[00:00] A group of physicists just found a way to overcome&nbsp; one of the biggest limitations of quantum physics.&nbsp;&nbsp; So big it's a theorem. The no-cloning theorem.&nbsp; This means that contrary to what we thought,&nbsp;&nbsp;
[00:14] quantum information can be copied. That's&nbsp; quite something. Have physicists just been&nbsp;&nbsp; wrong for half a century? Will this finally make&nbsp; quantum computers work? Let's have a look. &nbsp;&nbsp;
[00:26] When I can't sleep at night, and that's happened a lot&nbsp; recently, I wonder whether Captain Kirk dies when&nbsp;&nbsp; he goes through the teleporter. One way to think&nbsp; about what the teleporter does is to convert Kirk&nbsp;&nbsp;
[00:38] into pure information, send this information&nbsp; elsewhere, and reassemble him. The other way&nbsp;&nbsp; is that the teleporter just reads the information,&nbsp; destroys the original, and then rebuilds the copy&nbsp;&nbsp;
[00:50] elsewhere. Except that quantum physics has what's&nbsp; called the no-cloning theorem that says you can't&nbsp;&nbsp; copy a quantum state without destroying the&nbsp; original. And since everything is ultimately&nbsp;&nbsp;
[01:04] quantum information, including you and I, doesn't&nbsp; this mean that there is only ever one real kirk?&nbsp;&nbsp; And as a corollary, you can't back up yourself&nbsp; onto a computer, which is unfortunate because I'd&nbsp;&nbsp;
[01:19] like a restore point before I read the comment&nbsp; section. In case that didn't already give you&nbsp;&nbsp; a headache, a group of physicists just reported&nbsp; they found a way around the no-cloning theorem.&nbsp;&nbsp;
[01:31] It's probably the most basic theorem of quantum&nbsp; physics, but it's one of the reasons why it's so&nbsp;&nbsp; hard to make quantum computers work. It says that&nbsp; you can't duplicate quantum states. And this means&nbsp;&nbsp; that the most obvious way to prevent errors, just&nbsp; make several copies and do the same calculation on&nbsp;&nbsp;
[01:49] all of them doesn't work on a quantum computer.&nbsp; You have to do something more difficult. The no&nbsp;&nbsp; cloning theorem was first proved in the early&nbsp; 1980s. So it's somewhat of a latecomer in the&nbsp;&nbsp; history of quantum physics. It's fairly easy&nbsp; to understand. Really you only need to know&nbsp;&nbsp;
[02:05] that in quantum physics we describe everything by&nbsp; a wave function usually denoted Psi. But if we have&nbsp;&nbsp; multiple wave functions that might be phi or ksi&nbsp; or some other weird Greek symbols. If you want to&nbsp;&nbsp;
[02:19] know something like what's the probability that&nbsp; a particle with wave function psi actually behaves&nbsp;&nbsp; like some other wave function phi then you take&nbsp; the square of the product of these wave functions.&nbsp;&nbsp;
[02:33] Okay, that sounds a little mysterious. What&nbsp; does it mean that a wave function behaves like&nbsp;&nbsp; some other wave function? Well, you might ask,&nbsp; for example, if I have a particle with a wave&nbsp;&nbsp; function that's smeared out all over the place,&nbsp; what's the probability that it behaves as if it&nbsp;&nbsp;
[02:51] was only over here? You do this by taking&nbsp; the product of the wave functions and then&nbsp;&nbsp; taking the absolute square. But in particular,&nbsp; the probability that psi behaves like itself is&nbsp;&nbsp;
[03:05] one. So the absolute square of any wave function&nbsp; is one. If you wanted to clone a wave function,&nbsp;&nbsp; you'd need a sort of cloning apparatus into which&nbsp; you shove a wave function and an empty slot that&nbsp;&nbsp;
[03:20] I'll call zero. And out comes the wave function.&nbsp; And the previously empty slot is now the same wave&nbsp;&nbsp; function. So you have duplicated it. For such a&nbsp; cloning machine to be possible in quantum physics,&nbsp;&nbsp;
[03:34] this operation must preserve probabilities. But&nbsp; you see this immediately creates a problem because&nbsp;&nbsp; suppose you shove a second wave function&nbsp; into the cloning machine. phi and zero goes&nbsp;&nbsp;
[03:48] in and out comes phi phi. Now this operation must&nbsp; preserve all probabilities in particular that&nbsp;&nbsp; of the psi 0 to appear like phi 0 that must&nbsp; still be the same after the cloning but 0 0 is&nbsp;&nbsp;
[04:05] just 1. So this means that phi psi is equal to phi psi &nbsp; square and this just is not the case for most&nbsp;&nbsp; states. This means the copy machine can't exist.&nbsp; The authors of the new paper now say that there&nbsp;&nbsp;
[04:19] is a clever workaround for this. They show&nbsp; that contrary to what we thought all along&nbsp;&nbsp; that one can make perfect copies of an unknown&nbsp; quantum state of say a quantum bit, a qubit,&nbsp;&nbsp;
[04:32] one just has to make sure that one can only ever&nbsp; read out one of the copies. And this isn't just&nbsp;&nbsp; maths. They actually showed that this works on&nbsp; an IBM quantum computer with about 150 qubits.&nbsp;&nbsp;
[04:46] They showed that indeed it works despite the&nbsp; hardware noise. So let me be clear. It's not that&nbsp;&nbsp; they found a mistake in the no cloning theorem.&nbsp; Rather, they demonstrated both mathematically and&nbsp;&nbsp;
[05:00] experimentally that it isn't as restrictive as we&nbsp; thought it is. I give this paper a 0 out of 10&nbsp;&nbsp; on the [&nbsp;__&nbsp;] meter. They ticked all the boxes.&nbsp; Good maths, good experiment, good interpretation.&nbsp;&nbsp;
[05:14] Sabine approves. What does this mean? First, it&nbsp; means we have to rethink what we thought we knew&nbsp;&nbsp; about quantum information. Second, it might&nbsp; have practical uses. For one thing, it might&nbsp;&nbsp;
[05:26] lead to better quantum computing algorithms. So,&nbsp; maybe we'll get some use out of them sooner than&nbsp;&nbsp; we expected. But it might also come in handy for&nbsp; future quantum internet. There is a deeper lesson&nbsp;&nbsp;
[05:39] in this that I also learned from my tax advisor:&nbsp; If you follow the rules precisely enough, you can&nbsp;&nbsp; do the thing you were told you can't do. I have&nbsp; two children in school and I have opinions about&nbsp;&nbsp;
[05:53] their maths and science education. In my opinion,&nbsp; it could be better. Luckily, today's sponsor,&nbsp;&nbsp; Brilliant, helps my kids to get ready for the&nbsp; future and they can help yours too. Brilliant is&nbsp;&nbsp;
[06:06] an online learning platform with a large number&nbsp; of courses on mathematics, science, and coding.&nbsp;&nbsp; They just recently introduced a new AI tutor&nbsp; that'll walk children through the problem sets,&nbsp;&nbsp;
[06:18] not by giving away the answers, but by helping&nbsp; them when they're stuck, reminding them of what&nbsp;&nbsp; they already know and explaining each step. I&nbsp; really think this is the future of education. My&nbsp;&nbsp;
[06:30] children especially like it that Brilliant doesn't&nbsp; tie them to a specific topic, but that they're&nbsp;&nbsp; free to explore whatever piques their curiosity.&nbsp; And of course, I have a special offer. You can get&nbsp;&nbsp;
[06:42] started with Brilliant's tutor for free and if you&nbsp; use my link brilliant.org/Sabine/ or scan the QR&nbsp;&nbsp; code, you can save 20% on an annual subscription.&nbsp; Thanks for watching. See you tomorrow.