1 00:00:00,000 --> 00:00:20,617 *36C3 preroll music* 2 00:00:20,617 --> 00:00:24,550 Herald: Welcome to the world of quantum computing. Well, most of you is gonna just 3 00:00:24,550 --> 00:00:28,750 gonna say, ah, that stuff is just for cracking RSA keys. But there is actually a 4 00:00:28,750 --> 00:00:38,090 little bit more to that. It's interesting stuff. And our next speaker, Yann Allain 5 00:00:38,090 --> 00:00:42,839 is going to introduce this world of quantum computing to us. And he's gonna 6 00:00:42,839 --> 00:00:49,070 show us a couple of application scenarios and how to build your or our own quantum 7 00:00:49,070 --> 00:00:52,753 computer. Yann. 8 00:00:52,753 --> 00:00:59,784 *Applause* 9 00:00:59,784 --> 00:01:04,210 Yann: Salut, hello everybody, Guten Tag, alle - this is the only word I know in 10 00:01:04,210 --> 00:01:13,110 Deutsch. We will begin this session by try to convince you that building a quantum 11 00:01:13,110 --> 00:01:20,480 computer at home is still possible. This is the agenda. We are in a INFOSEC 12 00:01:20,480 --> 00:01:24,730 security conference. Why bother with quantum computing when we work at 13 00:01:24,730 --> 00:01:32,890 cybersecurity? We will try to explain to you in a simply manner how to our quantum 14 00:01:32,890 --> 00:01:41,891 computer works. We will explain to you, how we build our own quantum computer. And 15 00:01:41,891 --> 00:01:47,200 of course, because we are CCC, we need to know how to hack into a quantum computer. 16 00:01:47,200 --> 00:01:53,910 So, let me introduce myself a little bit. So, I'm Yann Allain, I am French. I'm used 17 00:01:53,910 --> 00:01:59,950 to share my project with some security conferences: Hack In the Box, Blackhat. 18 00:01:59,950 --> 00:02:03,920 I was a speaker and trainer in this type of conference. It's the first time for me in 19 00:02:03,920 --> 00:02:09,420 CCC, so it's very cool. I'm mostly an entrepreneur, an engineer. And of course, 20 00:02:09,420 --> 00:02:15,299 my new company, NextGenQ, which stands for Next Generation of Quantum Computer is a 21 00:02:15,299 --> 00:02:23,470 quantum company. I work in the INFOSEC security since 25 years now. So I'm a 22 00:02:23,470 --> 00:02:30,910 veteran of these domain. I fight again "I love you" very seasoned slammer worm. If 23 00:02:30,910 --> 00:02:36,110 you remember those worm and my past activities are related to software and 24 00:02:36,110 --> 00:02:43,450 hardware security. So why bother with quantum computing when we work in 25 00:02:43,450 --> 00:02:53,010 cybersecurity? If you want to make some difficult calculation on areas like sci., for 26 00:02:53,010 --> 00:03:02,599 example, to factor... a large number on a classical computer, it will take 10 to the 27 00:03:02,599 --> 00:03:14,480 power of 34 steps. It's a big number and it will take on a normal computer 300 28 00:03:14,480 --> 00:03:23,036 trillion of years. It's a long, long time. It's why we say that RSA is secure. On a 29 00:03:23,036 --> 00:03:29,553 quantum computer we have a specific algorithm called Shor algorithm. It take 30 00:03:29,553 --> 00:03:35,960 only 10 to the power of 7 steps. It's a smaller number and it takes only 10 31 00:03:35,960 --> 00:03:45,489 seconds. However, you could think that the statement is a little bit overhyped. Yes 32 00:03:45,489 --> 00:03:52,970 and no. No, because Shor algorithm is able to break RSA. This is the goal of this 33 00:03:52,970 --> 00:04:01,029 algorithm in the human time. However, at the moment we speak - to break a big 34 00:04:01,029 --> 00:04:07,469 number with this algorithm you need to have a much bigger quantum computer that 35 00:04:07,469 --> 00:04:18,169 exists nowadays. For example, you need 4 thousands ideal qubits quantum computer. 36 00:04:18,169 --> 00:04:26,840 It doesn't exist for the moment. However, quantum computing could be used also for 37 00:04:26,840 --> 00:04:35,020 some benefits for our domain of INFOSEC cybersecurity. There is many advantage on 38 00:04:35,020 --> 00:04:40,850 the corner. You can use a quantum computer or quantum technology to generate true 39 00:04:40,850 --> 00:04:46,870 random number. This is useful for cryptography. You can deploy what is 40 00:04:46,870 --> 00:04:52,460 called blind quantum computing. In fact, blind quantum computing is the ultimate 41 00:04:52,460 --> 00:04:59,930 privacy for the cloud, for example. Some guys try to launch, what they call a 42 00:04:59,930 --> 00:05:06,449 Quantum Internet. It's not so easy as cable networks and with a particular 43 00:05:06,449 --> 00:05:14,422 future for us, that could be cool to use if you use a Quantum Internet. Everyone 44 00:05:14,422 --> 00:05:22,610 that tried to spy you on the line will be detected. So it could be very useful. And 45 00:05:22,610 --> 00:05:30,669 of course, quantum computing brings to the massive new power of processing. 46 00:05:30,669 --> 00:05:42,830 But, how those computer works? This is the 1 slide quantum mechanic course. Why does 47 00:05:42,830 --> 00:05:49,910 a fancy new quantum computer are so powerful? In classical computing, we use 48 00:05:49,910 --> 00:06:01,020 bits. A bit is only in two states, 1 or 0. In quantum computing we replace the bits 49 00:06:01,020 --> 00:06:11,050 by the quantum bits, which we call then qubits. These qubits follow the quantum 50 00:06:11,050 --> 00:06:17,920 mechanical principle called superposition. And this principle is able to do, to 51 00:06:17,920 --> 00:06:26,980 provide to the user several step at the same time. So if you use a qubit, the 52 00:06:26,980 --> 00:06:34,039 qubit could be in the state of 0 and 1 nearly at the same time. It's not exactly 53 00:06:34,039 --> 00:06:39,190 what it is, but for us as a computer scientist, we could understand that it's a 54 00:06:39,190 --> 00:06:48,340 zero and 1 at the same time. And of course, if a quantum computer. This is a 55 00:06:48,340 --> 00:06:55,629 quantum computer wants to manage to deal with all this qubits. It deals with all 56 00:06:55,629 --> 00:07:02,000 the solution of the quantum register at the same time. So it will speed up the 57 00:07:02,000 --> 00:07:10,069 process of data computing because you take all the space generated by these quantum 58 00:07:10,069 --> 00:07:16,539 register and in one clock time the computer process all 59 00:07:16,539 --> 00:07:26,780 solution. This is mainly why and how the quantum computing is so powerful. So it's 60 00:07:26,780 --> 00:07:36,430 cool. So I want to build my own qubits. So this is my journey to build my own quantum 61 00:07:36,430 --> 00:07:43,160 computer. And you will see that there is some success and failure in most of the 62 00:07:43,160 --> 00:07:54,539 time. And I'm in the middle of this. So, I need to choose a technology to build 63 00:07:54,539 --> 00:08:00,509 my own qubit's hardware. This talk is mainly about hardwear. How to build your own 64 00:08:00,509 --> 00:08:07,310 hardrwear, to build your own quantum computer. So my ingredients, I need to 65 00:08:07,310 --> 00:08:14,770 find a suport at the hardware level that behave like quantum mechanic. Say you need 66 00:08:14,770 --> 00:08:20,679 to be able to do a quantum computer. So I need to find something that's behaving at 67 00:08:20,679 --> 00:08:28,860 atomic scale. I need to be able to build it. So I want to be able to use my do-it- 68 00:08:28,860 --> 00:08:35,677 yourself skills. And I want that my quantum computer works at room 69 00:08:35,677 --> 00:08:45,152 temperature. If it could be stable machine, it could be the best. There is 70 00:08:45,152 --> 00:08:55,080 many, many technology to build you on qubits. This one called superconducting 71 00:08:55,080 --> 00:09:05,200 qubits is used by a small startup like IBM, Google. Mainly the big one use this 72 00:09:05,200 --> 00:09:12,389 technology. Microsoft tried to use this technology. This technology with Diamond 73 00:09:12,389 --> 00:09:23,459 vacancy is used by university in Australia and in Ireland, I think. And of course, I 74 00:09:23,459 --> 00:09:31,000 use this technology. I use the technology called trap ions. So I trapped ions to 75 00:09:31,000 --> 00:09:39,730 make a quantum computer. So my low level hardware support and device to do some 76 00:09:39,730 --> 00:09:50,130 calculation with my quantum computer is atom. Why I choose an atom to make some 77 00:09:50,130 --> 00:09:55,680 fancy new qantum computer? The main reason is because I think I may be able to 78 00:09:55,680 --> 00:10:01,500 build it in my garage. It's enough affordable and well spread technology 79 00:10:01,500 --> 00:10:09,720 because we use technology that has been developed in 1945. There is a lot of 80 00:10:09,720 --> 00:10:16,279 experience with this type of technology. And the main reason, in fact, the qubit 81 00:10:16,279 --> 00:10:21,529 quality is better than any other technology. We have a long coherence time. 82 00:10:21,529 --> 00:10:27,380 If you have a long quantum coherence time, you can make much larger program, for 83 00:10:27,380 --> 00:10:37,029 example. So we need to share a bit of theory to understand how this type of 84 00:10:37,029 --> 00:10:43,120 computer works or I have a choice. I made a choice. I could have take time to make 85 00:10:43,120 --> 00:10:49,610 dozens of equations mainly I don't understand those equation to explain to 86 00:10:49,610 --> 00:10:55,470 you how to make some calculation with ions. But I found a video on YouTube and I 87 00:10:55,470 --> 00:11:01,880 would like to share you this two minutes only video to let you understand how at a 88 00:11:01,880 --> 00:11:08,310 theoretical point of view a quantum computer based on ion trap works. Let's 89 00:11:08,310 --> 00:11:10,310 see if it works. 90 00:11:10,310 --> 00:11:11,310 *background music starts* 91 00:11:11,310 --> 00:11:13,649 Video: .. electrically charged atoms make for excellent qubits. This kind of 92 00:11:13,649 --> 00:11:20,100 research has paved the way for a quantum computer prototype. Like an ordinary bit, 93 00:11:20,100 --> 00:11:25,579 a qubit can be a 1 or a 0. A Qubit differs from a bit because it can also be in 94 00:11:25,579 --> 00:11:32,889 combinations of these two states. An ion qubit is made from two of its energy 95 00:11:32,889 --> 00:11:37,490 levels. Ions of the same type are identical. So adding more qubits is 96 00:11:37,490 --> 00:11:43,190 simple. You just need to add more ions to the system. This is a major plus because a 97 00:11:43,190 --> 00:11:48,420 quantum computer will need lots and lots of qubits. Qubits must be configured in 98 00:11:48,420 --> 00:11:53,880 certain quantum states in order to perform quantum tasks. In an iron trap, taylored 99 00:11:53,880 --> 00:11:59,540 Laser pulses can change the energy of an eye on setting it into quantum state 1, 0 100 00:11:59,540 --> 00:12:09,230 or a combination of the two. The qubit surrounding environment sometimes sneakes 101 00:12:09,230 --> 00:12:15,769 in and destroys the qubit state, a covert act that can ruin a computation. But some 102 00:12:15,769 --> 00:12:20,440 ion energy levels are naturally isolated and scientists have come up with clever 103 00:12:20,440 --> 00:12:26,149 ways of adding an extra layers of protection. Quantum computer calculations 104 00:12:26,149 --> 00:12:31,259 are made from steps called Logic Gates. This will often involve more than one 105 00:12:31,259 --> 00:12:36,490 qubit, which means the qubits should be connected in some way in an iron trap. 106 00:12:36,490 --> 00:12:41,289 Neighboring ion qubits are connected through their collective motion. This 107 00:12:41,289 --> 00:12:46,180 happens because of their electrical repulsion. Laser pulses target the motion, 108 00:12:46,180 --> 00:12:52,139 enabling gates between any pair of qubits. To get the result of a calculation, 109 00:12:52,139 --> 00:12:57,340 scientists need to tell whether a qubit is in state 1 or 0. Shining laser pulses onto 110 00:12:57,340 --> 00:13:02,940 the ions makes only one of the two qubit levels flouress. So the result, light or 111 00:13:02,940 --> 00:13:08,480 no light, gives information about the calculation. Because many qubits are 112 00:13:08,480 --> 00:13:13,400 needed, quantum devices must be designed to be scalable. Researchers can only cram 113 00:13:13,400 --> 00:13:17,440 so many ions next to each other in a single ion trap before they get too 114 00:13:17,440 --> 00:13:22,970 unruly. But with modules each containing tens or hundreds of ions, they can start 115 00:13:22,970 --> 00:13:28,440 to wire up a large scale quantum computer. Flight from individual ion modules can be 116 00:13:28,440 --> 00:13:32,399 collected, allowing ion qubits from separate modules to communicate using 117 00:13:32,399 --> 00:13:38,480 photons rather than their motion. So far, scientists wired up two such modules and 118 00:13:38,480 --> 00:13:42,370 they are getting ready to deploy larger devices using several more. 119 00:13:42,370 --> 00:13:46,299 *background music stops* 120 00:13:46,299 --> 00:13:55,029 Yann: So now, congratulation, you are experts in ion trap quantum computing. A 121 00:13:55,029 --> 00:14:02,420 two minute video only is necessary. However, we like to build this quantum 122 00:14:02,420 --> 00:14:11,280 computer. So the plan is the following. We need some ions. You know that now. You 123 00:14:11,280 --> 00:14:16,361 need an ion trap. You need a vacuum chamber because we need to isolate our 124 00:14:16,361 --> 00:14:22,329 atom from the environment to maintain the quantum states. We need some laser, as you 125 00:14:22,329 --> 00:14:27,040 show in the video, to manipulate the quantum states. We need some low level 126 00:14:27,040 --> 00:14:34,581 software, to timely send the pulse of laser to manipulate the ions. And we 127 00:14:34,581 --> 00:14:45,120 need a camera to measure the ions quantum states. It's easy, no? So let's go to the 128 00:14:45,120 --> 00:14:55,049 difficult parts, I think mainly I would like to say that it's a work in progress. 129 00:14:55,049 --> 00:15:04,889 It's good, well, to say that it doesn't finished. And, just an alert, we need to 130 00:15:04,889 --> 00:15:14,190 manipulate, very high power electric voltages. So if you want to do this at home, do it 131 00:15:14,190 --> 00:15:22,630 at your own risk. It's not my fault. So how to create? First, we need to create an 132 00:15:22,630 --> 00:15:28,670 ion trap. How to create an ion trap and what is an iron trap? An iron trap is 133 00:15:28,670 --> 00:15:38,250 mainly a bunch of electrode with specific 3D or 2D geometry. We send to the 134 00:15:38,250 --> 00:15:44,516 electrode medium to high power voltage. AC voltage, alternative voltage. From 200 V 135 00:15:44,516 --> 00:15:55,980 to 6 kV. It's a big number for a voltage. We will use moderate to high frequency. 136 00:15:55,980 --> 00:16:02,690 This is due to the trap theory. Someone have won the Nobel Prize to explain that 137 00:16:02,690 --> 00:16:09,810 to trap an atom, you need to use an alternating voltage. And this electric 138 00:16:09,810 --> 00:16:15,050 voltage will make an electric field, and the goal of the electric field with the 139 00:16:15,050 --> 00:16:23,240 trap is just to maintain all the atom in a chain that will float over the air, over 140 00:16:23,240 --> 00:16:31,850 the trap. So how to achieve that at a small, small company budget, we say, 141 00:16:31,850 --> 00:16:41,629 because it's not for our best, I think. Let's go. So I use my ultra high tech 142 00:16:41,629 --> 00:16:52,990 military grade garage. I use 3D printer, local CNC machine, PCB milling techniques, 143 00:16:52,990 --> 00:16:58,830 only open source software. KiCad, FreeCAD, FlatCAM. KiCad for the electronics, 144 00:16:58,830 --> 00:17:05,170 FreeCAD for the mechanics and FlatCAM for the CNC. I used some high voltage 145 00:17:05,170 --> 00:17:14,449 transformer, classical electronics and of course isolated gloves. Security first. 146 00:17:14,449 --> 00:17:22,890 Safety first, sometime. And of course, I use eBay as a main procurement utility. 147 00:17:22,890 --> 00:17:31,130 First try. I need to make a classical Paul trap, of course, when I 148 00:17:31,130 --> 00:17:39,130 don't know, how it works, I go to Google and I find that some institution like CERN 149 00:17:39,130 --> 00:17:50,130 have a project to make an ion trap from 3D printed parts. I use conductive ink and 150 00:17:50,130 --> 00:17:57,420 only high voltage power supply. So I need to build this. There is the high voltage... 151 00:17:57,420 --> 00:18:06,240 air, two electrodes and one ring electrodes. The goal is to trap ions with 152 00:18:06,240 --> 00:18:14,570 that. So this is the main laboratory I use. So you have a variac. We take the 153 00:18:14,570 --> 00:18:21,870 electric plug from your domestic electric network. The high transformer and air. So 154 00:18:21,870 --> 00:18:31,080 the 3D printed, you have two electrode and the camera. This is the electrode. It's a 155 00:18:31,080 --> 00:18:41,429 very safe wiring system. For safety reason, I put some resistance here just to 156 00:18:41,429 --> 00:18:47,308 limit the currents. The first time. 157 00:18:47,308 --> 00:18:50,712 *laughter* 158 00:18:50,712 --> 00:18:55,090 Yann: In a more closer way you will see that the high voltages is coming from 159 00:18:55,090 --> 00:19:05,340 this. We will apply the voltages to the electrode. And the camera is here. Just 160 00:19:05,340 --> 00:19:16,880 see what the electrode will do. It works. I'm succeeding trapping some macro 161 00:19:16,880 --> 00:19:21,070 particles. This is not ion for demonstrational purposes, but we succeed to 162 00:19:21,070 --> 00:19:28,140 trap in the electrode some particles. Macro particles. But we have a first 163 00:19:28,140 --> 00:19:34,110 failure because with this geometry, we couldn't shine correctly the laser to 164 00:19:34,110 --> 00:19:40,840 manipulate the quantum state. First failure. Second try. We need to to make, 165 00:19:40,840 --> 00:19:48,450 another ion trap based on a new topology or new geometry of electrode. And this 166 00:19:48,450 --> 00:19:55,890 time we use a linear port to facilitate the laser shining. So again, I need to 167 00:19:55,890 --> 00:20:03,020 design on my own this new type because the CERN don't provide me the 3D 168 00:20:03,020 --> 00:20:10,540 printed parts. I use conductive ink and high voltages. So the goal is to design 169 00:20:10,540 --> 00:20:18,910 this. And in this trap you will see that we will trap the ion in the chain in 170 00:20:18,910 --> 00:20:32,220 the middle of the trap. So I use my 3D printer. I make some rods. The supports. 171 00:20:32,220 --> 00:20:41,600 Some electrodes. I built all the system and I plug the cable, the wiring and the 172 00:20:41,600 --> 00:20:49,770 trap. The particle will be trapped in these regions. For this second trap I 173 00:20:49,770 --> 00:20:56,539 didn't use a resistance to limit the currents, so it's impossible to touch this 174 00:20:56,539 --> 00:21:00,940 electrode because of death. 175 00:21:00,940 --> 00:21:06,780 *laughter* 176 00:21:06,780 --> 00:21:11,176 Yann: And it works, again. 177 00:21:11,176 --> 00:21:17,429 *applause* 178 00:21:17,429 --> 00:21:24,430 Yann: And in fact, this is a chain of particles that nearly clearly aligned. And 179 00:21:24,430 --> 00:21:33,289 this is my first quantum register of eight particles. But, this is the biggest 180 00:21:33,289 --> 00:21:40,760 failure, I need to put this ion trap in the vacuum chamber. A vacuum chamber is 181 00:21:40,760 --> 00:21:54,240 this type of thing. It's a big bunch of metal and we put the iron trap inside 182 00:21:54,240 --> 00:22:03,580 this. However, first, why we need a vacuum chamber? To be able to isolate particle 183 00:22:03,580 --> 00:22:10,589 from the other atoms in atmosphere to avoid collision between atoms. Because if 184 00:22:10,589 --> 00:22:15,159 we have collision between atoms, the quantum state is destroyed and the quantum 185 00:22:15,159 --> 00:22:23,430 processing is destroyed also. So we need a vacuum chamber. That's for them. 3D 186 00:22:23,430 --> 00:22:28,810 printing parts are not compatible with Ultra High Vacuum (UHV) environments. So 187 00:22:28,810 --> 00:22:38,940 it's a big fail? Are we doomed? Maker is our job, really. So we need to find a new 188 00:22:38,940 --> 00:22:44,390 solution. We have found one. So I need to find some materials that are compatible 189 00:22:44,390 --> 00:22:57,790 with UHV environments to build an ion trap. I ask the NASA, because NASA sends 190 00:22:57,790 --> 00:23:06,630 electronics in space. Space is like a big vacuum chamber. So they have a list of 191 00:23:06,630 --> 00:23:12,070 materials publicly available to be able to use some material that are compatible with 192 00:23:12,070 --> 00:23:21,779 a space condition. They are professionals. So what are the candidates, the material 193 00:23:21,779 --> 00:23:29,320 candidates for my ion trap? I need to use some gold for electronic conductor. I need 194 00:23:29,320 --> 00:23:36,330 to use ceramic for mechanical supports and Kapton cable for wiring inside the vacuum 195 00:23:36,330 --> 00:23:45,780 chamber. So maker is really, really our job, because I need to find an idea to 196 00:23:45,780 --> 00:23:55,730 transform my 3D printed linear ion trap to somthing that is compatible with UHV 197 00:23:55,730 --> 00:24:04,130 environments. So I need to read the manual. There is a lot of literature on 198 00:24:04,130 --> 00:24:13,590 quantum computers on Google, on Internet. So I have a bunch of books about quantum 199 00:24:13,590 --> 00:24:23,590 mechanics and research paper are full of details. I found this: some guys 200 00:24:23,590 --> 00:24:31,549 in 62 transform a linear Paul trap with rods to a planar ion trap with planar 201 00:24:31,549 --> 00:24:48,380 or surface electrodes. That's cool. So I need to transform this. To that. Oh, boy. 202 00:24:48,380 --> 00:24:59,970 I need to make my own chip. Price for complete chip factories are around 200 M$. 203 00:24:59,970 --> 00:25:07,490 I called Intel, they don't want to sell me one. And it's a bit out of my budget 204 00:25:07,490 --> 00:25:15,470 scope. A bit. Let's think this five minutes through to find a solution. In 205 00:25:15,470 --> 00:25:21,950 fact, it took me two months to find an affordable solution to do that. So I 206 00:25:21,950 --> 00:25:30,499 wanted to make a new design like a boss, of ion trap. I use a CNC, a 300$ CNC, come 207 00:25:30,499 --> 00:25:38,490 from Amazon. And then I found an empty ceramic chip carrier on eBay from a 208 00:25:38,490 --> 00:25:47,190 Norwegian guy. And I designed a simple KiCad PCB. So I use this. This is the 209 00:25:47,190 --> 00:25:54,500 ceramic chip supports. And what you see in yellow - it's gold. I designed in 210 00:25:54,500 --> 00:26:01,679 KiCad this PCB and this time we apply electric field, high voltage electric 211 00:26:01,679 --> 00:26:06,740 field to this electrode, this one and those one, and it creates an electric 212 00:26:06,740 --> 00:26:21,824 field to align all the macro particles of the ion in this line. And this is 213 00:26:21,824 --> 00:26:29,377 how I made my quantum computer chip. 214 00:26:29,377 --> 00:26:33,490 *applause* 215 00:26:33,490 --> 00:26:38,220 Yann: Thank you. And the better is that it works. 216 00:26:38,220 --> 00:26:46,177 *applause* 217 00:26:46,177 --> 00:26:49,010 So, tada, I have my first quantum 218 00:26:49,010 --> 00:26:57,210 computer done on my garage and just keep calm and except I'm a boss. 219 00:26:57,210 --> 00:26:58,620 *laughter* 220 00:26:58,620 --> 00:27:04,549 Yann: And it is not just a slide. Well, because if you want to see one of my 221 00:27:04,549 --> 00:27:12,700 prototype, I bring it so you can touch it and see how it works. But when you design 222 00:27:12,700 --> 00:27:20,250 such complex things; I am not a physicist, I'm just an engineer. A crazy one. But how 223 00:27:20,250 --> 00:27:26,809 to be sure that I am on the right road. I went to the Science Museum in London few 224 00:27:26,809 --> 00:27:34,700 few months ago, and there's this exhibition from our friend of GCHQ. Do you 225 00:27:34,700 --> 00:27:42,559 know what GCHQ is? It's like doing stuff of the UK and they made an exhibition 226 00:27:42,559 --> 00:27:53,029 about cryptography. And in this museum, they present a quantum computer based on 227 00:27:53,029 --> 00:28:03,539 ion trap technology. Thanks. This is the experimental part they show in this museum 228 00:28:03,539 --> 00:28:10,340 about quantum computer. In the right corner of this exhibition there is a 229 00:28:10,340 --> 00:28:17,070 wafer. On the wafer you have the electric design that they done to make their own 230 00:28:17,070 --> 00:28:31,758 iron trap. This is the design of the GCHQ. This is mine. I think I'm on the right road. 231 00:28:31,758 --> 00:28:39,010 *applause* 232 00:28:39,010 --> 00:28:42,640 Yann: Of course, I need to build my own vacuum chamber, it's not the difficult 233 00:28:42,640 --> 00:28:47,659 part, the vacuum chamber. It's just metal, you need not... you need some nuts, bolts, 234 00:28:47,659 --> 00:28:54,269 thin metal and pumps. A lot of pumps to suck out all the air in the vacuum. So I 235 00:28:54,269 --> 00:29:01,779 bought off eBay a different type of pumps. I like my vacuum chamber. This one, pretty 236 00:29:01,779 --> 00:29:10,500 one. And I put the ion trap inside the vacuum chamber. And for now I'm working on 237 00:29:10,500 --> 00:29:15,519 the laser, an optical setup. And this is the main difficult part for this quantum 238 00:29:15,519 --> 00:29:22,330 computer because with fancy, new, numerous wavelength for laser and we need to have a 239 00:29:22,330 --> 00:29:28,220 very precise wavelength to be able to manage all the atom, the energy level of 240 00:29:28,220 --> 00:29:37,149 the atom to make some calculation. So, of course, I could have and I have asked some 241 00:29:37,149 --> 00:29:46,940 professional of these devices to send me some proposal. A laser costs around 25 k€. 242 00:29:46,940 --> 00:29:56,860 A rule at least for this type of instrumentation. Or you can do it yourself 243 00:29:56,860 --> 00:30:08,350 from 2 k€. So I decide to make my own laser setup. I'm not a laser, optical or 244 00:30:08,350 --> 00:30:15,320 laser specialist. The first time I play with laser. And there is everything on 245 00:30:15,320 --> 00:30:22,130 the web. You can learn everything with the web. And I found this type of schematic 246 00:30:22,130 --> 00:30:29,570 you'd use either laser diode, some fancy optical lense. grating mirror that lets 247 00:30:29,570 --> 00:30:37,440 you choose... mainly choose what the reference frequency you want to use. 248 00:30:37,440 --> 00:30:45,200 There is a sort of loop control with PID control, which is for an electrician like 249 00:30:45,200 --> 00:30:53,760 me normal things to do. I don't know why all those fancy commercial product cost a 250 00:30:53,760 --> 00:31:00,840 lot. I don't know yet. Perhaps I will have some failure in the future, but I don't 251 00:31:00,840 --> 00:31:11,139 know. So I ask a guy on the Internet that's sold me a laser in kits. You can 252 00:31:11,139 --> 00:31:20,260 buy and mount you own laser. And this laser is controlled by an arduino. So you 253 00:31:20,260 --> 00:31:30,450 have fancy mirror. The HeNe aluminum laser tube and you can make your own laser at 254 00:31:30,450 --> 00:31:35,460 home also. I need a bunch of optical mounts and 255 00:31:35,460 --> 00:31:43,960 supports to support the lens and mirror, etc.. And as I bought a 3D printer for my 256 00:31:43,960 --> 00:31:49,519 iron trap that I can not use anymore because I use a vacuum chamber. I use the 257 00:31:49,519 --> 00:31:57,330 3D printer to make all the optical mounts, in fact. So it saved me my money again. 258 00:31:57,330 --> 00:32:02,730 However, you need to know that it's still a long road to have a complete quantum 259 00:32:02,730 --> 00:32:07,400 computer because I need to set up all these fancy optical and laser. This is my 260 00:32:07,400 --> 00:32:15,290 job at the moment. Mearly I have 6 months to one year of works. But the good 261 00:32:15,290 --> 00:32:21,730 news is that at the software level, everything exists. If you need to have a 262 00:32:21,730 --> 00:32:29,150 quantum compiler to make your code, it exist. At the moment, it's open source. If 263 00:32:29,150 --> 00:32:34,669 you need to have some firmware to make some pulse and laser control, it 264 00:32:34,669 --> 00:32:42,559 exists. And it is open source. So I am trying to convince you. Let me know if you 265 00:32:42,559 --> 00:32:50,960 agree with me, that doing a quantum computer at home - it's doable. You agree? 266 00:32:50,960 --> 00:32:58,232 *applause* 267 00:32:58,232 --> 00:33:06,992 Yann: But we are at the CCC. How to hack into a quantum computer. This is 268 00:33:06,992 --> 00:33:17,889 different part. It's easy. Just do what we do when we are infosec guy. Do the same 269 00:33:17,889 --> 00:33:27,510 things we do as usual: Hack the weakest link. You must know that when you build 270 00:33:27,510 --> 00:33:35,980 the quantum computer, there is few things that behave in the quantum mechanical 271 00:33:35,980 --> 00:33:42,950 regime. You just only need this chip, for example, and some laser. But all the 272 00:33:42,950 --> 00:33:50,049 equipment surrounding the quantum parts of the quantum computer is classical 273 00:33:50,049 --> 00:33:58,289 system. This is wave generator. Classical computer. Some IoTs, some programmable 274 00:33:58,289 --> 00:34:05,280 industrial systems. Sometimes they have IP address. If they have IP address, they are 275 00:34:05,280 --> 00:34:10,090 vulnerable. So, the main I have to know to hack into a quantum computer is to 276 00:34:10,090 --> 00:34:21,620 act the surrounding classical embedded system. So. Small company that is a 277 00:34:21,620 --> 00:34:29,880 competitor of me. It's cool. It's a startup called IBM. They used 278 00:34:29,880 --> 00:34:36,060 superconducting technology to build their own quantum computer. Their processor is 279 00:34:36,060 --> 00:34:43,870 just behind this delusion of refrigerator because they need to cool down their 280 00:34:43,870 --> 00:34:52,450 processor to be able to use the superconducting capability. Mine work at 281 00:34:52,450 --> 00:34:59,941 room temperature. And surrounded this processor, the researcher explained, this 282 00:34:59,941 --> 00:35:07,190 is a very good video to understand how it works. And surrounded this quantum 283 00:35:07,190 --> 00:35:17,280 part of their quantum computer you have a bunch of instruments. And if you zoom in, 284 00:35:17,280 --> 00:35:24,960 you see. If you zoom in this wave generator, it's a wave generator to send 285 00:35:24,960 --> 00:35:39,042 pulse to the superconducting processor. There is a sticker. And this sticker, in fact.. 286 00:35:39,042 --> 00:35:52,680 *laughter* 287 00:35:52,680 --> 00:36:01,830 Yann: So, of course, for security reason, I make some X to not show the complete 288 00:36:01,830 --> 00:36:14,330 passwords. So as a conclusion. I'm trying to convince you that quantum computing and 289 00:36:14,330 --> 00:36:19,340 quantum computer hardware is doable at home. So far cybersecurity or so-called 290 00:36:19,340 --> 00:36:27,146 cybersecurity specialist, you need to make, to adapt your own risk analysis. 291 00:36:27,146 --> 00:36:37,860 Because it's doable at home. Just understand that - is doable at home. They 292 00:36:37,860 --> 00:36:43,422 will, all these quantum computer will be used for good, bad and ugly. Just 293 00:36:43,422 --> 00:36:53,090 remember, GCHQ has a prototype in the museum. It would have fun if I could have 294 00:36:53,090 --> 00:37:02,204 seen the production quantum computer of the GCHQ. Of course quantum computers are 295 00:37:02,204 --> 00:37:10,270 hackable as any normal computer. So it's a good news for the cybersecurity industry. 296 00:37:10,270 --> 00:37:17,420 But you need as a community of maker in CCC, we need to be prepared to learn and 297 00:37:17,420 --> 00:37:23,820 how to use them, how to ask them how to program them and at the software level, 298 00:37:23,820 --> 00:37:30,300 just, you need to unlock your classical brain, the classical software brain 299 00:37:30,300 --> 00:37:36,590 because if I want to mention something at the software level, if you want to do some 300 00:37:36,590 --> 00:37:45,760 control codes, you need to be able to use your code without any variables. You can't 301 00:37:45,760 --> 00:37:51,460 use variables in quantum codes because if you use variable, you make a copy of a 302 00:37:51,460 --> 00:37:59,430 quantum state. Making a copy of a quantum state is impossible. So you can't use them 303 00:37:59,430 --> 00:38:05,310 to make a vairables or use variables in a new program and you can't debug it 304 00:38:05,310 --> 00:38:09,710 because if you debug it, you make a measurement. If you make a measurement, 305 00:38:09,710 --> 00:38:15,760 you destroy the quantum states. So be prepared to allow your brain to be able 306 00:38:15,760 --> 00:38:25,020 to make some code in the quantum world. But it's fun. Some time. Thanks for your 307 00:38:25,020 --> 00:38:30,500 attention. And if you have any question, it will be a pleasure. And as I'm French, 308 00:38:30,500 --> 00:38:47,291 I need to have a two hour lunchtime. 309 00:38:47,291 --> 00:38:52,430 Herald: Fantastic. Merci beaucoup. We have a lot of time now for your 310 00:38:52,430 --> 00:38:55,970 questions, answers. Line up at the microphones, please. And let's have a look 311 00:38:55,970 --> 00:39:01,170 if there is something from the Internet. Yes, there is. So please. First one from 312 00:39:01,170 --> 00:39:07,508 the Internet. Yann: Where is the Internet? 313 00:39:07,508 --> 00:39:12,650 Signal-Angel: All right. The Internet's quite impressed by your talk. So that's 314 00:39:12,650 --> 00:39:16,451 just a statement. Like everyone's very happy and pleased with your talk. 315 00:39:16,451 --> 00:39:19,339 Yann: Thanks to the Internet. *light laughter* 316 00:39:19,339 --> 00:39:23,850 Signal-Angel: All right. You have a few questions. So the first one is what 317 00:39:23,850 --> 00:39:29,200 properties should the element be chosen for the ion trap? 318 00:39:29,200 --> 00:39:32,620 Yann: What? Sorry. Signal-Angel: So what are the properties 319 00:39:32,620 --> 00:39:37,860 that should be looked at for choosing the element for the iron trap? 320 00:39:37,860 --> 00:39:45,360 Yann: What atom? I think the person asked what atom I used. I used the atom from 321 00:39:45,360 --> 00:39:52,820 calcium because those atoms have a specific.. because there is a lot of 322 00:39:52,820 --> 00:40:00,440 literature available. So it's easy for me to understand how it works. Researchers have 323 00:40:00,440 --> 00:40:05,180 done all the work before. And I used the atom because there are some energy level 324 00:40:05,180 --> 00:40:15,426 in this atom that is better protected from the environment. 325 00:40:15,426 --> 00:40:20,140 Herald: OK, let's quickly switch to microphone number 3. 326 00:40:20,140 --> 00:40:26,196 Mic3: Thank you for it. Thank you for your talk. My question is, what's the 327 00:40:26,196 --> 00:40:34,110 catch? If your design already exists in prototypes out there and it seems so much 328 00:40:34,110 --> 00:40:38,310 easier than working with superconductors, then why isn't everyone already doing 329 00:40:38,310 --> 00:40:42,300 this? Yann: Why someone choose superconducting 330 00:40:42,300 --> 00:40:45,760 and not ion trap technologies? Is that your question? 331 00:40:45,760 --> 00:40:47,760 Mic3: Correct. Yann: I don't know. 332 00:40:47,760 --> 00:40:53,970 *light laughter* Yann: Every time there is this type of 333 00:40:53,970 --> 00:40:58,740 question, why the big one used superconducting technology and why are you 334 00:40:58,740 --> 00:41:07,090 using iron trap technology? Mainly the answer could be that the big one is from 335 00:41:07,090 --> 00:41:13,500 the microelectronics domain. So a superconducting qubit is done on a 336 00:41:13,500 --> 00:41:19,350 wafer. So it's usual for this type of company to be able to build these 337 00:41:19,350 --> 00:41:26,734 type of qubits. I think it produced a habit. 338 00:41:26,734 --> 00:41:29,963 Mic3: Okay, thank you. Herald: Okay. Microphone number two, 339 00:41:29,963 --> 00:41:32,928 please. Mic2: I'm very impressed. But. 340 00:41:32,928 --> 00:41:37,150 Okay. You mentioned that hobbyists can't really afford this. A small company can. 341 00:41:37,150 --> 00:41:42,460 So just as a ballpark figure, I would like to ask the question. Nice. How much? 342 00:41:42,460 --> 00:41:57,520 Yann: All I've shown you here. It cost only less than 15 k€ of material for the 343 00:41:57,520 --> 00:42:05,934 moment. It is not for a hobbyist - for small company. 344 00:42:05,934 --> 00:42:11,563 Herald: Okay, one question from the Internet. Signal-Angel, please. 345 00:42:11,563 --> 00:42:16,460 Signal-Angel: All right. The next question is: is your next step going to be singling 346 00:42:16,460 --> 00:42:21,310 out individual ions? Yann: Sorry, can you repeat? 347 00:42:21,310 --> 00:42:26,150 Signal-Angel: Would your next step be singling out individual ions for the next 348 00:42:26,150 --> 00:42:32,220 step in your quantum computer? Yann: We try to manipulate single ions, 349 00:42:32,220 --> 00:42:38,650 but in fact, it's the goal with laser. With laser you shine a laser of individual 350 00:42:38,650 --> 00:42:46,890 qubits. And with another laser, you make a link between the ions with the common mode 351 00:42:46,890 --> 00:42:56,050 motion of the ion chain and you change the state of an individual ions, you transfer 352 00:42:56,050 --> 00:43:00,990 the state of these individual ions to the chain, which move because ions are 353 00:43:00,990 --> 00:43:07,260 electrically charged. So they repell each other and this act as a bus and you 354 00:43:07,260 --> 00:43:13,920 transfer the quantum state information to to a second ion to make a logic. So the 355 00:43:13,920 --> 00:43:24,200 goal effectively is to be able to manipulate one ions. We shine a laser on 356 00:43:24,200 --> 00:43:31,480 the individual atoms. This is the goal. Herald: Okay. Microphone number four, 357 00:43:31,480 --> 00:43:37,070 please. Mic4: Google announced recently that they 358 00:43:37,070 --> 00:43:43,840 achieved the quantum supremacy. What is your opinion on this theme? 359 00:43:43,840 --> 00:43:53,700 Yann: They have done a very good job for that. I think they show to the world 360 00:43:53,700 --> 00:43:58,960 for the first time that a quantum computer is able to do a calculation that a 361 00:43:58,960 --> 00:44:08,400 classical computer will never be able to do in the classical world. However, is 362 00:44:08,400 --> 00:44:16,990 that calculation useful? I'm not sure, except for one thing, it's able to 363 00:44:16,990 --> 00:44:25,470 certify the randomness of a number and it could be useful for the cyber security 364 00:44:25,470 --> 00:44:33,710 world. So it's I think and for my company, I have no money to spend to marketing 365 00:44:33,710 --> 00:44:39,570 thanks to Google because they show the world the power of quantum computer. So 366 00:44:39,570 --> 00:44:45,270 it's cool for me. Herald: Okay. Microphone number two, 367 00:44:45,270 --> 00:44:49,190 please. Mic2: Hello. Thanks for the nice talk. 368 00:44:49,190 --> 00:44:53,360 I'm a material scientist from Offline Gießen. Maybe you heard about our incident 369 00:44:53,360 --> 00:45:00,691 here. I was asking what are your current problems with this? For example, I mean, I 370 00:45:00,691 --> 00:45:05,300 think I have too many questions to ask here now. But for example, we saw that you 371 00:45:05,300 --> 00:45:11,521 had some like little pellets that were floating over your structure. But these 372 00:45:11,521 --> 00:45:16,080 are not the atoms that you are trying to to confine with each other so you can make 373 00:45:16,080 --> 00:45:21,990 calculations. So you didn't say anything about how you are trying to achieve this? 374 00:45:21,990 --> 00:45:26,790 And what is your current state? I mean, have you- could you start some crude 375 00:45:26,790 --> 00:45:31,960 calculations on this already or... ? Yann: Not for the moment because I need 376 00:45:31,960 --> 00:45:36,750 to shine the laser in the right direction. So for the moment, I am building the 377 00:45:36,750 --> 00:45:41,220 optical setup. Mic4: Okay. All right. Maybe there are 378 00:45:41,220 --> 00:45:44,310 some possibilities how I could help you with your project. 379 00:45:44,310 --> 00:45:48,260 Yann: You're welcome. Mic4: I have an access. If I could ask the 380 00:45:48,260 --> 00:45:54,930 right people, I'm not in a position to promise something to you now. But for 381 00:45:54,930 --> 00:45:59,960 example, we have an nano scribe laser system with this like a 3D printer. But 382 00:45:59,960 --> 00:46:03,474 you can build things on nano levels, on nano meter scale. 383 00:46:03,474 --> 00:46:08,800 Yann: What is the cost of using that? Mic4: The cost of the printer is around 384 00:46:08,800 --> 00:46:11,800 300.000€. Yann: Oh... I take it. 385 00:46:11,800 --> 00:46:17,910 Mic4: All right. Yann: Thanks. Thanks. Thanks for your 386 00:46:17,910 --> 00:46:20,330 help. Mic4: Maybe after the talk we can get in 387 00:46:20,330 --> 00:46:23,330 contact. Yann: Oh yes, we have a dinner. *laughter* 388 00:46:23,330 --> 00:46:26,018 Mic4: All right. Herald: Two new friends, actually. 389 00:46:26,018 --> 00:46:36,700 *applause* Question from the Internet, please. 390 00:46:36,700 --> 00:46:42,076 Signal Angel: All right. So how many qubits is it possible to make in the garage? 391 00:46:42,076 --> 00:46:49,710 Yann: For the prototype, we think we are able to do some 10 to 15 qubits with 392 00:46:49,710 --> 00:46:59,851 one ion trap. The goal is to chain the ion trap. So we have many, not as many as we 393 00:46:59,851 --> 00:47:08,420 want, but we could raise the number of qubits to 100 qubits. 394 00:47:08,420 --> 00:47:13,630 Heral: Okay. Microphone number three, please. 395 00:47:13,630 --> 00:47:17,430 Mic3: Which calculations do you plan to perform on your quantum computer? 396 00:47:17,430 --> 00:47:23,810 Yann: I don't care. I build thing and software guy do their code. It's not my 397 00:47:23,810 --> 00:47:30,510 job. Herald: Okay. Microphone number four, please. 398 00:47:30,510 --> 00:47:34,200 There is somebody. Mic4: Hello. So your optical setup 399 00:47:34,200 --> 00:47:38,730 reminded me of atomic force microscopes. Are you aware of what they are? 400 00:47:38,730 --> 00:47:43,170 Yann: Perhaps. Mic4: They are essentially an optical 401 00:47:43,170 --> 00:47:48,530 setup with a micro scale tip at the edge that rasters, that scans across the 402 00:47:48,530 --> 00:47:53,590 surface and can detect nanoscale features. But the cool thing is that even though 403 00:47:53,590 --> 00:47:57,960 this is a scientific instrument, there is also open hardware designs for that. And 404 00:47:57,960 --> 00:48:02,110 maybe you can see the ideas from that for your optical setup, because once again 405 00:48:02,110 --> 00:48:07,470 you've got precise lasers, at least on the geometrical side. They have to be 406 00:48:07,470 --> 00:48:10,960 precisely alined and everything. Yann: Thanks. Thanks for the information. 407 00:48:10,960 --> 00:48:15,280 And of course, we use a lot of spectrography techniques in this type of 408 00:48:15,280 --> 00:48:19,870 computer. Heral: Okay, we have somebody over there 409 00:48:19,870 --> 00:48:24,840 at microphone number three. Mic4: Did you consider optical quantum 410 00:48:24,840 --> 00:48:28,800 computers with entangled photons and such stuff? 411 00:48:28,800 --> 00:48:39,870 Yann: I did- This was my first choice in fact. However as far as I know, I'm not a 412 00:48:39,870 --> 00:48:45,100 physicist, it's difficult to make some and trick entanglements and, not 413 00:48:45,100 --> 00:48:51,660 entanglement... It's difficult to make some photon to talk to each other. Let's 414 00:48:51,660 --> 00:48:59,357 say that. So it's a complicated way to do something with multiple qubits. But 415 00:48:59,357 --> 00:49:06,685 photonic is a good technology because it works also at room temperature. But I 416 00:49:06,685 --> 00:49:09,190 prefer to have a vacuum chamber in my garage. 417 00:49:09,190 --> 00:49:19,712 Herald: Okay, let's interrogate the Internet again. 418 00:49:19,712 --> 00:49:24,820 Signal Angel: So you've mentioned that you should not be doing measurements on the 419 00:49:24,820 --> 00:49:29,310 quantum computer. So have you tried doing any measurements on your prototype? 420 00:49:29,310 --> 00:49:36,150 Yann: Measurement of what? Herald: This is hard. I think the Internet 421 00:49:36,150 --> 00:49:44,830 cannot really reply now. So can we... Yann: Internet is limited. I think we can 422 00:49:44,830 --> 00:49:50,850 give the guy that ask the question. He wants to send me the question I can answer 423 00:49:50,850 --> 00:49:53,900 just after all. Signal angel: But I think they are talking 424 00:49:53,900 --> 00:49:58,480 about electric field. Yann: Ah, no. I just I don't make any 425 00:49:58,480 --> 00:50:05,410 measurement. I'm an engineer. And as I am a good engineer, I just plug things. And 426 00:50:05,410 --> 00:50:14,370 just see what happens. I have no idea of the electric field generated. No idea. 427 00:50:14,370 --> 00:50:19,860 Again. Herald: OK. Microphone number two, please. 428 00:50:19,860 --> 00:50:24,910 Mic2: Hello. Thank you for the talk. So after you generate the vacuum in your 429 00:50:24,910 --> 00:50:29,480 vacuum chamber, how do you actually introduce the right number of ions and how 430 00:50:29,480 --> 00:50:32,660 do you keep them in the place where you need to have them? 431 00:50:32,660 --> 00:50:41,810 Yann: It's a good question. In fact, we don't introduce the ions. We put a calcium 432 00:50:41,810 --> 00:50:48,490 stone, sort of calcium stone, in a sort of oven, it's just a tube. We send current in 433 00:50:48,490 --> 00:50:53,690 this tube, the tube heats the calcium. They make some vapor and we shine a laser 434 00:50:53,690 --> 00:51:01,760 on the vapor of neutral atom of calcium. And this creates the ions. And this ion is 435 00:51:01,760 --> 00:51:07,170 trapped because it's now electric charged by the electrostatic field we make with 436 00:51:07,170 --> 00:51:14,140 the ion trap. So we just introduced before closing all the vacuum viewport and all 437 00:51:14,140 --> 00:51:20,550 the nuts and bolts. We just put a piece of stone of calcium, neutral atom. So 438 00:51:20,550 --> 00:51:27,980 everything is in the chamber before we turn on the quantum computer or the 439 00:51:27,980 --> 00:51:33,010 chamber. Herald: OK. We stay at microphone number 440 00:51:33,010 --> 00:51:38,990 two. There is another one. Mic2: OK, second question. What you're 441 00:51:38,990 --> 00:51:44,340 describing is you have a linear array of right now macroscopic particles. You will 442 00:51:44,340 --> 00:51:50,460 have a linear array of ions that are then coupled by kind of common vibrational 443 00:51:50,460 --> 00:51:56,700 modes. So they need to see each other's electrical fields. So I am wondering what 444 00:51:56,700 --> 00:52:04,450 the characteristic length scale between macroscopic particles versus ions would be 445 00:52:04,450 --> 00:52:08,300 if you want to have some meaningful vibrational modes that don't immediately 446 00:52:08,300 --> 00:52:15,020 get drowned by external thermal noise. Yann: So if I understand correctly the 447 00:52:15,020 --> 00:52:18,570 question, you ask me what is the dimension between the ions? 448 00:52:18,570 --> 00:52:24,110 Mic2: Yes. I mean you you are pretty big compared to the IBM guys. 449 00:52:24,110 --> 00:52:36,400 Yann: Yes. I'm big. Yes. You're right. The main dimension we use between ion it's few 450 00:52:36,400 --> 00:52:49,419 micron. And if some researcher succeeds to align 100 ions. So you have a chain of 100 451 00:52:49,419 --> 00:52:55,180 ions multiplied by five to ten microns between ions. This is the length. 452 00:52:55,180 --> 00:52:58,420 Mic2: But I mean, on your substrate, you ou have a fraction of a 453 00:52:58,420 --> 00:53:02,910 millimeter. Yeah. Between the... Yann: It's because it's prototype. 454 00:53:02,910 --> 00:53:08,030 Mic2: Okay. Yann: You're right. I need to squeeze the 455 00:53:08,030 --> 00:53:21,080 design a little bit, okay? It just need to buy a better CNC machine. 456 00:53:21,080 --> 00:53:27,150 Herald: OK, we got some question from the Internet again. 457 00:53:27,150 --> 00:53:36,600 Signal angel: All right so this one is... This is more towards knowing about the 458 00:53:36,600 --> 00:53:40,306 GCHQ exhibition. Is it still open do you know? 459 00:53:40,306 --> 00:53:44,750 Yann: Yes. I think, you have a free ticket if you want. It's free. In 460 00:53:44,750 --> 00:53:48,840 fact, it's free. Signal Angel: I guess people will contact 461 00:53:48,840 --> 00:53:52,719 you on Twitter for that. Yann: Yeah. I make some touristic business 462 00:53:52,719 --> 00:53:56,586 or so. I can help. Signal Angel: Everyone was impressed with 463 00:53:56,586 --> 00:54:01,330 your GCHQ hack. Herald: Ok, any more questions. How many 464 00:54:01,330 --> 00:54:06,571 people are working in your garage? Yann: There is me and sometimes one of 465 00:54:06,571 --> 00:54:12,712 my daughters, which is, 10 years old. Herald: Pro team. 466 00:54:12,712 --> 00:54:16,973 Yann: Yeah, a big one. Herald: Okay. Any more questions from 467 00:54:16,973 --> 00:54:25,141 the audience, from the Internet? We have time. Okay, I'm gonna close that session 468 00:54:25,141 --> 00:54:27,940 now, thank you very much. Big applause again for Yann. 469 00:54:27,940 --> 00:54:35,224 *applause* 470 00:54:35,224 --> 00:54:50,050 *36C3 outro music* 471 00:54:50,050 --> 00:55:02,000 Subtitles created by c3subtitles.de in the year 2021. Join, and help us!