WEBVTT 1 00:00:01.199 --> 00:00:06.200 Welcome to the Sentient Code, where intelligence is engineered, autonomy 2 00:00:06.280 --> 00:00:10.439 is emerging, and a line between human and machine grows thinner. 3 00:00:10.800 --> 00:00:15.359 Each episode, we decode the algorithms, explore the robotics, and 4 00:00:15.439 --> 00:00:21.920 examine the ideas shaping the future of artificial minds. 5 00:00:23.920 --> 00:00:28.960 Imagine for a second, just a single master Selton key. Okay, 6 00:00:29.000 --> 00:00:30.679 but I'm not talking about a key for you know, 7 00:00:30.839 --> 00:00:34.679 physical doors. I want you to picture a mathematical skeleton 8 00:00:34.799 --> 00:00:38.000 key that could instantly unlock every digital vault in the 9 00:00:38.159 --> 00:00:38.799 entire world. 10 00:00:38.960 --> 00:00:42.799 Yeah. Yeah, every single bank account, every secure email server, 11 00:00:43.359 --> 00:00:45.560 every classified government database, just. 12 00:00:45.560 --> 00:00:46.920 All laid bare exactly. 13 00:00:46.960 --> 00:00:51.560 We're talking about trillions of dollars locked in cryptocurrency, global 14 00:00:51.600 --> 00:00:55.719 financial systems, the entire backbone of e commerce, and I 15 00:00:55.719 --> 00:00:59.159 mean even your own personal medical records just click wide open, 16 00:00:59.359 --> 00:01:01.479 the digital entirely shattered. 17 00:01:01.799 --> 00:01:04.200 It sounds like, you know, the inciting incident of some 18 00:01:04.400 --> 00:01:07.799 cyber thriller movie, but it's actually the very real, very 19 00:01:07.920 --> 00:01:10.439 urgent reality of where our technology is heading right now. 20 00:01:10.680 --> 00:01:13.640 Yeah, because in late March of twenty twenty six, so recently, 21 00:01:13.879 --> 00:01:17.040 a consortium of researchers from Caltech, UC Berkeley and this 22 00:01:17.959 --> 00:01:19.040 quantum startup. 23 00:01:18.680 --> 00:01:20.920 Called Ori Atomic led by Medlin Kane, right. 24 00:01:20.799 --> 00:01:24.439 Led by Madeline Kane. Yeah, they prove that building this 25 00:01:24.560 --> 00:01:28.280 master key is not just some theoretical Pike dream anymore. 26 00:01:28.680 --> 00:01:33.239 It is drastically easier and drastically closer to reality than 27 00:01:33.280 --> 00:01:34.799 anyone previously. 28 00:01:34.280 --> 00:01:36.319 Thought, which is terrifying, it really is. 29 00:01:36.719 --> 00:01:40.040 We are looking at this rapidly shrinking countdown toward what 30 00:01:40.079 --> 00:01:43.000 the cybersecurity world calls Q day, right, Q. 31 00:01:43.159 --> 00:01:46.120 Day, which is the day a quantum computer finally becomes 32 00:01:46.159 --> 00:01:50.680 powerful enough to just break modern encryption. Yep. But you know, 33 00:01:50.879 --> 00:01:53.359 to understand the magnitude of what the Ari Atomic and 34 00:01:53.400 --> 00:01:55.400 Caltech research has just pulled off, we kind of need 35 00:01:55.439 --> 00:01:58.079 to understand the locks that are currently on our digital 36 00:01:58.120 --> 00:02:01.719 doors right now. Makes sense so rely heavily on classical 37 00:02:01.799 --> 00:02:06.560 encryption standards, things like RSA and ECC, which stands for 38 00:02:06.840 --> 00:02:08.360 elliptic curve cryptography. 39 00:02:08.520 --> 00:02:11.199 Right, So, whenever your browser shows that little padlock icon 40 00:02:11.240 --> 00:02:12.759 at the top, or when you send a message on 41 00:02:12.800 --> 00:02:14.680 a secure app, or even when you just buy something 42 00:02:14.680 --> 00:02:17.800 with a credit card online, you're relying on those exact 43 00:02:17.879 --> 00:02:21.919 algorithms exactly. And the way those traditional algorithms protect your 44 00:02:22.039 --> 00:02:25.240 data is well, I think it's fascinating. I always like 45 00:02:25.319 --> 00:02:28.199 to picture classical encryption like mixing paint. 46 00:02:28.360 --> 00:02:29.439 Oh, I like that analogy. 47 00:02:29.560 --> 00:02:32.360 Yeah. So if I take a bucket of bright yellow 48 00:02:32.400 --> 00:02:35.520 paint and I mix it with a very specific shade 49 00:02:35.520 --> 00:02:39.039 of dark blue, I easily get a new bucket of 50 00:02:39.080 --> 00:02:42.319 green paint. That's the encryption process. It's very easy to 51 00:02:42.360 --> 00:02:43.120 do in one direction. 52 00:02:43.280 --> 00:02:43.439 Right. 53 00:02:43.439 --> 00:02:45.159 But if I hand you that bucket of green paint 54 00:02:45.240 --> 00:02:48.479 and tell you to unmix it, do luck exactly to 55 00:02:48.479 --> 00:02:52.240 give me back the exact original shades of yellow and 56 00:02:52.280 --> 00:02:55.319 blue just by looking at the green. Yeah, you are 57 00:02:55.360 --> 00:02:57.639 going to have a practically impossible time. 58 00:02:57.840 --> 00:02:59.520 Yeah, you can't just pull the colors apart. 59 00:02:59.639 --> 00:03:02.360 And that's what a normal classical computer is trying to 60 00:03:02.360 --> 00:03:05.680 do when it attempts to break encryption. Even the most 61 00:03:05.680 --> 00:03:08.240 powerful supercomputer in the world would have to sit there 62 00:03:08.759 --> 00:03:12.240 just trying combination after combination of blue and yellow for 63 00:03:12.800 --> 00:03:16.919 literally millions of years, just guessing. That's guessing. The math 64 00:03:16.960 --> 00:03:21.599 behind unmixing the paint, which usually involves facturing unimaginably large numbers, 65 00:03:22.000 --> 00:03:23.879 is just it's too hard to brute force. 66 00:03:23.759 --> 00:03:28.120 Enter quantum computing, right, So spistically, this mathematical concept called 67 00:03:28.199 --> 00:03:32.280 Shores algorithm, which is actually formulated way back in nineteen 68 00:03:32.319 --> 00:03:35.159 ninety four. Yeah, it's been around a while for decades. 69 00:03:35.240 --> 00:03:38.319 It's been sort of the boogeyman of cryptography. Now. A 70 00:03:38.360 --> 00:03:41.520 lot of people think Shor's algorithm just works by I 71 00:03:41.520 --> 00:03:45.719 don't know, guessing the paint colors really really fast. 72 00:03:45.560 --> 00:03:47.759 Right, like a superpowered normal computer. 73 00:03:48.039 --> 00:03:50.400 Yeah, but that's not how it works at all, is 74 00:03:50.400 --> 00:03:50.800 it not? 75 00:03:50.879 --> 00:03:54.639 At all? A quantum computer doesn't just guess faster. Shores 76 00:03:54.759 --> 00:03:58.840 algorithm approaches the problem from an entirely different dimension. It 77 00:03:58.960 --> 00:04:01.520 utilizes something called quantum interference. 78 00:04:01.840 --> 00:04:03.000 Okay, interference. 79 00:04:03.120 --> 00:04:05.400 Yeah, the best way to understand interference is to think 80 00:04:05.400 --> 00:04:07.039 about noise canceling headphones. 81 00:04:07.120 --> 00:04:09.400 Oh interesting, So those headphones. 82 00:04:08.960 --> 00:04:11.039 They listen to the ambient noise around you, and they 83 00:04:11.080 --> 00:04:14.039 create an exact opposite sound wave to cancel it out right, 84 00:04:14.560 --> 00:04:15.639 leaving you with silence. 85 00:04:15.719 --> 00:04:19.160 So Shores algorithm acts like a mathematical noise canceler. 86 00:04:19.319 --> 00:04:23.360 Exactly. When a quantum computer runs Shores algorithm, it sets 87 00:04:23.439 --> 00:04:26.319 up this massive calculation where all the incorrect. 88 00:04:25.920 --> 00:04:27.800 Answers, the wrong paint colors. 89 00:04:27.759 --> 00:04:31.120 Right, the wrong pink colors. They literally interfere with each 90 00:04:31.120 --> 00:04:34.040 other and cancel themselves out. Yeah, and at the exact 91 00:04:34.040 --> 00:04:37.000 same time, the correct answer is amplified. 92 00:04:37.199 --> 00:04:37.680 Wow. 93 00:04:37.759 --> 00:04:40.639 So instead of trying every combination one by one, the 94 00:04:40.720 --> 00:04:43.639 quantum computer functions more like a prism. You shine the 95 00:04:43.639 --> 00:04:46.920 green paint through the prism and it instantly separates the 96 00:04:47.000 --> 00:04:49.319 light back into the original yellow and blue wavelengths. 97 00:04:49.360 --> 00:04:52.040 It just solves the problem almost instantly, instantly, which is 98 00:04:52.079 --> 00:04:55.560 absolutely terrifying. But you know, for the last thirty years, 99 00:04:55.600 --> 00:04:58.279 we've had a very comforting safety net, right we have. Yeah, 100 00:04:58.319 --> 00:05:01.199 because to run shores algorithm on a scale large enough 101 00:05:01.199 --> 00:05:04.879 to actually break global encryption, you need a massive, highly 102 00:05:04.920 --> 00:05:09.360 capable quantum computer, and building one is incredibly. 103 00:05:08.720 --> 00:05:14.399 Difficult, so difficult because quantum bits or quibbits are notoriously unstable. 104 00:05:14.720 --> 00:05:19.560 They're incredibly fragile. Slight temperature change, a stray microwave, or 105 00:05:19.600 --> 00:05:21.319 even like a cosmic ray. 106 00:05:21.160 --> 00:05:23.079 From space cosmic ray. 107 00:05:23.120 --> 00:05:26.800 Yeap, literally a particle from space can cause a quibot 108 00:05:26.920 --> 00:05:30.240 to lose its quantum state. That interference is called noise, 109 00:05:30.319 --> 00:05:32.319 and it causes massive calculation errors. 110 00:05:32.439 --> 00:05:34.240 Okay, so they're super sensitive. 111 00:05:33.920 --> 00:05:38.319 Extremely so to get one reliable, error free logical quibt, 112 00:05:38.879 --> 00:05:43.399 you need a massive number of noisy physical quibots to 113 00:05:43.600 --> 00:05:46.720 basically act as backups and handle the error correction. 114 00:05:46.959 --> 00:05:51.439 Which brings us to honestly, the incredible shrinking number. The 115 00:05:51.480 --> 00:05:53.959 timeline of this threat is mind blowing when you map 116 00:05:53.959 --> 00:05:56.000 it out, it really is. Let's walk through the data, 117 00:05:56.079 --> 00:05:59.279 because to understand why this March twenty twenty six discovery 118 00:05:59.360 --> 00:06:02.519 is causing such a global shock wave, you have to 119 00:06:02.560 --> 00:06:05.199 see how incredibly fast the goalposts are moving here. 120 00:06:05.360 --> 00:06:06.639 Yeah, let's look at the history. 121 00:06:06.680 --> 00:06:09.199 Back in twenty twelve, experts looked at that noise problem 122 00:06:09.240 --> 00:06:12.079 you mentioned and estimated it would take roughly one billion 123 00:06:12.120 --> 00:06:15.680 physical quibbets to run Shor's algorithm and break standard encryption. 124 00:06:15.920 --> 00:06:20.199 Right, one billion physical quibits is just a staggering engineering hurdle. 125 00:06:20.720 --> 00:06:23.720 At the time, looking at that number, QDA felt like 126 00:06:23.759 --> 00:06:26.000 it was a century way. Yeah, it felt like a 127 00:06:26.040 --> 00:06:29.000 problem for you know, our great grandchildren to figure out. 128 00:06:29.160 --> 00:06:32.000 But then the math started getting better. By twenty nineteen, 129 00:06:32.240 --> 00:06:35.839 Google's Craig Gidney and his team managed to drastically optimize 130 00:06:35.839 --> 00:06:38.639 the error correction process and they brought that estimate down 131 00:06:38.680 --> 00:06:41.639 from one billion to about twenty million quibuits. 132 00:06:41.399 --> 00:06:44.319 Which is a massively, hugely, But Gidney and his team 133 00:06:44.560 --> 00:06:49.000 didn't stop there. By twenty twenty five, through even further refinements, 134 00:06:49.079 --> 00:06:52.959 they slashed that requirement again to under one million noisy equibits, 135 00:06:53.000 --> 00:06:53.160 and the. 136 00:06:53.199 --> 00:06:56.800 Number just keep tumbling. It's crazy. Early in twenty twenty six, 137 00:06:56.839 --> 00:07:00.959 a company called Iceberg Quantum introduced their Pinnacle architecture, which 138 00:07:01.000 --> 00:07:04.279 brought the number down to roughly one hundred thousand kubits. Yeah, 139 00:07:04.399 --> 00:07:06.639 think about that for a second. We went from a 140 00:07:06.680 --> 00:07:09.720 billion to one hundred thousand in just over a decade. 141 00:07:09.879 --> 00:07:11.600 Its exponential progress. 142 00:07:11.680 --> 00:07:14.519 And now this ari Atomic and caltech research drops the 143 00:07:14.519 --> 00:07:17.920 floor completely out from under us completely. Their study reveals 144 00:07:17.959 --> 00:07:21.279 that to crack ECC two fifty six, which for context, 145 00:07:21.639 --> 00:07:26.319 is the exact encryption securing Bitcoin ethereum countless blockchain wallet, 146 00:07:26.480 --> 00:07:28.879 all the crypto, all the crypto, it could take as 147 00:07:28.959 --> 00:07:31.839 few as ten thousand to twenty six thousand cubits, and 148 00:07:31.920 --> 00:07:34.040 it would only take about ten days of computing time 149 00:07:34.120 --> 00:07:35.199 to shatter the locks. 150 00:07:35.639 --> 00:07:38.720 Ten days. And for RSA twenty forty eight, which is 151 00:07:38.720 --> 00:07:44.480 the backbone of HTTPS web traffic, your secure VPN's global. 152 00:07:44.079 --> 00:07:46.480 Banking, the really important stuff, right, they. 153 00:07:46.399 --> 00:07:48.920 Estimate it requires just about one hundred and two thousand cubits, 154 00:07:49.199 --> 00:07:50.800 cracking it in around ninety seven days. 155 00:07:50.800 --> 00:07:52.439 Okay, let me stop you there and just push back 156 00:07:52.439 --> 00:07:54.000 on this for a second, because I'm trying to put 157 00:07:54.000 --> 00:07:55.879 myself in the shoes of an engineer listening to this. 158 00:07:56.360 --> 00:08:00.000 Ten thousand cubits still sounds like a really massive machine 159 00:08:00.079 --> 00:08:03.560 to physically build it. Does We see companies like IBM 160 00:08:03.600 --> 00:08:06.079 and Google in the news and they're building systems with 161 00:08:06.360 --> 00:08:09.160 what a few hundred maybe one thousand kubus right around there, 162 00:08:09.199 --> 00:08:12.399 so they were still clearly in the prototype phase. Why 163 00:08:12.600 --> 00:08:16.360 is ten thousand the absolute panic button number? Shouldn't we 164 00:08:16.360 --> 00:08:18.120 still feel safe for a few more decades. 165 00:08:18.319 --> 00:08:20.839 Well, that is the most common misconception right now, and 166 00:08:20.879 --> 00:08:23.319 it is a really dangerous one. We can't look at 167 00:08:23.319 --> 00:08:28.160 a ten thousand cubit requirement as some distant theoretical target anymore. 168 00:08:28.160 --> 00:08:31.560 Why not because it is actively on the corporate roadmap 169 00:08:31.920 --> 00:08:34.960 of the biggest tech companies in the world. IBM is 170 00:08:35.120 --> 00:08:39.080 already publicly targeting systems with ten thousand or more kubits 171 00:08:39.159 --> 00:08:40.519 by the year twenty twenty nine. 172 00:08:40.720 --> 00:08:42.679 Wait, by twenty twenty nine, that is just a few 173 00:08:42.759 --> 00:08:43.279 years away. 174 00:08:43.440 --> 00:08:46.759 Exactly. That is why this cybersecurity world is on high alert. 175 00:08:47.080 --> 00:08:50.600 The finish line for breaking global encryption has just been 176 00:08:50.639 --> 00:08:54.279 moved right to where the runners currently are. Wow, this 177 00:08:54.399 --> 00:08:58.799 ari atomic study violently shifts the timeline for a cryptographically 178 00:08:59.080 --> 00:09:04.679 relevant quantum computer from maybe fifty years away to the 179 00:09:04.759 --> 00:09:06.919 late twenty twenties or early twenty thirties. 180 00:09:06.960 --> 00:09:10.679 So how did these researchers actually achieve this massive reduction? 181 00:09:10.799 --> 00:09:12.600 I mean, to get from a million quibits down to 182 00:09:12.639 --> 00:09:14.759 ten thousand. They couldn't just write better software. 183 00:09:14.919 --> 00:09:18.600 No, No, they had to completely rethink the physical hardware itself. Okay, 184 00:09:18.759 --> 00:09:21.799 they moved away from traditional systems, which usually rely on 185 00:09:21.840 --> 00:09:25.720 superconducting circuits and microscopic wires cool to absolute zero, and 186 00:09:25.759 --> 00:09:28.399 they turn to something called neutral atom quantum. 187 00:09:28.080 --> 00:09:31.600 Platforms, which I have to say sounds completely wild. 188 00:09:31.840 --> 00:09:35.159 It is brilliant engineering. Instead of manufacturing quibits on a 189 00:09:35.200 --> 00:09:39.559 silicon chip like traditional processors, a neutral atom platform uses 190 00:09:39.679 --> 00:09:44.240 actual individual atoms suspended in a completely empty vacuum chamber. 191 00:09:44.080 --> 00:09:46.360 Suspended by lasers. I was reading about this and it 192 00:09:46.399 --> 00:09:49.440 honestly blew my mind. They literally use a rays of 193 00:09:49.600 --> 00:09:53.120 highly focused lasers. I think they're often called optical tweezers. 194 00:09:53.399 --> 00:09:54.919 Optical tweezers to trap and. 195 00:09:54.919 --> 00:09:58.320 Hold individual atoms in place, so they can act as 196 00:09:58.399 --> 00:10:03.039 the quibbits trying to hold a grain of sand with tweezers, 197 00:10:03.559 --> 00:10:06.159 but the grain of sand is a single atom, and 198 00:10:06.200 --> 00:10:08.120 the tweezers are made of pure light. 199 00:10:08.399 --> 00:10:11.720 It's incredible and the flexibility this offers is what allowed 200 00:10:11.720 --> 00:10:15.279 the Ari Atomic team to drastically lower the quibbit requirement. 201 00:10:15.399 --> 00:10:19.759 Oh so they utilize what they call reconfigurable atomic arrays. 202 00:10:20.519 --> 00:10:25.399 See in a traditional superconducting quantum computer, the quibots are 203 00:10:25.399 --> 00:10:29.799 physically wired together in a fixed position on a chip. 204 00:10:29.559 --> 00:10:30.879 Like hardwired right. 205 00:10:30.919 --> 00:10:32.720 So if quibbit A on the left side of the 206 00:10:32.799 --> 00:10:35.639 chip needs to interact with quibit Z on the right side, 207 00:10:35.759 --> 00:10:38.639 the information has to travel through a bunch of intermedia equibits. 208 00:10:39.159 --> 00:10:41.399 Every time you pass that information, it takes time and 209 00:10:41.440 --> 00:10:44.000 it introduces massive opportunities for noise and errors. 210 00:10:44.039 --> 00:10:45.919 Okay, actually came up with an analogy for this. 211 00:10:46.000 --> 00:10:47.200 Tell me at this track, so let's hear it. 212 00:10:47.200 --> 00:10:49.799 Think of a traditional quantum system like a fixed assembly 213 00:10:49.840 --> 00:10:52.559 line in a factory. A part has to travel down 214 00:10:52.600 --> 00:10:57.639 a very long, rigid conveyor belt, passing station after static 215 00:10:57.720 --> 00:11:00.080 station just to get to the end of the line. 216 00:11:00.159 --> 00:11:04.320 It's slow and it's inflexible. Right, But this new neutral 217 00:11:04.360 --> 00:11:08.360 atom system, because the quibbets are held by lasers, the 218 00:11:08.440 --> 00:11:12.360 lasers can actually physically move the atoms around in real time. 219 00:11:12.919 --> 00:11:15.759 So instead of a rigid assembly line, it's more like 220 00:11:15.799 --> 00:11:18.879 an intricate ballroom dance. Oh I could They can pluck 221 00:11:18.879 --> 00:11:21.759 an atom from one side of the array, gracefully, move 222 00:11:21.759 --> 00:11:24.000 it across the floor, and place it right next to 223 00:11:24.039 --> 00:11:26.480 another atom, so they can interact with exactly the right 224 00:11:26.519 --> 00:11:28.360 partner at the exact right time. 225 00:11:28.600 --> 00:11:32.919 That analogy tracks perfectly, and that dynamic ballroom dance allows 226 00:11:32.960 --> 00:11:36.960 for massive parallel operations. It's a brilliant optimization of. 227 00:11:36.960 --> 00:11:39.879 Physical space because they don't need the conveyor belt exactly. 228 00:11:39.960 --> 00:11:42.559 Because they can move the quippets around so freely, they 229 00:11:42.639 --> 00:11:46.399 significantly reduce the overhead needed for things called magic state factories. 230 00:11:46.879 --> 00:11:49.440 Okay, magic state factories. That sounds like something out of 231 00:11:49.440 --> 00:11:52.519 a fantasy novel. What does a factory actually do inside 232 00:11:52.519 --> 00:11:53.399 a quantum computer? 233 00:11:53.840 --> 00:11:56.399 Well, think of a magic state factory like a massive 234 00:11:56.440 --> 00:12:00.720 prep kitchen in a very busy restaurant. Okay, I'm computing 235 00:12:01.360 --> 00:12:06.559 certain complex calculations require very specific, highly prepared quantum states 236 00:12:07.200 --> 00:12:11.360 the ingredients, so to speak. Historically, you needed thousands of 237 00:12:11.360 --> 00:12:15.080 extra kubits just acting as static prep kitchens, taking up 238 00:12:15.200 --> 00:12:17.639 enormous amounts of space on the chip. But because the 239 00:12:17.720 --> 00:12:20.320 lasers can move the atoms around, you do not need 240 00:12:20.360 --> 00:12:23.320 a thousand separate kitchens anymore. Oh I see, Yeah, The 241 00:12:23.360 --> 00:12:26.960 atoms can physically travel to one highly efficient central kitchen, 242 00:12:27.639 --> 00:12:30.799 grab the prepared ingredients, and move back into the calculation. 243 00:12:31.720 --> 00:12:34.480 That alone eliminates the need for hundreds of thousands of 244 00:12:34.519 --> 00:12:35.440 physical kubits. 245 00:12:35.519 --> 00:12:37.960 That is so smart. And the researchers also paired this 246 00:12:38.039 --> 00:12:40.399 ballroom dance with a new type of error correction called 247 00:12:40.440 --> 00:12:43.120 Lebe codes. Right, yes, what makes a BB code better 248 00:12:43.159 --> 00:12:44.360 than what we were using before? 249 00:12:44.679 --> 00:12:48.759 So BB codes act is an extraordinarily efficient autocorrect. In 250 00:12:48.879 --> 00:12:52.320 older systems, an autocorrect protocol might only check one or 251 00:12:52.320 --> 00:12:55.159 two adjacent quibits to see if an error occurred. 252 00:12:54.919 --> 00:12:56.720 Like checking your spelling word by word. 253 00:12:56.960 --> 00:13:01.200 Right, But BB codes can check multiple neighboring bit simultaneously, 254 00:13:01.600 --> 00:13:05.279 suppressing errors beautifully. When you combine the moving atoms with 255 00:13:05.360 --> 00:13:08.679 this specific BB code autocorrect. They're able to operate at 256 00:13:08.679 --> 00:13:11.039 a physical error rate of just zero point one percent. 257 00:13:11.360 --> 00:13:11.799 Wow. 258 00:13:11.960 --> 00:13:14.840 Yeah, that combination is what drops the baseline for a 259 00:13:14.879 --> 00:13:20.159 cryptographically relevant implementation down to an astonishing nine hundred and 260 00:13:20.159 --> 00:13:20.679 sixty one. 261 00:13:20.639 --> 00:13:24.440 Quipots nine hundred and sixty one. That is so specific 262 00:13:24.480 --> 00:13:27.559 and so low. Now we do need to ground this 263 00:13:27.639 --> 00:13:29.279 a bit. I don't want anyone listening to think that 264 00:13:29.360 --> 00:13:32.759 can just order a ten thousand kubet neutral atom machine 265 00:13:32.759 --> 00:13:33.559 on Amazon to market. 266 00:13:33.600 --> 00:13:34.360 Oh, definitely not. 267 00:13:34.519 --> 00:13:37.120 There are still formidable engineering hurdles, right. 268 00:13:37.080 --> 00:13:41.200 Absolutely, Moving atoms with lasers in a vacuum sounds elegant, 269 00:13:41.600 --> 00:13:45.080 but managing that intense laser control without accidentally causing the 270 00:13:45.120 --> 00:13:47.159 atoms to bump into each other or lose their quantum 271 00:13:47.159 --> 00:13:50.320 state is incredibly difficult. I can imagine that loss of 272 00:13:50.320 --> 00:13:53.159 state is called decoherence. Think of a qubit like a 273 00:13:53.159 --> 00:13:56.080 spinning top. As long as it is spinning perfectly, it 274 00:13:56.120 --> 00:13:59.440 holds its quantum information. But the slightest bump from a 275 00:13:59.480 --> 00:14:02.440 laser or a stray particle of light causes the top 276 00:14:02.480 --> 00:14:03.480 to wobble and fall. 277 00:14:03.320 --> 00:14:05.480 Over, and that ruins the calculation. 278 00:14:05.840 --> 00:14:09.360 Right, That is decoherence. They also have to master real 279 00:14:09.440 --> 00:14:13.799 time decoding, meaning the classical computers tracking the errors have 280 00:14:13.879 --> 00:14:19.159 to process the autocorrect data at blistering speeds. So yeah, 281 00:14:19.240 --> 00:14:21.519 it's not a solve problem from an engineering standpoint. 282 00:14:21.639 --> 00:14:24.519 No, but the blueprint is clearly there. Yes it is, 283 00:14:24.720 --> 00:14:27.200 and it is a blueprint that works with numbers. The 284 00:14:27.240 --> 00:14:30.279 industry is already scaling toward now. It's really easy to 285 00:14:30.320 --> 00:14:35.399 hear about these engineering hurdles, vacuum chambers, decoherence, spinning pops 286 00:14:35.440 --> 00:14:37.879 falling over and think to yourself, Okay, well they still 287 00:14:37.879 --> 00:14:40.200 have some science to figure out. I don't need to 288 00:14:40.240 --> 00:14:43.360 worry about my bank account or my personal data until 289 00:14:43.639 --> 00:14:44.639 the twenty thirties. 290 00:14:44.679 --> 00:14:48.399 And that is a very dangerous illusion. Yeah, because humanity 291 00:14:48.480 --> 00:14:52.480 is currently facing a concept in the cybersecurity world known 292 00:14:52.519 --> 00:14:54.639 as harvest now decrypt later. 293 00:14:54.879 --> 00:14:57.879 This this gave me absolute chills when I fully understood it. 294 00:14:58.080 --> 00:15:00.080 Walk us through the harvest now concept. Because of this 295 00:15:00.120 --> 00:15:02.120 means the threat isn't actually waiting for us in the future. 296 00:15:02.120 --> 00:15:03.320 The threat is happening today. 297 00:15:03.639 --> 00:15:06.639 It is bad actors aren't waiting for Qday to start 298 00:15:06.639 --> 00:15:10.720 their attacks. State sponsored hackers and massive cyber criminal syndicates 299 00:15:11.000 --> 00:15:15.039 are actively intercepting and siphoning up massive troves of encrypted 300 00:15:15.120 --> 00:15:16.399 data right now. 301 00:15:16.279 --> 00:15:18.519 Even though they can't read it exactly. 302 00:15:18.840 --> 00:15:20.879 To them, it just looks like the scrambled green paint 303 00:15:20.879 --> 00:15:23.879 we talked about earlier. It's securely locked by RSA or ECC, 304 00:15:24.480 --> 00:15:27.600 but they do not care. They are stockpiling it. They 305 00:15:27.600 --> 00:15:32.000 are hoarding exabytes of your encrypted data on massive server farms. 306 00:15:31.919 --> 00:15:33.559 Just waiting for the machine to be built. 307 00:15:33.600 --> 00:15:37.840 Exactly when Q day arrives and that ten thousand cubit 308 00:15:37.919 --> 00:15:42.360 machine comes online, they will use Shore's algorithm as the prism. 309 00:15:42.679 --> 00:15:46.240 They will instantly retroactively decrypt everything they have been harvesting 310 00:15:46.279 --> 00:15:47.200 for the past decade. 311 00:15:47.279 --> 00:15:50.759 So essentially, our digital past isn't actually secure, not at all. 312 00:15:50.799 --> 00:15:52.720 It's just sitting in a digital freezer, waiting to be 313 00:15:52.759 --> 00:15:55.720 thought out on Q day. Something I send securely today 314 00:15:55.840 --> 00:15:58.000 under the assumption that it is locked forever, could be 315 00:15:58.039 --> 00:16:00.799 easily decrypted and used against me ten years from now. 316 00:16:01.000 --> 00:16:03.840 That is exactly what is at stake. And consider what 317 00:16:03.960 --> 00:16:07.240 is currently protected by these algorithms. We're not just talking 318 00:16:07.279 --> 00:16:10.879 about abstract military communications. We are talking about the foundation 319 00:16:10.960 --> 00:16:13.240 of the Internet. Yeah, every time you log into a 320 00:16:13.240 --> 00:16:18.080 portal V HTTPS, every secure VPN session you use to 321 00:16:18.159 --> 00:16:22.440 access your company's network remotely, your secure emails, your confidential 322 00:16:22.480 --> 00:16:24.360 electronic medical records. 323 00:16:24.000 --> 00:16:26.919 And the financial side is terrifying. We are talking about 324 00:16:26.960 --> 00:16:30.919 bank transfers, credit card transactions, the power grid, and let's 325 00:16:30.919 --> 00:16:34.320 not forget blockchain, all right, trillions of dollars en dormant 326 00:16:34.320 --> 00:16:37.679 crypto wallets. If a quantum computer can derive your private 327 00:16:37.759 --> 00:16:40.759 key just by looking at your public address, the entire 328 00:16:40.799 --> 00:16:43.240 foundation of cryptocurrency collapses overnight. 329 00:16:43.399 --> 00:16:44.840 It completely breaks the system. 330 00:16:44.960 --> 00:16:48.039 So what is the defense playbook here? Humanity isn't just 331 00:16:48.120 --> 00:16:50.519 sitting around waiting for the locks to break. There has 332 00:16:50.600 --> 00:16:52.279 to be a global response. 333 00:16:52.440 --> 00:16:56.639 There is, thankfully. The primary defense is a massive transition 334 00:16:57.000 --> 00:17:01.960 to something called post quantum cryptography. Yeah. The National Institute 335 00:17:02.000 --> 00:17:05.279 of Standards and Technology or NIST has already been working 336 00:17:05.319 --> 00:17:09.599 on this for years and has officially standardized several PQC algorithms. 337 00:17:09.720 --> 00:17:11.799 Yeah. I was looking at the names of these algorithms 338 00:17:11.839 --> 00:17:14.920 and they sound like spaceships from a sci fi movie. 339 00:17:14.960 --> 00:17:15.559 They really do. 340 00:17:15.839 --> 00:17:20.400 Crystal's Kiber dilithium falcon. But here's what I don't understand. 341 00:17:21.160 --> 00:17:24.160 If Shor's algorithm is so powerful that it acts as 342 00:17:24.160 --> 00:17:27.640 a magical prism to break encryption. Why can't it just 343 00:17:27.960 --> 00:17:31.039 break Crystal's Kiber two? Why do these new locks work? 344 00:17:31.200 --> 00:17:34.200 Because the new locks do not use the paint mixing math. 345 00:17:34.799 --> 00:17:38.319 Unlike RSA, which relies on the difficulty of factoring large numbers, 346 00:17:38.680 --> 00:17:43.160 these new PQC algorithms rely on entirely different branches of mathematics. Okay, 347 00:17:43.240 --> 00:17:46.640 like what specifically something called lattice based cryptography. 348 00:17:46.720 --> 00:17:47.839 How does a lattice work? 349 00:17:48.079 --> 00:17:51.079 Imagine a massive multi dimensional grid, a lattice made up 350 00:17:51.079 --> 00:17:53.599 of hundreds of thousands of dots. It is like being 351 00:17:53.680 --> 00:17:55.960 dropped into a thick fog in a five hundred dimensional 352 00:17:55.960 --> 00:17:59.319 space and you're asked to find one specific hidden point 353 00:17:59.359 --> 00:17:59.799 of origin. 354 00:18:00.079 --> 00:18:01.599 Sounds impossible, It pretty much is. 355 00:18:02.119 --> 00:18:05.559 Quantum computers are incredible at finding the hidden periods or 356 00:18:05.599 --> 00:18:08.960 patterns in numbers, which is why they destroy factoring, But 357 00:18:09.119 --> 00:18:12.920 finding a hidden point in a five hundred dimensional spatial fog, 358 00:18:13.720 --> 00:18:17.400 quantum computers have no cheat code for that. Oh wow, 359 00:18:17.480 --> 00:18:20.319 They're just as hopelessly lost in the lattice fog as 360 00:18:20.359 --> 00:18:21.240 a normal computer. 361 00:18:21.680 --> 00:18:24.519 Okay, so we have the new locks. The math is done, 362 00:18:24.799 --> 00:18:27.400 NISS says, here you go here are the quantum proof locks. 363 00:18:27.839 --> 00:18:30.960 Governments in the US, the EU, the UK, China, they're 364 00:18:31.000 --> 00:18:32.920 all pouring billions into this. 365 00:18:32.880 --> 00:18:34.359 Transition, huge amounts of money. 366 00:18:34.440 --> 00:18:37.680 Yes, Google even issued warnings in early twenty twenty six 367 00:18:37.799 --> 00:18:41.440 urging the entire tech industry to adopt crypto agility, meaning 368 00:18:41.640 --> 00:18:43.440 you know, design your system so you can swap out 369 00:18:43.480 --> 00:18:47.119 encryption algorithms quickly. But if we have the algorithms ready 370 00:18:47.119 --> 00:18:50.640 to go today, why do cybersecurity experts estimate the transition 371 00:18:50.759 --> 00:18:54.400 timeline will take five to fifteen years. If the math works, 372 00:18:54.480 --> 00:18:58.720 why can't massive organizations just push a global software update overnight? 373 00:18:58.839 --> 00:19:02.599 Because upgrading global cryptography is an absolute logistical nightmare. 374 00:19:02.640 --> 00:19:03.599 Why is it so hard? 375 00:19:04.000 --> 00:19:09.039 You have to think about legacy systems. Large enterprises, multinational banks, 376 00:19:09.039 --> 00:19:12.640 and governments do not just have one single encryption switch 377 00:19:12.720 --> 00:19:16.319 they can flip. They have millions of deeply embedded, sometimes 378 00:19:16.480 --> 00:19:22.000 entirely hidden, cryptographic instances across decades old mainframes, custom software, 379 00:19:22.200 --> 00:19:23.839 and third party vendor networks. 380 00:19:24.000 --> 00:19:26.359 So a bank might literally have to go into their 381 00:19:26.400 --> 00:19:30.039 digital basement, find code written twenty years ago by an 382 00:19:30.039 --> 00:19:32.799 engineer who doesn't even work there anymore, and try to 383 00:19:32.880 --> 00:19:35.680 change the lock on a system that is quietly securing 384 00:19:35.759 --> 00:19:36.799 a massive. 385 00:19:36.519 --> 00:19:41.279 Database precisely, and the new lattice based PQC algorithms aren't 386 00:19:41.319 --> 00:19:44.079 always a clean swap. What do you mean, well, cryptographic 387 00:19:44.119 --> 00:19:47.480 keys for these post quantum methods are physically much larger 388 00:19:47.480 --> 00:19:50.400 in terms of data size, They require more processing power 389 00:19:50.559 --> 00:19:53.480 and have much higher performance overheads. Ah. Think of a 390 00:19:53.559 --> 00:19:56.200 legacy system's data pipeline like a mail slot on a 391 00:19:56.200 --> 00:19:59.079 front door. The old RSA keys fit perfectly through the 392 00:19:59.079 --> 00:20:01.960 mail slot. The new Crystal's Kiber key is the size 393 00:20:01.960 --> 00:20:04.319 of a shipping container. Oh man, Yeah, if you try 394 00:20:04.319 --> 00:20:07.119 to force that massive new key through an old legacy system, 395 00:20:07.200 --> 00:20:08.519 the system will completely. 396 00:20:08.160 --> 00:20:12.000 Crash, which means they have to rebuild the door entirely exactly. 397 00:20:12.079 --> 00:20:15.599 So what does a company actually do right now? How 398 00:20:15.599 --> 00:20:19.160 do we survive this five to fifteen year gap while 399 00:20:19.200 --> 00:20:21.599 the hackers are actively harvesting our data? 400 00:20:22.039 --> 00:20:25.359 Experts are stressing a hybrid approach. You do not just 401 00:20:25.440 --> 00:20:29.680