Building bridges between quantum and classical computing

WEBVTT
1
00:00:08.000 --> 00:00:11.359
- Hello, and welcome to the Physics World weekly
2
00:00:11.359 --> 00:00:11.859
- podcast.
3
00:00:13.174 --> 00:00:15.974
- I'm Hamish Johnston, and my guest in this
4
00:00:15.974 --> 00:00:16.474
- episode
5
00:00:16.774 --> 00:00:17.754
- is Yonatan
6
00:00:18.054 --> 00:00:18.554
- Cohen,
7
00:00:18.934 --> 00:00:19.994
- who is cofounder
8
00:00:20.375 --> 00:00:21.754
- and chief technology
9
00:00:22.134 --> 00:00:22.634
- officer
10
00:00:23.094 --> 00:00:24.474
- of Quantum Machines.
11
00:00:25.589 --> 00:00:28.969
- The company was started in Israel in 2018,
12
00:00:29.350 --> 00:00:30.890
- and it now has facilities
13
00:00:31.429 --> 00:00:31.929
- worldwide.
14
00:00:32.869 --> 00:00:34.489
- As Yonatan explains,
15
00:00:35.109 --> 00:00:37.850
- Quantum Machines develops control systems
16
00:00:38.355 --> 00:00:39.575
- for quantum computers,
17
00:00:40.195 --> 00:00:43.975
- providing a link between the quantum and classical
18
00:00:44.274 --> 00:00:44.774
- worlds.
19
00:00:53.370 --> 00:00:54.270
- Hi, Yonatan.
20
00:00:54.729 --> 00:00:55.950
- Welcome to the podcast.
21
00:00:56.969 --> 00:00:58.990
- Hi, Hamish. Thanks for having me.
22
00:00:59.690 --> 00:01:02.510
- So, Yonatan, can you give us a brief
23
00:01:02.570 --> 00:01:04.109
- introduction to the company?
24
00:01:04.810 --> 00:01:07.564
- When was it founded? How many employees do
25
00:01:07.564 --> 00:01:10.704
- you have? And what's the main business focus?
26
00:01:12.525 --> 00:01:14.305
- Yeah. So at Quantum Machines,
27
00:01:14.605 --> 00:01:17.965
- we are focused on, accelerating the realization of
28
00:01:17.965 --> 00:01:19.265
- useful quantum computers
29
00:01:19.780 --> 00:01:20.519
- by providing,
30
00:01:22.659 --> 00:01:25.719
- the control systems for quantum computers.
31
00:01:26.019 --> 00:01:28.899
- We're providing control systems for many, many different
32
00:01:28.899 --> 00:01:30.680
- players in the quantum industry.
33
00:01:31.299 --> 00:01:32.659
- The control system is the,
34
00:01:33.305 --> 00:01:35.305
- one of the core elements of a quantum
35
00:01:35.305 --> 00:01:38.105
- computer. It's a system it's a classical system,
36
00:01:38.105 --> 00:01:40.504
- not the quantum system, but the classical system
37
00:01:40.504 --> 00:01:42.444
- that operates the quantum hardware,
38
00:01:42.984 --> 00:01:46.105
- and orchestrates the the quantum operations that that
39
00:01:46.105 --> 00:01:47.885
- we perform on the quantum hardware.
40
00:01:48.430 --> 00:01:50.369
- We were founded in 2018.
41
00:01:51.150 --> 00:01:52.369
- We're three founders,
42
00:01:53.229 --> 00:01:53.729
- physicists.
43
00:01:55.150 --> 00:01:56.450
- We all did our PhDs,
44
00:01:56.909 --> 00:01:58.909
- together in the same lab at the Weizmann
45
00:01:58.909 --> 00:02:00.369
- Institute, in Israel.
46
00:02:00.834 --> 00:02:01.334
- And,
47
00:02:01.875 --> 00:02:02.194
- and,
48
00:02:02.915 --> 00:02:04.375
- yeah. And today, we are,
49
00:02:05.075 --> 00:02:06.854
- almost more than 300 people,
50
00:02:07.474 --> 00:02:10.775
- around the world. We're spreading over, 21 countries,
51
00:02:11.555 --> 00:02:14.754
- and we, have, more than 500 customers around
52
00:02:14.754 --> 00:02:16.789
- the world using our products. And,
53
00:02:18.289 --> 00:02:19.590
- yeah, it's very exciting.
54
00:02:20.530 --> 00:02:22.789
- And and why is it important to
55
00:02:23.169 --> 00:02:25.109
- integrate quantum and
56
00:02:25.489 --> 00:02:26.549
- classical systems?
57
00:02:28.449 --> 00:02:29.729
- Why do you have to do that to
58
00:02:29.729 --> 00:02:30.469
- do effective
59
00:02:30.885 --> 00:02:33.205
- quantum computing? And and what are some of
60
00:02:33.205 --> 00:02:34.665
- the challenges involved?
61
00:02:36.165 --> 00:02:37.625
- Yeah. So to begin with,
62
00:02:38.004 --> 00:02:39.224
- you know, we are classical
63
00:02:39.685 --> 00:02:40.185
- beings.
64
00:02:40.645 --> 00:02:42.485
- We live in the classical world, and we
65
00:02:42.485 --> 00:02:44.105
- need something to that bridges,
66
00:02:44.560 --> 00:02:46.419
- you know, our classical language,
67
00:02:47.199 --> 00:02:49.539
- and controls quantum quantum
68
00:02:50.000 --> 00:02:52.879
- hardware, at the quantum level. So that's that's
69
00:02:52.879 --> 00:02:55.439
- the main task of the control systems. Again,
70
00:02:55.439 --> 00:02:56.819
- it's a classical system
71
00:02:57.305 --> 00:02:59.944
- that but that interfaces the quantum hardware. So
72
00:02:59.944 --> 00:03:02.905
- it's really sits in this interface between quantum
73
00:03:02.905 --> 00:03:05.305
- and the the quantum hardware and the the
74
00:03:05.305 --> 00:03:06.284
- classical world.
75
00:03:06.905 --> 00:03:09.405
- So that's the first thing. But besides that,
76
00:03:09.784 --> 00:03:12.284
- it's very important these days to also connect
77
00:03:12.540 --> 00:03:13.439
- quantum processing,
78
00:03:14.300 --> 00:03:15.760
- with classical processing,
79
00:03:16.780 --> 00:03:19.500
- because almost everything we do on a quantum
80
00:03:19.500 --> 00:03:22.300
- computer involves some classical processing as well. So
81
00:03:22.300 --> 00:03:23.599
- it it starts from,
82
00:03:24.139 --> 00:03:26.835
- you know, bringing up, booting the quantum computer
83
00:03:26.835 --> 00:03:28.455
- where we run a sequence of
84
00:03:28.835 --> 00:03:30.995
- quantum programs, but then we take some data
85
00:03:30.995 --> 00:03:32.995
- from the quantum processor, and then we analyze
86
00:03:32.995 --> 00:03:34.375
- it on classical processors,
87
00:03:34.754 --> 00:03:36.935
- CPUs, GPUs, etcetera, to extract,
88
00:03:38.290 --> 00:03:39.030
- the parameters,
89
00:03:40.290 --> 00:03:43.010
- of of of the quantum system. And based
90
00:03:43.010 --> 00:03:45.650
- on that, we, we call it calibration of
91
00:03:45.650 --> 00:03:48.210
- the quantum computer. So we calibrate the quantum
92
00:03:48.210 --> 00:03:50.745
- computer. So all the way from calibrations
93
00:03:51.205 --> 00:03:53.784
- through then you have, things like error correction,
94
00:03:54.324 --> 00:03:56.745
- where you take data from the quantum processor
95
00:03:56.805 --> 00:03:59.364
- in real time, and then you, crunch the
96
00:03:59.364 --> 00:04:01.384
- numbers on the on the classical processor
97
00:04:02.485 --> 00:04:03.944
- to decode the errors,
98
00:04:04.800 --> 00:04:06.800
- and understand what what errors you have on
99
00:04:06.800 --> 00:04:07.860
- your quantum hardware,
100
00:04:08.159 --> 00:04:10.560
- in your quantum algorithm and fix them. And
101
00:04:10.560 --> 00:04:12.479
- all the way up to applications where at
102
00:04:12.479 --> 00:04:13.060
- the end,
103
00:04:14.960 --> 00:04:17.519
- many of the applications are hybrid applications by
104
00:04:17.519 --> 00:04:20.055
- themselves where the quantum processor only does,
105
00:04:21.014 --> 00:04:21.754
- a subtask,
106
00:04:23.495 --> 00:04:26.214
- sub subtask in the entire quantum application or
107
00:04:26.214 --> 00:04:27.514
- quantum classical application,
108
00:04:27.975 --> 00:04:30.454
- we should say. So everywhere from you see
109
00:04:30.454 --> 00:04:32.615
- from the calibration through the error equation all
110
00:04:32.615 --> 00:04:33.834
- the way up to the applications,
111
00:04:34.294 --> 00:04:37.060
- we need this kind of integration between quantum
112
00:04:37.060 --> 00:04:39.060
- and classical processing. We need to do some
113
00:04:39.060 --> 00:04:39.879
- quantum processing,
114
00:04:40.419 --> 00:04:42.339
- and then the results go to a classical
115
00:04:42.339 --> 00:04:44.500
- processor. We do some classical processing, and then
116
00:04:44.500 --> 00:04:46.579
- the result go again to the quantum processor.
117
00:04:46.579 --> 00:04:49.639
- So this interleaving of quantum and classical processing,
118
00:04:50.254 --> 00:04:51.615
- we see it all over the stack. So
119
00:04:51.615 --> 00:04:53.475
- it's very important to integrate
120
00:04:54.495 --> 00:04:54.995
- correctly,
121
00:04:55.894 --> 00:04:56.514
- these two,
122
00:04:57.535 --> 00:05:00.115
- different types of of, compute resources
123
00:05:00.574 --> 00:05:02.115
- in a way that we can orchestrate,
124
00:05:02.735 --> 00:05:03.235
- workflows,
125
00:05:04.279 --> 00:05:07.319
- of these quantum and classical processing tasks,
126
00:05:07.800 --> 00:05:08.620
- on the system.
127
00:05:09.480 --> 00:05:11.800
- And and I would imagine that there's there's
128
00:05:11.800 --> 00:05:14.040
- some big challenges here. I mean, I'm I'm
129
00:05:14.040 --> 00:05:16.120
- assuming one of the biggest challenges is the
130
00:05:16.120 --> 00:05:16.939
- fact that
131
00:05:17.605 --> 00:05:18.105
- quantum,
132
00:05:18.564 --> 00:05:21.444
- a quantum processor is a very delicate thing.
133
00:05:21.444 --> 00:05:23.285
- You know, it could be, you know, for
134
00:05:23.285 --> 00:05:24.824
- example, based on superconductors
135
00:05:25.204 --> 00:05:28.805
- that are cooled to near absolute zero. And,
136
00:05:28.805 --> 00:05:29.305
- actually,
137
00:05:29.910 --> 00:05:32.710
- just connecting to those things and getting the
138
00:05:32.710 --> 00:05:34.330
- information in and out
139
00:05:34.710 --> 00:05:36.889
- in an efficient way is
140
00:05:37.830 --> 00:05:39.430
- is is a big challenge. I mean, is
141
00:05:39.430 --> 00:05:41.509
- that something that you spend a lot of
142
00:05:41.509 --> 00:05:42.009
- time
143
00:05:42.324 --> 00:05:44.745
- thinking about? Or or are there other
144
00:05:45.285 --> 00:05:45.785
- important,
145
00:05:46.645 --> 00:05:47.865
- factors as well?
146
00:05:48.964 --> 00:05:51.925
- No. That's absolutely right. That's the first, that's
147
00:05:51.925 --> 00:05:54.500
- that's what we call the the analog front
148
00:05:54.500 --> 00:05:56.439
- end or sometimes we call it the quantum
149
00:05:56.579 --> 00:05:58.819
- the QPU front end. The front end of
150
00:05:58.819 --> 00:05:59.639
- our system,
151
00:06:00.339 --> 00:06:03.220
- where it integrates with the quantum system, that's
152
00:06:03.220 --> 00:06:04.839
- one of our our, main,
153
00:06:05.220 --> 00:06:05.800
- main, technology
154
00:06:06.500 --> 00:06:08.439
- developments and and main challenges.
155
00:06:08.854 --> 00:06:10.294
- And as you said, like, this is where
156
00:06:10.294 --> 00:06:11.274
- we really produce,
157
00:06:11.735 --> 00:06:13.035
- very delicate signals.
158
00:06:13.415 --> 00:06:15.415
- They need to be very clean signals. They're
159
00:06:15.415 --> 00:06:17.975
- still classical signals, but once they interact with
160
00:06:17.975 --> 00:06:18.714
- the qubits,
161
00:06:19.095 --> 00:06:21.514
- or the quantum elements on the quantum processor,
162
00:06:21.860 --> 00:06:23.939
- then they perform these quantum operations. And if
163
00:06:23.939 --> 00:06:24.600
- the signals
164
00:06:25.060 --> 00:06:27.079
- that our control system produces,
165
00:06:27.779 --> 00:06:29.240
- are not clean enough,
166
00:06:30.740 --> 00:06:33.779
- and, not shaped correctly, and and and the
167
00:06:33.779 --> 00:06:35.319
- noise level is too high,
168
00:06:36.345 --> 00:06:38.845
- then you then the basic quantum operations,
169
00:06:39.384 --> 00:06:41.225
- are not accurate enough. We call it the
170
00:06:41.225 --> 00:06:43.464
- fidelity of the operation. So the fidelity of
171
00:06:43.464 --> 00:06:44.525
- our quantum operations
172
00:06:45.064 --> 00:06:46.925
- would, could be actually affected
173
00:06:47.384 --> 00:06:49.464
- by the control system if the control system
174
00:06:49.464 --> 00:06:50.365
- is not designed,
175
00:06:50.769 --> 00:06:51.269
- correctly.
176
00:06:51.970 --> 00:06:53.970
- So and the same in the readout, you
177
00:06:53.970 --> 00:06:56.449
- mentioned, you know, both getting data in into
178
00:06:56.449 --> 00:06:58.769
- our qubits and then taking data back from
179
00:06:58.769 --> 00:07:00.290
- the qubits. How do we how do you
180
00:07:00.290 --> 00:07:01.269
- perform the readout
181
00:07:01.649 --> 00:07:02.629
- of your qubits?
182
00:07:03.165 --> 00:07:04.384
- How do you measure them,
183
00:07:05.004 --> 00:07:07.564
- with, the least amount of interference with your
184
00:07:07.564 --> 00:07:08.625
- quantum processor?
185
00:07:09.084 --> 00:07:11.345
- Again, you need to, have a very,
186
00:07:11.964 --> 00:07:12.845
- well thought of,
187
00:07:13.725 --> 00:07:15.105
- front end of the system.
188
00:07:15.644 --> 00:07:16.384
- And this
189
00:07:16.764 --> 00:07:18.509
- has has a lot to do with,
190
00:07:19.870 --> 00:07:22.050
- you know, generating and also,
191
00:07:23.310 --> 00:07:24.670
- acquiring and and and,
192
00:07:25.230 --> 00:07:26.129
- analog signals.
193
00:07:26.670 --> 00:07:28.189
- So this is the way we call it
194
00:07:28.189 --> 00:07:30.350
- the analog front end. So that's one big
195
00:07:30.350 --> 00:07:33.074
- challenge that we face. And, also, that challenge
196
00:07:33.134 --> 00:07:35.375
- becomes, like, 10 times harder, a 100 times
197
00:07:35.375 --> 00:07:38.495
- harder as you scale. Because today, you also
198
00:07:38.495 --> 00:07:40.754
- wanna scale the quantum hardware.
199
00:07:41.134 --> 00:07:43.294
- So, also, the analog front end of the
200
00:07:43.294 --> 00:07:45.474
- control system needs to scale dramatically.
201
00:07:46.009 --> 00:07:47.610
- And so not only are we trying to
202
00:07:47.610 --> 00:07:50.009
- generate very, very clean signals, we're trying to
203
00:07:50.009 --> 00:07:52.649
- generate many, many, many of these signals. And
204
00:07:52.649 --> 00:07:54.889
- you you wanna lower the cost per quick
205
00:07:54.889 --> 00:07:57.470
- qubit control. You wanna lower the power consumption
206
00:07:57.689 --> 00:08:00.350
- per qubit control. You wanna, lower the size
207
00:08:00.410 --> 00:08:01.470
- per qubit control.
208
00:08:01.814 --> 00:08:03.735
- All the, you know, things that you always,
209
00:08:04.055 --> 00:08:05.915
- do when you scale up a a technology,
210
00:08:06.455 --> 00:08:08.134
- while at the same time, you have to
211
00:08:08.134 --> 00:08:08.634
- keep,
212
00:08:09.334 --> 00:08:11.915
- producing very, very accurate signals.
213
00:08:12.855 --> 00:08:14.154
- So that's a big challenge.
214
00:08:15.180 --> 00:08:16.800
- Other than that, there are other challenges.
215
00:08:17.419 --> 00:08:19.979
- For example, how do you connect that control
216
00:08:19.979 --> 00:08:20.479
- system,
217
00:08:20.939 --> 00:08:22.000
- again, with classical,
218
00:08:22.699 --> 00:08:24.399
- high performance compute resources?
219
00:08:25.259 --> 00:08:26.720
- So how do you build the architecture
220
00:08:27.259 --> 00:08:29.504
- that the the chip that orchestrates
221
00:08:29.884 --> 00:08:32.845
- those analog pulses, analog signals that we send
222
00:08:32.845 --> 00:08:34.304
- to our quantum processor?
223
00:08:34.845 --> 00:08:36.225
- That chip needs to also,
224
00:08:36.845 --> 00:08:38.865
- communicate with classical processors.
225
00:08:39.404 --> 00:08:41.644
- What's the architecture of that chip? How do
226
00:08:41.644 --> 00:08:43.664
- we sequence those quantum operations?
227
00:08:44.250 --> 00:08:46.330
- How do we communicate with other,
228
00:08:46.970 --> 00:08:49.209
- compute resources in the in the in the
229
00:08:49.209 --> 00:08:50.029
- in the stack
230
00:08:50.570 --> 00:08:51.629
- in the right way?
231
00:08:52.009 --> 00:08:53.610
- As I mentioned, it will allow us to
232
00:08:53.610 --> 00:08:56.830
- orchestrate these these complex workflows. That's another,
233
00:08:57.210 --> 00:08:59.629
- whole, different channels challenge,
234
00:09:00.584 --> 00:09:02.664
- which is, so there are there are really
235
00:09:02.664 --> 00:09:04.424
- challenges across the stack. And then, of course,
236
00:09:04.424 --> 00:09:06.345
- you wanna create, like, good software on top
237
00:09:06.345 --> 00:09:08.904
- of it and good user experience because a
238
00:09:08.904 --> 00:09:10.985
- lot of a lot of what's happening in
239
00:09:10.985 --> 00:09:12.985
- the field is still, you know, of course,
240
00:09:12.985 --> 00:09:14.044
- research and development.
241
00:09:14.589 --> 00:09:17.089
- And it's very important that we bring,
242
00:09:17.710 --> 00:09:18.690
- you know, good,
243
00:09:19.789 --> 00:09:23.250
- good, user experience, what we call developer experience
244
00:09:24.190 --> 00:09:26.909
- to people that develop quantum computers to shorten
245
00:09:26.909 --> 00:09:28.210
- the r and d cycle
246
00:09:29.014 --> 00:09:30.715
- and give them a lot of flexibility.
247
00:09:31.495 --> 00:09:33.575
- Every day, there is a new paper to
248
00:09:33.654 --> 00:09:35.254
- we can do decoding this way. We can
249
00:09:35.254 --> 00:09:36.695
- do error correction this way. We can do
250
00:09:36.695 --> 00:09:39.654
- calibration that way. So keeping the the system
251
00:09:39.654 --> 00:09:40.875
- also very flexible,
252
00:09:41.990 --> 00:09:43.289
- is very, very important.
253
00:09:44.230 --> 00:09:45.370
- And and you understand,
254
00:09:45.830 --> 00:09:48.330
- my understanding is when you were doing academic
255
00:09:48.470 --> 00:09:49.850
- research, you were looking
256
00:09:50.470 --> 00:09:53.690
- at qubits. Did did you realize then that
257
00:09:53.830 --> 00:09:56.090
- when you were doing your research that there's
258
00:09:56.384 --> 00:09:58.785
- a gap in the market here? That, okay,
259
00:09:58.785 --> 00:10:01.684
- there's people trying to build really good qubits,
260
00:10:01.825 --> 00:10:05.665
- but maybe people weren't thinking too much about,
261
00:10:05.665 --> 00:10:07.925
- you know, the interface with the real world,
262
00:10:08.144 --> 00:10:10.165
- and that's where you and your colleagues
263
00:10:10.625 --> 00:10:11.684
- got your ideas
264
00:10:13.159 --> 00:10:15.639
- to to to form a company. Is is
265
00:10:15.639 --> 00:10:17.980
- that was that the the eureka moment?
266
00:10:19.639 --> 00:10:21.740
- Yeah. So, we are three founders,
267
00:10:22.600 --> 00:10:23.980
- and, one of us,
268
00:10:24.360 --> 00:10:25.820
- Nissim, my cofounder,
269
00:10:26.279 --> 00:10:28.375
- did his post doc at Yale. So he
270
00:10:28.375 --> 00:10:31.495
- actually finished his PhD a few years, before
271
00:10:31.495 --> 00:10:33.495
- Itamar and I, and he went to do
272
00:10:33.495 --> 00:10:34.714
- his postdoc at Yale,
273
00:10:35.815 --> 00:10:37.975
- with, probably, you know, the the sort of
274
00:10:37.975 --> 00:10:38.714
- most famous,
275
00:10:39.095 --> 00:10:39.595
- superconducting
276
00:10:39.975 --> 00:10:40.475
- qubits,
277
00:10:41.014 --> 00:10:41.514
- group
278
00:10:42.089 --> 00:10:42.990
- in in academia,
279
00:10:44.009 --> 00:10:46.490
- at Yale, with professor Rob Shulkov and and,
280
00:10:46.970 --> 00:10:49.370
- and professor Michel de Verere that just won
281
00:10:49.370 --> 00:10:51.709
- the Nobel Prize recently in physics.
282
00:10:52.570 --> 00:10:54.269
- And when he did his postdoc,
283
00:10:54.970 --> 00:10:56.684
- that's he worked on,
284
00:10:57.164 --> 00:10:58.865
- on the quantum error correction.
285
00:10:59.245 --> 00:11:00.384
- And he had to develop,
286
00:11:00.924 --> 00:11:03.184
- a lot of things in the control system
287
00:11:03.245 --> 00:11:05.745
- over there in order to be able to
288
00:11:05.804 --> 00:11:08.065
- enable this kind of experiments. So,
289
00:11:09.029 --> 00:11:10.490
- yes. When when when,
290
00:11:11.269 --> 00:11:12.250
- when we started,
291
00:11:13.029 --> 00:11:15.450
- when when when when we started searching,
292
00:11:15.830 --> 00:11:16.730
- for our startup,
293
00:11:17.750 --> 00:11:18.809
- Itamar and I,
294
00:11:19.350 --> 00:11:20.090
- who we,
295
00:11:20.629 --> 00:11:22.549
- we called Nissim at some point when we
296
00:11:22.549 --> 00:11:24.425
- start started focusing on quantum.
297
00:11:24.805 --> 00:11:27.285
- And then, we interviewed him, and we interviewed
298
00:11:27.524 --> 00:11:29.845
- we knew his work really well, but we
299
00:11:29.845 --> 00:11:32.165
- interviewed also him and what's what's happening in
300
00:11:32.165 --> 00:11:34.004
- the lab. And we interviewed other people that
301
00:11:34.004 --> 00:11:35.370
- that did similar works.
302
00:11:36.809 --> 00:11:38.909
- And then we, yes, we realized that,
303
00:11:39.610 --> 00:11:41.289
- not only do we sort of have a
304
00:11:41.289 --> 00:11:43.929
- sort of an unfair advantage because this really
305
00:11:43.929 --> 00:11:46.490
- brought a lot of knowledge in, the control
306
00:11:46.490 --> 00:11:46.990
- systems,
307
00:11:47.544 --> 00:11:48.044
- But,
308
00:11:48.504 --> 00:11:51.544
- also, we, we realized that, yes, there's going
309
00:11:51.544 --> 00:11:53.485
- to be a serious unmet need
310
00:11:54.024 --> 00:11:54.524
- for,
311
00:11:54.904 --> 00:11:58.205
- next generation control systems, for quantum control,
312
00:11:59.065 --> 00:12:00.845
- because people were using,
313
00:12:01.610 --> 00:12:04.089
- things that were not dedicated for quantum at
314
00:12:04.089 --> 00:12:04.830
- that time.
315
00:12:05.529 --> 00:12:07.529
- And even the ones who did who did
316
00:12:07.529 --> 00:12:08.750
- develop their own
317
00:12:09.209 --> 00:12:10.589
- quantum control systems,
318
00:12:11.289 --> 00:12:13.450
- were still basing it on on sort of
319
00:12:13.450 --> 00:12:16.355
- very simple architectures that we thought are not
320
00:12:16.355 --> 00:12:17.335
- going to give,
321
00:12:17.715 --> 00:12:18.215
- enough,
322
00:12:19.554 --> 00:12:21.955
- flexibility in what's needed in order to to
323
00:12:21.955 --> 00:12:24.595
- to build, for instance, quantum error correction or
324
00:12:24.595 --> 00:12:26.274
- other things that that would be needed in
325
00:12:26.274 --> 00:12:28.769
- the field. So, yes, we we that that
326
00:12:29.090 --> 00:12:31.009
- so it wasn't in our PhD, but it
327
00:12:31.009 --> 00:12:32.529
- was in in Nissim's postdoc,
328
00:12:33.009 --> 00:12:34.690
- based on on on what he did in
329
00:12:34.690 --> 00:12:36.690
- his postdoc. We that was the the the
330
00:12:36.690 --> 00:12:37.670
- beginning of
331
00:12:38.210 --> 00:12:38.769
- the the,
332
00:12:39.410 --> 00:12:42.465
- the understanding that that, that there is a
333
00:12:42.785 --> 00:12:45.425
- potential serious potential here to to to build
334
00:12:45.425 --> 00:12:47.504
- something which will be very valuable for the
335
00:12:47.504 --> 00:12:48.004
- community.
336
00:12:48.865 --> 00:12:51.665
- And you mentioned superconducting qubits, and I I
337
00:12:51.665 --> 00:12:52.404
- I think that's
338
00:12:52.785 --> 00:12:53.285
- mostly
339
00:12:54.225 --> 00:12:56.085
- what, quantum machines
340
00:12:56.625 --> 00:12:59.009
- works with. But there are other, you know,
341
00:12:59.009 --> 00:13:02.149
- sort of qubit designs. There's trapped ions,
342
00:13:02.850 --> 00:13:04.149
- spins in solids.
343
00:13:04.690 --> 00:13:05.090
- And,
344
00:13:06.049 --> 00:13:07.970
- I don't think it's clear which of these
345
00:13:07.970 --> 00:13:10.070
- qubit technologies will ultimately
346
00:13:11.009 --> 00:13:11.750
- be useful.
347
00:13:13.284 --> 00:13:15.544
- Does this make it difficult to develop
348
00:13:16.804 --> 00:13:18.584
- quantum classical integration
349
00:13:19.445 --> 00:13:21.784
- technologies, you know, looking to the future,
350
00:13:22.725 --> 00:13:24.985
- when, you know, you don't really know,
351
00:13:25.365 --> 00:13:27.625
- you know, ten, twenty years down the road,
352
00:13:27.764 --> 00:13:28.139
- which
353
00:13:28.779 --> 00:13:31.899
- qubits people will be working with? Or or
354
00:13:31.899 --> 00:13:33.279
- is there sort of a universal
355
00:13:33.980 --> 00:13:35.920
- nature to what you're developing
356
00:13:36.299 --> 00:13:38.320
- in terms of qubit technology?
357
00:13:39.820 --> 00:13:41.740
- Yeah. So that's a great question. So first
358
00:13:41.740 --> 00:13:43.820
- of all, at Quantum Machines, we are working
359
00:13:43.820 --> 00:13:44.115
- with
360
00:13:44.754 --> 00:13:45.654
- many different,
361
00:13:46.274 --> 00:13:46.774
- QPUs,
362
00:13:47.235 --> 00:13:49.634
- many different qubit types. We have customers in
363
00:13:49.634 --> 00:13:51.394
- in all of the qubit types that you
364
00:13:51.394 --> 00:13:53.794
- mentioned and actually more. So we work with
365
00:13:53.794 --> 00:13:56.754
- trapped ions, quantum computers. We work we, we
366
00:13:56.754 --> 00:13:57.254
- support
367
00:13:57.659 --> 00:14:00.459
- neutral atoms based quantum computers. We have many
368
00:14:00.459 --> 00:14:03.200
- customers in in neutral atoms based quantum computers,
369
00:14:04.299 --> 00:14:07.200
- superconducting qubits, as you mentioned, spin qubits,
370
00:14:08.700 --> 00:14:09.179
- and,
371
00:14:09.579 --> 00:14:12.459
- and and more and more, NVCenters and and
372
00:14:12.459 --> 00:14:13.279
- and others.
373
00:14:13.634 --> 00:14:14.134
- So,
374
00:14:14.675 --> 00:14:17.394
- yeah, when we started thinking about how we
375
00:14:17.394 --> 00:14:20.134
- wanna build our control system and our control,
376
00:14:20.675 --> 00:14:21.815
- solutions and platforms,
377
00:14:22.355 --> 00:14:24.035
- we, yeah, we we one of the goals
378
00:14:24.035 --> 00:14:26.375
- we set for ourselves is to provide solutions
379
00:14:26.434 --> 00:14:28.455
- to all of these different qubit types.
380
00:14:28.820 --> 00:14:30.419
- And as you said, it's it's not a
381
00:14:30.419 --> 00:14:34.179
- trivial, decision because, there are differences between those
382
00:14:34.179 --> 00:14:36.980
- qubit types. So there are obviously differences also
383
00:14:36.980 --> 00:14:38.200
- in the control system,
384
00:14:38.820 --> 00:14:39.320
- requirements,
385
00:14:39.860 --> 00:14:41.480
- from one qubit to another.
386
00:14:41.865 --> 00:14:43.565
- However, there are also a lot of commonalities.
387
00:14:44.585 --> 00:14:46.125
- And so we were able
388
00:14:46.585 --> 00:14:47.085
- with,
389
00:14:47.945 --> 00:14:48.605
- you know,
390
00:14:49.225 --> 00:14:50.924
- a single control platform,
391
00:14:52.345 --> 00:14:55.004
- that does have different features and different,
392
00:14:55.850 --> 00:14:58.410
- front end modules, for example, for different qubit
393
00:14:58.410 --> 00:15:01.070
- types, but still a single control platform
394
00:15:01.610 --> 00:15:03.929
- where I think at least 50% of the
395
00:15:03.929 --> 00:15:04.429
- development
396
00:15:04.809 --> 00:15:07.210
- is is common. At least 50% of the
397
00:15:07.210 --> 00:15:08.429
- platform is common,
398
00:15:09.615 --> 00:15:11.375
- to to different cubic,
399
00:15:11.774 --> 00:15:14.675
- types. We're able to to really provide solutions,
400
00:15:14.975 --> 00:15:17.055
- to all of this. So, yes, it cost
401
00:15:17.055 --> 00:15:19.455
- in extra development and extra effort, but I
402
00:15:19.455 --> 00:15:20.995
- think it also has huge,
403
00:15:21.455 --> 00:15:22.355
- huge advantage.
404
00:15:22.690 --> 00:15:24.389
- Besides the fact that we're not betting
405
00:15:24.769 --> 00:15:26.610
- on who's going to be the the winner
406
00:15:26.610 --> 00:15:27.110
- yet,
407
00:15:27.889 --> 00:15:30.050
- which is great from a business perspective at
408
00:15:30.050 --> 00:15:31.509
- this stage, I believe, because,
409
00:15:32.210 --> 00:15:34.050
- yeah, we don't know yet who's going to
410
00:15:34.050 --> 00:15:35.970
- be the the the definite winner or if
411
00:15:35.970 --> 00:15:37.570
- there is going to be even a definite
412
00:15:37.570 --> 00:15:38.054
- winner.
413
00:15:39.014 --> 00:15:40.855
- But besides that, I think that also there
414
00:15:40.855 --> 00:15:42.315
- is a lot of value in,
415
00:15:43.095 --> 00:15:44.955
- you know, even just, like, understanding
416
00:15:45.414 --> 00:15:47.335
- what should be common to all the qubits,
417
00:15:47.335 --> 00:15:49.274
- how to build such an architecture,
418
00:15:49.894 --> 00:15:52.534
- where are the interfaces that are agnostic, where
419
00:15:52.534 --> 00:15:54.269
- are the interfaces that are not
420
00:15:54.649 --> 00:15:56.350
- agnostic? And so on and so forth.
421
00:15:57.929 --> 00:15:59.870
- And at the end of the day, also,
422
00:16:00.250 --> 00:16:04.250
- everyone to everyone wants to eventually integrate those
423
00:16:04.250 --> 00:16:06.190
- different types of quantum processors
424
00:16:07.049 --> 00:16:09.070
- to the the general data center.
425
00:16:09.544 --> 00:16:12.264
- So by building a platform where, you know,
426
00:16:12.264 --> 00:16:13.245
- the the higher,
427
00:16:14.184 --> 00:16:16.444
- layers of of our stack are agnostic
428
00:16:16.824 --> 00:16:19.304
- sort of agnostic, and the lower layers of
429
00:16:19.304 --> 00:16:22.105
- our stack are less agnostic and more specific
430
00:16:22.105 --> 00:16:24.750
- per qubit type, We are actually creating this
431
00:16:24.750 --> 00:16:27.730
- bridge between all of these, different qubit types
432
00:16:28.029 --> 00:16:30.690
- to, how you can integrate with a single
433
00:16:30.750 --> 00:16:32.929
- control layer to, the general
434
00:16:33.309 --> 00:16:34.450
- data center.
435
00:16:35.789 --> 00:16:38.125
- And that's something that is is very I
436
00:16:38.125 --> 00:16:40.305
- think, would become very valuable in the fields.
437
00:16:40.365 --> 00:16:41.965
- Right? At the end of the day, when
438
00:16:41.965 --> 00:16:45.245
- you integrate quantum computers to high performance compute
439
00:16:45.245 --> 00:16:47.884
- centers or to the, you know, big data
440
00:16:47.884 --> 00:16:50.465
- centers, big cloud cloud providers, the infrastructure,
441
00:16:51.019 --> 00:16:52.940
- You would want to talk to them with
442
00:16:52.940 --> 00:16:54.779
- a single language from a single,
443
00:16:55.340 --> 00:16:55.840
- platform.
444
00:16:56.940 --> 00:16:59.019
- I see. And so, I mean, it almost
445
00:16:59.019 --> 00:16:59.759
- sounds like
446
00:17:00.220 --> 00:17:02.720
- your vision of the future is that
447
00:17:03.514 --> 00:17:05.595
- there there could be you know, we could
448
00:17:05.595 --> 00:17:08.714
- be using multiple qubit technologies. You know, for
449
00:17:08.714 --> 00:17:10.095
- example, we could be using
450
00:17:10.394 --> 00:17:12.414
- trapped ion qubits to solve
451
00:17:12.795 --> 00:17:14.815
- a certain part of a problem, whereas
452
00:17:15.220 --> 00:17:16.039
- maybe superconducting
453
00:17:16.340 --> 00:17:17.559
- qubits are better
454
00:17:17.940 --> 00:17:20.500
- at solving another part of that problem, and
455
00:17:20.500 --> 00:17:22.600
- I don't know, linear optical computing
456
00:17:23.220 --> 00:17:25.559
- can do something else. I mean, is
457
00:17:25.940 --> 00:17:26.500
- is that,
458
00:17:27.214 --> 00:17:28.755
- I mean, is that something that,
459
00:17:29.775 --> 00:17:31.775
- you know, that you see as as as
460
00:17:31.775 --> 00:17:33.315
- a possibility in the future?
461
00:17:33.934 --> 00:17:35.795
- Yeah. I think it's definitely a possibility.
462
00:17:36.815 --> 00:17:39.134
- There are people that are working on on
463
00:17:39.134 --> 00:17:40.994
- these types of of things, and,
464
00:17:42.130 --> 00:17:44.470
- yeah, it's it's it's it's definitely a possibility
465
00:17:44.930 --> 00:17:47.009
- that we would use different types of quantum
466
00:17:47.009 --> 00:17:47.490
- computers,
467
00:17:47.809 --> 00:17:48.690
- that would be,
468
00:17:49.329 --> 00:17:51.410
- each would be better at at at its
469
00:17:51.410 --> 00:17:52.150
- own tasks,
470
00:17:52.529 --> 00:17:54.690
- and we would need to integrate all of
471
00:17:54.690 --> 00:17:56.150
- them under a single umbrella.
472
00:17:56.535 --> 00:17:57.035
- Right?
473
00:17:57.494 --> 00:18:00.455
- Where users eventually they might not even know
474
00:18:00.455 --> 00:18:02.215
- that they are working on this or that
475
00:18:02.215 --> 00:18:03.035
- quantum computer,
476
00:18:04.134 --> 00:18:06.154
- and we need to build this entire stack
477
00:18:06.535 --> 00:18:08.695
- that can can really integrate to these different
478
00:18:08.695 --> 00:18:09.994
- types of of hardware.
479
00:18:11.200 --> 00:18:12.019
- I see.
480
00:18:12.480 --> 00:18:15.039
- And and I wanted to talk, a bit
481
00:18:15.039 --> 00:18:18.660
- about, you know, sort of specific pro products,
482
00:18:19.759 --> 00:18:22.720
- from quantum machines. Early earlier this year, you
483
00:18:22.720 --> 00:18:23.220
- launched,
484
00:18:23.920 --> 00:18:24.660
- the open
485
00:18:24.960 --> 00:18:25.460
- acceleration
486
00:18:26.160 --> 00:18:26.660
- stack.
487
00:18:27.015 --> 00:18:29.654
- Can you describe this product and and talk
488
00:18:29.654 --> 00:18:31.275
- about how it's being used?
489
00:18:32.695 --> 00:18:35.575
- Yeah. So the open acceleration stack allows us
490
00:18:35.575 --> 00:18:36.714
- to connect
491
00:18:37.015 --> 00:18:40.075
- our control system, which essentially means to connect
492
00:18:40.880 --> 00:18:43.440
- all the QPUs, the quantum processors that we
493
00:18:43.440 --> 00:18:44.579
- are connected with,
494
00:18:45.119 --> 00:18:46.819
- to classical accelerators
495
00:18:47.599 --> 00:18:48.980
- classical compute accelerators
496
00:18:49.599 --> 00:18:51.700
- with very, very low latency links,
497
00:18:52.559 --> 00:18:53.539
- with high bandwidth.
498
00:18:54.345 --> 00:18:54.845
- So,
499
00:18:55.944 --> 00:18:57.644
- whether it's an NVIDIA GPU
500
00:18:58.184 --> 00:18:59.644
- or an AMD FPGA
501
00:19:00.345 --> 00:19:02.684
- or an Intel CPU, etcetera, etcetera,
502
00:19:02.984 --> 00:19:04.285
- we can now connect
503
00:19:04.744 --> 00:19:06.605
- our control system to that accelerator
504
00:19:07.399 --> 00:19:09.099
- with microsecond latency,
505
00:19:10.119 --> 00:19:10.619
- links
506
00:19:11.000 --> 00:19:12.919
- so that we can communicate. So then the
507
00:19:12.919 --> 00:19:13.419
- QPU
508
00:19:13.880 --> 00:19:16.140
- and the GPU, for instance, can communicate
509
00:19:16.599 --> 00:19:19.000
- over this very low latency link, a high
510
00:19:19.000 --> 00:19:20.859
- bandwidth, and share information.
511
00:19:21.345 --> 00:19:23.424
- And as I mentioned, this is very important,
512
00:19:23.744 --> 00:19:26.224
- because we need to do this interleaving between
513
00:19:26.224 --> 00:19:27.924
- quantum and classical processing.
514
00:19:28.544 --> 00:19:30.785
- So now we can virtually connect to any
515
00:19:30.785 --> 00:19:31.765
- classical accelerator,
516
00:19:33.345 --> 00:19:36.390
- from the QPU side, And, and this gives
517
00:19:36.390 --> 00:19:38.069
- a lot of value to our customers. They're
518
00:19:38.069 --> 00:19:38.549
- building,
519
00:19:38.869 --> 00:19:42.569
- calibration workflows, error correction workflows, and eventually application
520
00:19:42.630 --> 00:19:43.130
- workflows
521
00:19:43.669 --> 00:19:45.289
- using these these hybrid systems.
522
00:19:46.549 --> 00:19:48.744
- I see. And and so is that is
523
00:19:48.744 --> 00:19:50.525
- that being used using,
524
00:19:51.705 --> 00:19:54.445
- a number of different types of of quantum
525
00:19:54.505 --> 00:19:55.005
- processor?
526
00:19:56.105 --> 00:19:58.045
- Yes. Definitely. We have customers,
527
00:19:59.224 --> 00:20:02.220
- with spin qubits, for example. Dirac is a
528
00:20:03.740 --> 00:20:05.200
- company from Australia
529
00:20:05.579 --> 00:20:07.819
- that is heavily using this kind of, low
530
00:20:07.819 --> 00:20:08.880
- latency links
531
00:20:09.339 --> 00:20:09.740
- to,
532
00:20:10.220 --> 00:20:13.759
- stabilize their quantum processors, calibrate their quantum processors,
533
00:20:14.220 --> 00:20:16.640
- and, again, later on do the error correction,
534
00:20:17.419 --> 00:20:20.674
- using this link between our control system and
535
00:20:20.674 --> 00:20:21.075
- the,
536
00:20:21.714 --> 00:20:22.855
- the the accelerators.
537
00:20:24.115 --> 00:20:26.454
- In the same way, we have superconducting qubits
538
00:20:26.514 --> 00:20:27.014
- and
539
00:20:27.315 --> 00:20:28.694
- neutral atom companies,
540
00:20:29.075 --> 00:20:30.210
- doing the same thing.
541
00:20:30.529 --> 00:20:31.809
- So so yes.
542
00:20:32.529 --> 00:20:35.429
- Again, each each Qubitab has slightly different requirements,
543
00:20:35.569 --> 00:20:37.970
- but, I think this part of the stack
544
00:20:37.970 --> 00:20:39.349
- is actually very general
545
00:20:39.890 --> 00:20:42.470
- to to a lot to to everyone.
546
00:20:44.234 --> 00:20:44.634
- And,
547
00:20:45.035 --> 00:20:47.914
- another announcement that you made, earlier this year
548
00:20:47.914 --> 00:20:49.214
- is that your OPX
549
00:20:49.835 --> 00:20:50.654
- 1,000
550
00:20:50.795 --> 00:20:52.654
- platform was used to control
551
00:20:53.194 --> 00:20:55.615
- a Rigetti nine qubit processor,
552
00:20:56.419 --> 00:20:59.079
- achieving a 99.5
553
00:20:59.460 --> 00:20:59.960
- median
554
00:21:00.419 --> 00:21:02.279
- two qubit gate fidelity.
555
00:21:03.140 --> 00:21:04.919
- Can can you explain the significance
556
00:21:05.460 --> 00:21:06.179
- of this,
557
00:21:06.500 --> 00:21:07.640
- specific achievement?
558
00:21:09.974 --> 00:21:13.195
- Yeah. So this is, a great example,
559
00:21:14.615 --> 00:21:15.015
- that,
560
00:21:15.575 --> 00:21:16.075
- shows
561
00:21:16.855 --> 00:21:19.275
- something which is quite new in the field.
562
00:21:19.654 --> 00:21:21.255
- You know, there's a lot of,
563
00:21:21.900 --> 00:21:23.500
- I mean, so so far in the field
564
00:21:23.500 --> 00:21:25.440
- that there is there's been a lot of,
565
00:21:25.660 --> 00:21:29.099
- vertical integration. Right? So which is very natural
566
00:21:29.099 --> 00:21:31.039
- at the beginning of of of a new
567
00:21:31.180 --> 00:21:32.960
- technological field. You have,
568
00:21:33.580 --> 00:21:34.960
- you know, companies doing
569
00:21:35.420 --> 00:21:35.920
- everything,
570
00:21:36.460 --> 00:21:39.244
- the entire stack and doing vertical integration,
571
00:21:40.424 --> 00:21:42.285
- before you start seeing more,
572
00:21:43.305 --> 00:21:44.205
- sort of horizontal
573
00:21:44.904 --> 00:21:45.404
- expertise
574
00:21:46.265 --> 00:21:48.205
- being developed, in the stack.
575
00:21:49.465 --> 00:21:49.965
- So,
576
00:21:50.809 --> 00:21:52.170
- it used to be, you know, that the
577
00:21:52.170 --> 00:21:54.349
- company does both the quantum processor
578
00:21:54.809 --> 00:21:55.710
- and the calibration
579
00:21:56.009 --> 00:21:58.890
- stack or maybe even the control system themselves
580
00:21:58.890 --> 00:22:01.130
- and then the calibration stack and then bringing
581
00:22:01.130 --> 00:22:02.890
- up the the qubits and seeing that that
582
00:22:02.890 --> 00:22:05.369
- that they can achieve high fidelity operations on
583
00:22:05.369 --> 00:22:06.029
- their quantum,
584
00:22:06.994 --> 00:22:08.214
- on their quantum processors.
585
00:22:09.875 --> 00:22:12.534
- And, this is an example where we we
586
00:22:12.595 --> 00:22:13.494
- we came in,
587
00:22:14.194 --> 00:22:15.095
- to a Rigetti,
588
00:22:15.554 --> 00:22:16.054
- QPU.
589
00:22:16.595 --> 00:22:19.315
- Rigetti is now providing their QPUs. They also
590
00:22:19.315 --> 00:22:20.994
- provide a full stack on the computer, but
591
00:22:20.994 --> 00:22:22.789
- they can also provide you with only the
592
00:22:22.789 --> 00:22:24.730
- QPU, only the quantum processor.
593
00:22:25.430 --> 00:22:27.049
- And so we show that
594
00:22:27.589 --> 00:22:29.830
- using our control system and actually also our
595
00:22:29.830 --> 00:22:31.130
- calibration stack,
596
00:22:31.430 --> 00:22:34.470
- you can achieve this high fidelity operations on
597
00:22:34.470 --> 00:22:37.164
- the Rigetti QPU. So showing that we can
598
00:22:37.164 --> 00:22:38.705
- do that by, taking,
599
00:22:39.164 --> 00:22:41.965
- components from, you know, different vendors, the QPU
600
00:22:41.965 --> 00:22:44.525
- from Rigetti, the control from QM, and the
601
00:22:44.525 --> 00:22:46.225
- the the calibrations from QM.
602
00:22:47.164 --> 00:22:49.085
- We can really achieve state of the art
603
00:22:49.085 --> 00:22:49.585
- fidelities,
604
00:22:50.660 --> 00:22:53.059
- that shows that the the the now we're
605
00:22:53.059 --> 00:22:55.539
- really getting into this place where things are
606
00:22:55.539 --> 00:22:57.240
- starting to become more standardized.
607
00:22:57.859 --> 00:22:59.559
- You can really build this horizontal,
608
00:23:00.179 --> 00:23:00.679
- expertise
609
00:23:01.059 --> 00:23:03.875
- and, and and and and build a a
610
00:23:03.875 --> 00:23:04.914
- more, rich,
611
00:23:05.315 --> 00:23:05.815
- ecosystem
612
00:23:06.434 --> 00:23:07.174
- of players
613
00:23:07.795 --> 00:23:09.894
- focusing on different parts of the stack.
614
00:23:11.154 --> 00:23:13.015
- You you've mentioned two collaborative
615
00:23:13.315 --> 00:23:13.815
- projects,
616
00:23:14.275 --> 00:23:16.615
- that Quantum Machines is involved with.
617
00:23:17.210 --> 00:23:18.910
- Can you can you talk about,
618
00:23:19.529 --> 00:23:21.390
- what what else you're doing worldwide?
619
00:23:25.049 --> 00:23:27.630
- Well, we're doing a lot of different things.
620
00:23:27.690 --> 00:23:28.190
- We
621
00:23:30.214 --> 00:23:32.695
- have many, many different partnerships. We run here
622
00:23:32.695 --> 00:23:35.414
- the Israeli Quantum Computing Center in, in Tel
623
00:23:35.414 --> 00:23:38.695
- Aviv, which, by itself is hosting different chips
624
00:23:38.695 --> 00:23:40.795
- from around the world. We have a collaboration,
625
00:23:40.855 --> 00:23:42.154
- for instance, with Colab.
626
00:23:42.470 --> 00:23:43.289
- It's a startup,
627
00:23:43.809 --> 00:23:45.849
- from The US, that,
628
00:23:46.390 --> 00:23:48.070
- what like, the the the one of the
629
00:23:48.070 --> 00:23:50.410
- founders there is John Martinez, who's,
630
00:23:50.950 --> 00:23:52.890
- also one of the Nobel Prize laureates,
631
00:23:53.430 --> 00:23:54.089
- in physics,
632
00:23:54.710 --> 00:23:57.295
- from, you know, '25. And I
633
00:23:57.695 --> 00:23:59.075
- we've been working very,
634
00:23:59.455 --> 00:24:01.154
- closely with John to,
635
00:24:02.015 --> 00:24:04.575
- we we're hosting the the collab chips in
636
00:24:04.575 --> 00:24:07.955
- our our Israeli Quantum Computing Center here, demonstrating
637
00:24:08.095 --> 00:24:10.195
- different types of hybrid quantum classical
638
00:24:14.779 --> 00:24:16.640
- applications on on that chip.
639
00:24:17.500 --> 00:24:18.000
- We,
640
00:24:18.779 --> 00:24:21.039
- have, of course, our collaboration with with NVIDIA
641
00:24:21.179 --> 00:24:23.419
- where we, for the first time, created these
642
00:24:23.419 --> 00:24:25.845
- low latency links between the the control system
643
00:24:25.845 --> 00:24:27.384
- and the NVIDIA GPUs.
644
00:24:28.164 --> 00:24:31.045
- We're working now with AMD, integrating with their,
645
00:24:31.365 --> 00:24:31.865
- FPGAs.
646
00:24:32.565 --> 00:24:33.684
- We're working with,
647
00:24:34.085 --> 00:24:35.765
- with many other partners on this,
648
00:24:36.964 --> 00:24:38.664
- quantum classical hybridization,
649
00:24:39.445 --> 00:24:40.585
- part of the stack.
650
00:24:41.700 --> 00:24:42.200
- So,
651
00:24:43.220 --> 00:24:45.319
- yeah, it's kind of across the board.
652
00:24:46.579 --> 00:24:49.799
- Quantum Machines has recently acquired the Dutch company
653
00:24:50.019 --> 00:24:50.519
- QHarbor,
654
00:24:51.140 --> 00:24:54.039
- which has developed a data storage and visualization
655
00:24:55.125 --> 00:24:57.065
- software platform for scientists.
656
00:24:58.005 --> 00:24:59.605
- On the face of it, this seems like
657
00:24:59.605 --> 00:25:00.265
- a strange
658
00:25:00.644 --> 00:25:04.085
- acquisition for a quantum computing company. Why why
659
00:25:04.085 --> 00:25:05.545
- did you buy QHarbor?
660
00:25:07.880 --> 00:25:08.859
- So QHarbor,
661
00:25:09.240 --> 00:25:11.559
- we we we bought them for a for
662
00:25:11.559 --> 00:25:12.380
- a few reasons.
663
00:25:13.159 --> 00:25:14.039
- The the first,
664
00:25:14.519 --> 00:25:16.679
- is really, as we always say, is just
665
00:25:16.679 --> 00:25:17.339
- the team.
666
00:25:18.119 --> 00:25:20.855
- QHarbor, I mean, it's a small team, but,
667
00:25:21.315 --> 00:25:22.355
- very, very strong,
668
00:25:23.154 --> 00:25:25.714
- team on the technical level, and we just
669
00:25:25.714 --> 00:25:27.954
- clicked with the people, you know, and, felt
670
00:25:27.954 --> 00:25:29.795
- that there is a, you know, we want
671
00:25:29.795 --> 00:25:30.934
- these guys on board.
672
00:25:31.394 --> 00:25:33.234
- So that that's always, I think, one of
673
00:25:33.234 --> 00:25:34.535
- the key key aspects,
674
00:25:35.559 --> 00:25:38.059
- especially when the team is small. Right?
675
00:25:38.440 --> 00:25:40.059
- It's it's really about bringing,
676
00:25:40.440 --> 00:25:43.240
- you know, fantastic people to the team. Besides
677
00:25:43.240 --> 00:25:45.720
- that, we keep building more and more software
678
00:25:45.720 --> 00:25:48.359
- tools for our customers. And the software that
679
00:25:48.359 --> 00:25:49.420
- QHarbor built
680
00:25:50.345 --> 00:25:52.105
- seemed to us like, it can provide a
681
00:25:52.105 --> 00:25:54.365
- lot of, extra value to our customers.
682
00:25:55.225 --> 00:25:57.644
- I mean, how they how they store, manage,
683
00:25:58.184 --> 00:26:00.365
- standardized data and data
684
00:26:00.664 --> 00:26:01.164
- viewing,
685
00:26:01.705 --> 00:26:02.765
- and data analysis.
686
00:26:03.839 --> 00:26:05.619
- As I mentioned, quantum computing,
687
00:26:05.920 --> 00:26:06.960
- there's a lot of,
688
00:26:07.680 --> 00:26:10.960
- need to do data analysis and, and, data
689
00:26:10.960 --> 00:26:11.460
- management.
690
00:26:12.079 --> 00:26:14.720
- And so this tool was just, you know,
691
00:26:14.720 --> 00:26:17.200
- something that we thought could integrate really nicely
692
00:26:17.200 --> 00:26:19.744
- with our software platform and provide more value
693
00:26:19.744 --> 00:26:20.565
- to our customers.
694
00:26:21.424 --> 00:26:23.744
- Other than that, we are also, growing in
695
00:26:23.744 --> 00:26:27.105
- Delft specifically. We, just opened an office in
696
00:26:27.105 --> 00:26:29.825
- Delft, and we are growing there. So this
697
00:26:29.825 --> 00:26:31.924
- sort of aligned also with our our
698
00:26:32.299 --> 00:26:33.359
- our growth ambitious,
699
00:26:33.980 --> 00:26:35.919
- ambitious in Europe in general, but,
700
00:26:36.700 --> 00:26:37.099
- also,
701
00:26:37.500 --> 00:26:40.000
- specifically in in in the in The Netherlands.
702
00:26:40.779 --> 00:26:43.259
- So that was kind of everything clicked together,
703
00:26:43.259 --> 00:26:43.759
- and
704
00:26:44.315 --> 00:26:45.775
- and we we made the acquisition.
705
00:26:47.115 --> 00:26:47.615
- And,
706
00:26:47.914 --> 00:26:49.595
- Yonatan, I was hoping that we could sort
707
00:26:49.595 --> 00:26:51.674
- of finish off this interview by talking a
708
00:26:51.674 --> 00:26:52.575
- bit about
709
00:26:52.875 --> 00:26:53.375
- careers.
710
00:26:54.554 --> 00:26:56.315
- And and first, I'd like to ask you
711
00:26:56.315 --> 00:26:58.015
- about, you know, your own
712
00:26:58.369 --> 00:26:58.869
- career.
713
00:26:59.170 --> 00:26:59.670
- You,
714
00:27:00.289 --> 00:27:02.869
- you did a PhD, and I believe you
715
00:27:03.170 --> 00:27:04.630
- looked at Mariana
716
00:27:05.410 --> 00:27:06.950
- cubits. Is that right?
717
00:27:07.330 --> 00:27:09.029
- Can can you talk a bit about,
718
00:27:09.890 --> 00:27:12.384
- your, you know, your your reese the research
719
00:27:12.384 --> 00:27:14.465
- that you did? You know, why why you
720
00:27:14.465 --> 00:27:16.244
- did it? Why it's important
721
00:27:16.625 --> 00:27:17.025
- for,
722
00:27:17.904 --> 00:27:20.325
- for the development of quantum computers?
723
00:27:22.144 --> 00:27:24.005
- Sure. Yeah. So my research,
724
00:27:24.305 --> 00:27:26.619
- was, I need at least some of my
725
00:27:26.619 --> 00:27:28.559
- my research work was, around
726
00:27:29.339 --> 00:27:30.480
- searching for topological
727
00:27:30.779 --> 00:27:33.119
- quantum states of matter and, specifically,
728
00:27:33.579 --> 00:27:35.119
- my run of fermions in
729
00:27:35.900 --> 00:27:36.400
- semiconducting,
730
00:27:36.940 --> 00:27:37.440
- superconducting
731
00:27:37.740 --> 00:27:38.240
- hybrid
732
00:27:38.619 --> 00:27:39.119
- devices.
733
00:27:40.384 --> 00:27:42.704
- This was sort of when this business of
734
00:27:42.704 --> 00:27:43.204
- superconducting,
735
00:27:43.505 --> 00:27:44.005
- semiconducting,
736
00:27:45.744 --> 00:27:46.724
- hybrid devices
737
00:27:47.265 --> 00:27:49.265
- started to really gain a lot of attention
738
00:27:49.265 --> 00:27:50.644
- by the physics community,
739
00:27:51.744 --> 00:27:54.164
- because of a a couple of, important,
740
00:27:54.859 --> 00:27:56.859
- you know, works that found the signature of
741
00:27:56.859 --> 00:27:58.400
- these Majorana fermions. And,
742
00:27:59.980 --> 00:28:00.480
- and,
743
00:28:02.220 --> 00:28:04.700
- I'm now nowadays, we're seeing more and more
744
00:28:04.700 --> 00:28:08.000
- results from Microsoft around this, this this this,
745
00:28:08.779 --> 00:28:09.900
- this, this,
746
00:28:10.384 --> 00:28:12.164
- you know, physical phenomena that
747
00:28:12.465 --> 00:28:14.305
- are promising in the sense that maybe one
748
00:28:14.305 --> 00:28:17.125
- day we could use that to create, topological
749
00:28:17.424 --> 00:28:17.924
- quantum,
750
00:28:18.225 --> 00:28:18.725
- computing,
751
00:28:19.184 --> 00:28:21.585
- which means that basically you, the so those
752
00:28:21.585 --> 00:28:25.125
- those topological quantum states and specifically Majorana fermions
753
00:28:25.184 --> 00:28:26.244
- can form qubits
754
00:28:27.080 --> 00:28:30.440
- that, are protected from, you know, environments or
755
00:28:30.440 --> 00:28:31.580
- protected from noise,
756
00:28:32.359 --> 00:28:33.180
- in the environment,
757
00:28:33.720 --> 00:28:35.880
- in a in an inherent way in the
758
00:28:35.880 --> 00:28:37.720
- in the hardware. So it's just almost as
759
00:28:37.720 --> 00:28:40.295
- if your your hardware is doing the quantum
760
00:28:40.295 --> 00:28:42.475
- error correction for you. Right?
761
00:28:44.055 --> 00:28:46.215
- And it's very hard to make those,
762
00:28:46.695 --> 00:28:47.195
- qubits,
763
00:28:47.734 --> 00:28:49.654
- and, I think that there is still a
764
00:28:49.654 --> 00:28:50.695
- long way to go,
765
00:28:51.335 --> 00:28:53.515
- which is still a long research,
766
00:28:55.220 --> 00:28:57.159
- way to go to really
767
00:28:58.019 --> 00:28:58.519
- build
768
00:28:59.460 --> 00:29:01.399
- robust qubits out of these,
769
00:29:01.700 --> 00:29:03.240
- these type of of of devices.
770
00:29:03.619 --> 00:29:05.220
- But I think there's a lot of interesting
771
00:29:05.220 --> 00:29:07.380
- progress that's happening, and I hope that one
772
00:29:07.380 --> 00:29:09.755
- day we we will see this as a
773
00:29:09.755 --> 00:29:11.835
- as a platform that really competes to to
774
00:29:11.835 --> 00:29:14.335
- become a, you know, a a quantum computer.
775
00:29:15.434 --> 00:29:16.815
- So to me, it's very exciting.
776
00:29:17.595 --> 00:29:19.774
- In the doing my PhD, it was
777
00:29:20.394 --> 00:29:22.875
- an extremely interesting topic on the sort of
778
00:29:22.875 --> 00:29:23.375
- fundamental
779
00:29:24.140 --> 00:29:25.519
- physics side of things.
780
00:29:26.539 --> 00:29:27.919
- How can you create these
781
00:29:28.220 --> 00:29:28.720
- exotic,
782
00:29:30.700 --> 00:29:32.559
- systems that are
783
00:29:32.859 --> 00:29:35.519
- by the nature, by the mathematics of the
784
00:29:35.755 --> 00:29:36.954
- the the Hamiltonian, the,
785
00:29:37.355 --> 00:29:39.194
- the the way that we sort of engineer
786
00:29:39.194 --> 00:29:39.855
- the Hamiltonian,
787
00:29:40.714 --> 00:29:43.994
- from different materials, from magnetic fields, we can
788
00:29:43.994 --> 00:29:47.434
- actually create these very exotic phenomena that,
789
00:29:48.289 --> 00:29:48.789
- that,
790
00:29:49.329 --> 00:29:51.190
- that that that that,
791
00:29:51.809 --> 00:29:52.309
- that,
792
00:29:52.690 --> 00:29:53.750
- that, you know,
793
00:29:55.169 --> 00:29:57.169
- that that doesn't just appear like that in
794
00:29:57.169 --> 00:29:59.984
- nature, if you don't engineer a system. So
795
00:30:00.065 --> 00:30:01.845
- I think that's a fascinating topic,
796
00:30:02.464 --> 00:30:02.964
- and,
797
00:30:03.904 --> 00:30:06.065
- yeah, especially for research. And I hope that
798
00:30:06.065 --> 00:30:08.065
- one day it will also become a become
799
00:30:08.065 --> 00:30:09.845
- a a qubit platform,
800
00:30:10.304 --> 00:30:11.924
- that can can really scale.
801
00:30:12.545 --> 00:30:14.464
- And and were you working in the lab,
802
00:30:14.464 --> 00:30:16.799
- or are you were you a theorist, or
803
00:30:16.799 --> 00:30:18.099
- was it a bit of both?
804
00:30:18.640 --> 00:30:21.359
- No. I was an experimental physicist. So I
805
00:30:21.359 --> 00:30:22.339
- did a lot of,
806
00:30:23.039 --> 00:30:25.059
- fabrication work in the clean room,
807
00:30:25.519 --> 00:30:26.740
- fabricating devices,
808
00:30:28.174 --> 00:30:31.154
- both indium arsenide nanowires coupled to superconductors.
809
00:30:31.535 --> 00:30:33.234
- I worked also in gallium arsenide,
810
00:30:34.654 --> 00:30:37.555
- quantum all effect, fractional quantum all effect. So
811
00:30:37.695 --> 00:30:39.474
- I was, you know, working
812
00:30:40.259 --> 00:30:41.460
- a lot on,
813
00:30:41.940 --> 00:30:44.259
- making new type of devices and then measuring
814
00:30:44.259 --> 00:30:45.720
- them in dilution refrigerators,
815
00:30:46.339 --> 00:30:48.200
- doing, all kinds of different,
816
00:30:51.220 --> 00:30:53.080
- measurements to explore different phenomena.
817
00:30:54.455 --> 00:30:57.115
- And so, yes, I was an experimental physicist,
818
00:30:57.734 --> 00:30:59.035
- looking for this exotic,
819
00:30:59.575 --> 00:31:00.075
- phenomena
820
00:31:00.455 --> 00:31:03.434
- in topological quantum, quantum states of matter.
821
00:31:04.375 --> 00:31:05.835
- And some of our listeners,
822
00:31:06.942 --> 00:31:09.609
- I I can guarantee you will be interested
823
00:31:09.670 --> 00:31:10.410
- in careers
824
00:31:10.950 --> 00:31:12.410
- in quantum computing.
825
00:31:12.950 --> 00:31:15.529
- What advice would you give to a physics
826
00:31:15.589 --> 00:31:16.089
- graduate
827
00:31:16.390 --> 00:31:16.789
- who,
828
00:31:17.670 --> 00:31:18.730
- would like a career?
829
00:31:19.845 --> 00:31:23.204
- Maybe at quantum machines or at another quantum
830
00:31:23.204 --> 00:31:24.184
- computing company?
831
00:31:24.644 --> 00:31:28.005
- I think that's doing some serious hands on
832
00:31:28.005 --> 00:31:28.505
- work,
833
00:31:29.684 --> 00:31:31.704
- on what really interests you,
834
00:31:32.724 --> 00:31:34.184
- is the most important thing.
835
00:31:34.919 --> 00:31:37.019
- You know, really going down to the details
836
00:31:37.079 --> 00:31:39.559
- of, of the of of of the thing
837
00:31:39.559 --> 00:31:42.119
- of interest in doing you know, really getting
838
00:31:42.119 --> 00:31:44.700
- sort of your hands dirty with with stuff,
839
00:31:45.559 --> 00:31:47.399
- and going to the really nuts and bolts.
840
00:31:47.399 --> 00:31:50.045
- That's the best, first thing to to start
841
00:31:50.045 --> 00:31:50.545
- from.
842
00:31:52.205 --> 00:31:52.605
- And,
843
00:31:53.404 --> 00:31:55.744
- and besides that, you know, it's the usual,
844
00:31:57.244 --> 00:31:59.664
- do do do something that really,
845
00:32:00.205 --> 00:32:01.049
- interests you,
846
00:32:01.850 --> 00:32:03.369
- Because you have to succeed in this field,
847
00:32:03.369 --> 00:32:05.150
- you have to also work very hard.
848
00:32:05.610 --> 00:32:08.250
- And, the field is moving very, very fast
849
00:32:08.250 --> 00:32:08.910
- these days.
850
00:32:09.289 --> 00:32:10.830
- There's a lot that's happening.
851
00:32:11.529 --> 00:32:13.130
- You wanna succeed in this field. You have
852
00:32:13.130 --> 00:32:15.575
- to work very hard. So it better be,
853
00:32:15.815 --> 00:32:18.315
- in something that, that you find very interesting
854
00:32:18.375 --> 00:32:21.355
- and, and and it gives you, the motivation
855
00:32:21.494 --> 00:32:21.994
- to
856
00:32:22.455 --> 00:32:22.955
- to
857
00:32:23.575 --> 00:32:25.335
- to wake up in in every morning and
858
00:32:25.335 --> 00:32:27.275
- do and and get your hands dirty.
859
00:32:28.700 --> 00:32:30.779
- I would also say that that that,
860
00:32:31.259 --> 00:32:32.720
- I think, you
861
00:32:33.019 --> 00:32:34.400
- know, you know, it's,
862
00:32:35.180 --> 00:32:36.240
- as things progress,
863
00:32:37.740 --> 00:32:39.819
- you have to be a little sort of
864
00:32:39.819 --> 00:32:40.319
- multidisciplinary
865
00:32:40.859 --> 00:32:41.714
- in the sense that,
866
00:32:42.275 --> 00:32:44.194
- you you okay. So if you're a physicist,
867
00:32:44.194 --> 00:32:45.875
- of course, you're coming from the physics, but
868
00:32:45.875 --> 00:32:47.734
- it's critical that you,
869
00:32:48.914 --> 00:32:49.414
- know,
870
00:32:49.875 --> 00:32:51.494
- how to program. You understand,
871
00:32:52.595 --> 00:32:53.794
- software. You understand,
872
00:32:54.515 --> 00:32:56.134
- technology stacks in general,
873
00:32:56.599 --> 00:32:58.700
- how the hardware is built, how the software
874
00:32:58.759 --> 00:33:00.460
- is built, how things integrate,
875
00:33:01.240 --> 00:33:03.420
- understand a little bit of the engineering language,
876
00:33:03.640 --> 00:33:05.799
- both on in hardware and software. Mhmm. And
877
00:33:05.799 --> 00:33:08.039
- I think it's it's becoming very, very important
878
00:33:08.039 --> 00:33:08.539
- that
879
00:33:08.904 --> 00:33:10.365
- that people sort of understand
880
00:33:10.744 --> 00:33:11.244
- systems,
881
00:33:12.664 --> 00:33:15.384
- especially the physicists, because I think that's one
882
00:33:15.384 --> 00:33:17.545
- of the I think that's the the that's
883
00:33:17.545 --> 00:33:20.025
- one of the strengths of physicists is that
884
00:33:20.025 --> 00:33:20.924
- they can understand
885
00:33:21.549 --> 00:33:24.130
- systems sort of from the, you know, physics
886
00:33:24.190 --> 00:33:25.730
- all the way up to to software.
887
00:33:26.430 --> 00:33:28.190
- So, you know, if you're a physicist, you're
888
00:33:28.190 --> 00:33:30.430
- probably not going to be the the the
889
00:33:30.430 --> 00:33:31.250
- best programmer,
890
00:33:31.789 --> 00:33:34.269
- in the world already here because you haven't
891
00:33:34.269 --> 00:33:37.115
- done that, But you can understand how everything
892
00:33:37.115 --> 00:33:37.615
- integrates,
893
00:33:37.994 --> 00:33:40.315
- pretty well. So I think that that's something
894
00:33:40.315 --> 00:33:42.335
- that physicists should I should look at.
895
00:33:43.755 --> 00:33:46.394
- Well, that's great, Yonatan. Thanks so much for
896
00:33:46.394 --> 00:33:47.454
- coming on the podcast.
897
00:33:48.714 --> 00:33:50.654
- Thank you so much, Hamish. Thank you.
898
00:33:57.789 --> 00:33:59.009
- That was Yonatan
899
00:33:59.309 --> 00:34:01.330
- Cohen of Quantum Machines.
900
00:34:01.950 --> 00:34:04.289
- Thanks to Yonatan for his insights
901
00:34:04.615 --> 00:34:06.875
- into the challenges and opportunities
902
00:34:07.654 --> 00:34:08.474
- in interfacing
903
00:34:08.934 --> 00:34:11.835
- quantum and classical computing systems.
904
00:34:12.695 --> 00:34:14.855
- I'm afraid that's all the time we have
905
00:34:14.855 --> 00:34:16.155
- for this week's podcast.
906
00:34:16.710 --> 00:34:19.989
- I'm Hamish Johnston, and our producer is Fred
907
00:34:19.989 --> 00:34:20.489
- Ailes.
908
00:34:21.109 --> 00:34:23.609
- The music in Physics World's podcasts
909
00:34:24.150 --> 00:34:24.889
- is called
910
00:34:25.190 --> 00:34:26.650
- one three seven,
911
00:34:26.949 --> 00:34:29.670
- and it was composed and performed by the
912
00:34:29.670 --> 00:34:30.170
- physicist
913
00:34:30.869 --> 00:34:31.690
- Philip Moriarty.
914
00:34:32.704 --> 00:34:34.565
- We'll be back again next week.