I think the future of computing we've got to get

903
00:50:46,120 --> 00:50:49,480
Speaker 3: our heads around is that not everything is deterministic, and

904
00:50:49,520 --> 00:50:52,080
Speaker 3: it's much more going to be probilistic. How do you

905
00:50:52,200 --> 00:50:55,439
Speaker 3: handle error bars? How do you put confidence? I think

906
00:50:55,480 --> 00:50:58,720
Speaker 3: a lot of those questions which you're referring to INAI

907
00:50:58,760 --> 00:51:02,439
Speaker 3: are going to completely imply and quantum. I actually think

908
00:51:02,480 --> 00:51:08,240
Speaker 3: it's a mistake to compare AI verse quantum. I actually

909
00:51:08,280 --> 00:51:12,440
Speaker 3: think of quantum as much. It's quantum verse classical compute,

910
00:51:12,480 --> 00:51:15,480
Speaker 3: and AI is going to come across on top. So

911
00:51:15,800 --> 00:51:18,480
Speaker 3: as we go forward and we get a better understanding thing,

912
00:51:18,680 --> 00:51:22,160
Speaker 3: I'm not going to say quantum is going to replace

913
00:51:22,200 --> 00:51:25,200
Speaker 3: the classical compute that enables AI, but I think some

914
00:51:25,280 --> 00:51:27,520
Speaker 3: of the math you do in AI will be able

915
00:51:27,520 --> 00:51:30,279
Speaker 3: to go to both. So what can we formally prove?

916
00:51:31,080 --> 00:51:34,000
Speaker 3: I can come up with a problem where I take

917
00:51:34,040 --> 00:51:36,440
Speaker 3: a circle and a color, half of it red and

918
00:51:36,520 --> 00:51:39,359
Speaker 3: half oft of blue, and then I say, I'm going

919
00:51:39,400 --> 00:51:43,359
Speaker 3: to apply an operation that takes those dots make it. Say,

920
00:51:43,440 --> 00:51:45,799
Speaker 3: let's say ten dots over here red, ten dots over

921
00:51:45,840 --> 00:51:48,560
Speaker 3: here blue, and I'm going to wind them around many,

922
00:51:48,560 --> 00:51:51,680
Speaker 3: many times. I can then show you that if you

923
00:51:51,760 --> 00:51:54,680
Speaker 3: feed that into a classical computer it's a classical random

924
00:51:54,760 --> 00:51:58,359
Speaker 3: number generator. You can give yourself as much data as

925
00:51:58,400 --> 00:52:00,920
Speaker 3: you want. You will never be able to say did

926
00:52:00,960 --> 00:52:03,879
Speaker 3: the red come from the left side or the right side.

927
00:52:03,920 --> 00:52:07,319
Speaker 3: You would take infinite data. It is like you would

928
00:52:07,360 --> 00:52:10,640
Speaker 3: have to break a classical random number generator. I can

929
00:52:10,680 --> 00:52:13,920
Speaker 3: show you a quantum algorithm that can do that deterministically.

930
00:52:15,000 --> 00:52:17,920
Speaker 3: So where we're thinking is when the data appears to

931
00:52:17,960 --> 00:52:23,080
Speaker 3: be completely unstructured or you looks essentially like a complete

932
00:52:23,160 --> 00:52:27,480
Speaker 3: random number to the classical methods, there are quantum methods

933
00:52:27,520 --> 00:52:30,239
Speaker 3: that can actually potentially find that structure.

934
00:52:34,719 --> 00:52:37,160
Speaker 2: That's it for this episode of Smart Talks with IBM.

935
00:52:37,640 --> 00:52:39,840
Speaker 2: If you haven't already, be sure to check out my

936
00:52:39,920 --> 00:52:45,320
Speaker 2: conversation with IBM Chairman and CEO Arvind Krishna, and stay tuned.

937
00:52:45,719 --> 00:52:50,719
Speaker 2: Another episode is coming soon. Smart Talks with IBM is

938
00:52:50,760 --> 00:52:54,880
Speaker 2: produced by Matt Romano, Amy Gains, McQuaid, Trina Menino, and

939
00:52:55,000 --> 00:52:59,680
Speaker 2: Jake Harper. Engineering by Nina Bird Lawrence, Mastering by Sarah Buguer,

940
00:53:00,120 --> 00:53:05,160
Speaker 2: music by Gramoscope, Strategy by Tatiana Lieberman, Cassidy Meyer and

941
00:53:05,280 --> 00:53:09,440
Speaker 2: Sofia Derlon. Smart Talks with IBM is a production of

942
00:53:09,560 --> 00:53:14,759
Speaker 2: Pushkin Industries and Ruby Studio at iHeartMedia. To find more

943
00:53:14,800 --> 00:53:19,520
Speaker 2: Pushkin podcasts, listen on the iHeartRadio app, Apple Podcasts, or

944
00:53:19,560 --> 00:53:24,040
Speaker 2: wherever you listen to podcasts. I'm Malcolm Godwell. This is

945
00:53:24,080 --> 00:53:28,560
Speaker 2: a paid advertisement from IBM. The conversations on this podcast

946
00:53:28,840 --> 00:53:38,600
Speaker 2: don't necessarily represent IBM's positions, strategies, or opinions.