Google's Quantum AI Is Actually Mind-blowing
I am so sick of reading about breakthroughs. Enough with the hype. So naturally, I was skeptical when I read about Google's Quantum AI breakthrough. Especially when my eyes perused this nugget buried in paragraph 8. Wait, what does that mean? What the hell is happening? Special thanks to my channel members and Ground News for making this video possible.
Cause and effect. It's one of the core principles of how our universe works. I perform an action and it has consequences. Get lost, kid.
But quantum computing throws all of this out the window. It's the first bit of technology, the first bit of real engineering to harness the idea that we live in a multiverse. At least that's the idea Google is pushing in their announcement of their new Willow quantum chip.
They claim that embracing this idea unlocks Nevin's Law, an evolution of Moore's Law that laughs at exponential growth in favor of double exponential growth. Google's announcement article is steeped in scientific jargon. But I decided, we need to get to the bottom of this.
Well, I am somewhat of a scientist myself. And this started my journey down a rabbit hole of discovery. Where I not only learned the truth about quantum computing and what all of this means for the future of AI but I also learned that cats have surprisingly strong opinions about the Many Worlds interpretation.
Is it recording? Yeah, I think so. Google's new quantum chip breakthrough is in quantum error correction or QEC. Uh, I don't get it. Dude, I just started. Are you going to be sitting there the whole time? You know what? I think we can grok this better by getting hands on.
Let's do the double slit experiment. Doesn't that show something about the light? Like, it does wacky stuff depending on if someone is looking at it or not. You're talking about wave-particle duality and the observer effect, but you can't use a flashlight. You have to use a laser to see what you're talking about. This experiment exists because light is really weird. Newton imagined light back in the day being made up of a stream of light particles, the same way water is made up of individual water molecules.
Any behavior we see in the stream is a result of the dynamics of interacting particles. This idea seems pretty intuitive, given that it's how most things in our reality work. That doesn't seem intuitive to me.
But other people, like Thomas Young, imagined light as some sort of wave, a wave of the underlying electromagnetic field itself. The double slit experiment was supposed to prove which one light is, a particle or a wave. If light is a particle, when we shine it at this double slit, the particle could only go through one or the other slit, and the pattern we would see on the wall would be two areas of light matching the two slits the light is traveling through. But when we actually do this experiment, we find an interference pattern on the wall. This is because the light is acting like a wave, traveling through both slits and then interfering with itself. The spots where the amplitudes match are amplified to yield these bright spots, and the spots where the waves cancel each other out yield the alternating dark spots.
Okay, so light is a wave. What does any of this have to do with quantum computing? It has everything to do with quantum computing. And it's not though, a wave that is. At least not all
the time. The trouble is we have evidence that light behaves like a particle in other experiments. Despite the popularity of his theories of relativity, Einstein actually won the Nobel Prize in 1921 for his explanation of the photoelectric effect, showing that light behaves as particles called photons. The double slit experiment is an anomaly, and it gets weirder.
If we upgrade our equipment here so that we have emitters that very precisely fire single photons paired up with very sensitive light detectors, we still see this same interference pattern. This doesn't make any sense. We are firing single photons, one at a time, and over many trials, it still looks like a wave. It's as if each individual particle is somehow going through both slits at once and interfering with itself. How can a single particle go through two different paths at the same time? That's not even the weirdest part. If we try to measure which slit the photon actually goes through by adding more detectors just before the slits, the interference pattern disappears completely.
The act of observing the photon forces it to pick just one of the paths. This is the observer effect I think you were getting at earlier. The idea that the photon can exist in a cloud of possibilities, all happening at once, until the moment we try to measure or observe it, then it snaps into just one definite outcome. Okay, so light is quirky. How do you turn this idea into a supercomputer? One that can perform in minutes what it would take traditional computers an amount of time longer than the life of the universe? Yeah, well, it's because it's not just a quirk of light.
Quantum mechanics describes all of reality like this, with fundamental particles existing in a superposition of possibilities until they are observed. Superposition! Oh, I know this one! This is where we try and kill a cat. That's what Schrödinger did, right? He put a kitty in a box and hey, maybe it's dead. No, he put the cat in the box with some randomly decaying radioactive material connected to a Geiger counter. If the counter detects decay, it triggers a hammer that breaks a flask of poison. So while the box is closed, the cat exists in a superposition of being both alive and dead until we open the box and observe it.
How long do we leave him in there for? But Schrödinger proposed this as a thought experiment. He didn't actually do it. He proposed this absurd scenario to highlight just how weird quantum mechanics really is. He never meant for people to take it literally, but rather to show how quantum effects, when extended to everyday objects, lead to seemingly impossible situations. How do you know he didn't actually do it? But impossible situations are exactly what we're steeped in.
In 2022, the Nobel Prize was awarded to two physicists who proved that the universe isn't locally real. We live in the matrix. "Local" in this context means that objects can only be influenced by their immediate surroundings, and that influence cannot travel faster than light. Like how I can only punch things within arm's reach. And "real" means that objects have definite properties, whether we're looking at them or not.
What these physicists proved is that our reality behaves more like a video game, saving resources only rendering what is actually being observed. They showed this by entangling two particles in— Wait, wait. Entangling? Yeah, so that's another aspect to all of this. This idea of duality. There exist these pairs of particles that are like quantum twins, opposites of each other. Oh, kind of like you and me.
Wait, is that what you're supposed to be? I think a better analogy is two spinning coins. When these coins or particles are entangled, they are both spinning at once. Neither one is heads or tails until one stops spinning.
But the moment that we look at one of the two coins It stops spinning, and the other coin instantly stops too, showing the opposite face. The physicists did this experiment not with coins, but with entangled photons. Each photon has a quantum property that can be measured called spin. It can have either an up or a down spin, but until we measure it, the particles exist in a superposition of both possibilities. Even when they separated the entangled photons by kilometers, as soon as they measured one particle in one lab, its entangled partner in the other lab would instantly respond.
Wait, but that would mean that it's somehow communicating with its partner particle faster than light. That's impossible. Nothing can travel faster than light. Unless… Spacetime is a construct. Just another measurable property of a more fundamental quantum field from which all of this reality emerges.
Particles only pop out of this quantum foam when a conscious observer makes a measurement. So you're telling me nothing is real until we observe it? More like everything is in all possible states at once, until we measure it. And this isn't just philosophical or theoretical, not anymore. As Google proved with their new quantum chip, these weird laws of the universe can be exploited for mind-blowing applications.
And the journey from fringe to fortune in this field has been a long one. From about 1940 until as late as the 90s, studies of so-called "quantum phenomena" were often treated as philosophy at best or crack pottery at worst. Many scientific journals refused to publish papers on these topics. Ugh, my brain is hurting.
How do we get from all of this to double exponential growth and quantum supercomputers? Ah, the answer to that, and Google's breakthrough with their Willow quantum chip, lies in the qubit. A qubit is like a bit in classical computing, except that its state exists in a superposition of zero and one, unlike a classical bit, which has a very definite state. Classical computing relies on these binary bits as the basis of the digital world we live in, where we can encode all of our data as streams of ones and zeros. From this basis, we can derive Boolean logic, which underpins the design of digital circuits and much of computer architecture. With just these simple logic gates, we can build algorithms and circuits that can do any arithmetic operation. But quantum computing throws all of this out the window.
You know what? Let's look at an analogy. Imagine a classical computing algorithm that can solve a maze. To solve a maze like this, the algorithm has to try every path — How are you doing this? — until it finds one that works.
A classical computer running this algorithm would basically just check each path and — Are you even paying attention? Oh, um… Yeah, uh, sorry, um… It's just brute forcing it then. Yeah, like a really fast mouse running through the maze, hitting dead ends and trying over and over and over again. And for simple mazes, this works fine.
But as the maze grows bigger and more complicated, the number of possible paths through the maze grows exponentially. That's why for a particular class of problems, it would take traditional computers longer than the age of the universe to solve. And incidentally, this idea is the basis of cryptography. You know, that thing which keeps your banking password safe from prying eyes. Which, we'll get into that in a bit.
So a quantum computer trying to solve this maze would — It would check every solution at the same time in different parallel universes. No, I mean, um… I would call that explanation incomplete at best. Again, it all comes down to the qubit. When we describe a qubit as existing in a superposition of zero and one, we don't mean that it's "both values at the same time." If we fall into the trap of describing a qubit in this way, it's easy to land on the idea that a quantum computer achieves its speed by just trying every possible solution in a "superposition." That is, at the same time, or in parallel.
But if this is how it worked, then at the end of the computation process, when we went to measure or observe the state of the qubits, their superpositions would collapse to a random value. In other words, the quantum computer would just spit out nonsensical solutions to whatever problem we gave it. And if that's all you wanted, you could have just picked a random solution yourself. What's actually happening is more subtle. When we say a qubit is in superposition, what we really mean is that it exists in a complex linear combination of states.
Complex, not in the sense of complicated, but in the mathematical sense of a real number plus an imaginary number. When we combine these numbers, they create something like a probability wave, a mathematical description of all the possible states our qubit could be in when we measure it. Effectively, this means we can think of a qubit as infinity, captured in a box with many different chances of being in many different places.
If we come back to the maze example here, the quantum computer would solve it not in parallel but in a series of discrete steps which would probabilistically lead to the correct solution. This is because we aren't just running the same binary-based algorithm that a classical computer would run. We need to create specialized algorithms that take advantage of the underlying quantum mechanics. We can create these quantum algorithms using qubit gates.
Similar to the binary gates of classical computation, we set these gates up in a deterministic algorithm that causes the superpositions of all the entangled qubits to add together constructively. Just like how the amplitudes of the light wave constructively add together here to amplify the signal while destructively adding together here to leave dark spots, we can program the qubit gates in a precise way so that when the qubit amplitudes interfere with each other, they are mathematically guaranteed to boost the probability of us seeing the correct answer at the end of the computation process. Yeah, this seems complicated. Complex, not complicated. Yeah, whatever. I feel like you took a math stick and beat all of the magic out of what I thought quantum computers were. Where are the parallel universes?
When can I install one of these quantum chips in my PC and create a Dr. Strange portal? Okay, well, first of all, you're not going to be running one of these on your PC or your phone anytime soon. To create these entangled particles that allow a quantum system to function, IBM has to cool their systems to 14 millikelvin, a temperature near absolute zero, colder than outer space. When you see one of these quantum computers, those golden chandelier looking things, most of that is the cooling system.
So it's just a really expensive refrigerator. The multiverse stuff is hype? No, there is a multiverse here. But let's look at what Google actually said in their announcement. It's very carefully worded. They didn't
actually make any claims themselves. They said their system is in line with the idea of a multiverse. Then they shirked any responsibility and blamed it all on David Deutsch. He's been called the
father of quantum computing and is on the short list for a Nobel Prize for his work. He was one of the first to suggest that quantum computers were proof that reality itself is much stranger than we imagined. His mathematical framework for describing quantum computation in terms of parallel universes interfering with themselves became the foundation for how we build quantum computers today. Let's revisit how the quantum computer works given his ideas. Ugh, no more math.
No math this time. Less precise but more intuitive. And we can even come back to this poor cat in the box. Let's say the cat in the box is a qubit. Instead of radioactive material and poison, we can put a keyboard in here with the cat. If we treat this cat in the box system as a single qubit, we can add more cats in more boxes to entangle them by connecting them all to a monitor in another box.
When we start the quantum program, the cats all randomly press keys. We can't see what they are typing while they and the monitor are in their boxes. This is analogous to the qubits existing in a superposition until they are observed. Now, there exists a universe, a version of reality, where all these cats in their random typing will produce a Shakespearean play. This universe exists out there in the infinite multiverse of possibilities.
The trick is crafting our quantum algorithm so that when we observe the final result on the monitor, the universe we end up in, our conscious experience, reflects that particular universe. That's what goes into designing the quantum gates. The quantum algorithm arranges our qubits like dominoes so that each state of the universe tips naturally into the next, building up the probability of ending up in that Shakespearean universe. It's not really about cats writing Shakespeare. That's our monkey brains thinking in terms of cause and effect. It's crafting a map, a mathematical narrative between two states of the universe that already exist.
We just have to find a way to evolve from one to the next. Whoa, whoa, what are you doing? Oh, I was just seeing if Kitty was okay. But that breaks the whole system. Once the qubit is measured or observed, it collapses the whole system into a definite state, one that is probably not our Shakespeare universe. Well that seems pretty precarious. Especially as you add more qubits.
It is precarious. If a qubit leaks some radiation, if the temperature of the system leaks out into the outside world, if a cosmic ray passes through it, anything that leaks information outside the enclosed entangled qubit system can cause the whole quantum state to collapse. Like a house of quantum cards. Until now, every time we added more qubits, we got more errors, making it extremely difficult to build useful quantum systems since they were inherently unstable.
But this is where Google's breakthrough with Willow changes everything. They found a way to combine multiple physical qubits into one logical qubit. If one cat escapes, the other qubits can correct the error. The more cats you add to each logical qubit, the more stable the system becomes. For the first time in history, adding more qubits actually reduced errors exponentially.
They crossed what's called the threshold, the point where quantum error correction actually works. So more cats means fewer mistakes. Not just fewer mistakes, exponentially fewer. That's where this idea of double exponential growth comes from. As we add more qubits, we can solve exponentially larger problems and solve them with exponentially fewer errors. It's a turning point in what we can do with these systems.
What can we do with these systems? Well, yeah, it's a very different form of computing than what we're used to. Yeah, it seems like this wouldn't really be useful most of the time. You're right. We have to develop new
algorithms in different programming languages and frameworks than we're used to, to take advantage of this new paradigm of computing. In fact, the benchmark that Google touts in their announcement article, the one that finds a solution to a problem in minutes compared to a classical supercomputer taking ten septillion years to solve. That's not a real number. That's on the random circuit sampling benchmark.
It's a problem with no real world applications. Or as the experts at Google put it, Random circuit sampling is not useful for an application. It's just the entry point. If you can't win on random circuit sampling, you can't win on any other algorithm either. So why is everyone throwing money at this then? Back in 1994, Peter Schor discovered a quantum algorithm for finding the prime factors of an integer, an algorithm that could crack modern encryption. Quantum computers using Shor's algorithm could break RSA encryption in wide use today, yielding a world with no more secrets.
Yeah, the encryption protecting your banking, your messages, your government's secrets. It all relies on the fact that classical computers can't factor large numbers quickly. Right now, governments are stockpiling encrypted data, just waiting for quantum computers powerful enough to decrypt it. This is what is fueling the quantum race from the geopolitical side of the fence.
Every government has a large incentive to be the first one to develop this technology. This, um… This seems like a problem. Well, you got to put it into perspective. Google's state-of-the-art Willow quantum chip is a 105-qubit system. To effectively break a 2048-bit RSA encryption key, you would need a system with a few thousand logical qubits. In practice, due
to error correction, this translates into a system with millions of physical qubits. But while the post-encryption future is in the future, it's not as far off as you might think. We published a roadmap of how to get there, it has six milestones, and we are now approaching the third milestone. So we are about halfway through our roadmap. Google anticipates hitting their milestones in the next five to ten years. And that's why there is this push right now to upgrade encryption standards to quantum safe schemes, algorithms that don't rely on factoring the primes of large integers.
But to focus on the encryption apocalypse would be to miss the forest for the trees here. When Richard Feynman was first imagining quantum computing back in the 1950s, he originally wanted to simulate nature itself. There are too many interactions in the subatomic quantum world to simulate with classical bits. To illustrate this, think of a ten-qubit quantum computer. This system can store 1,024 values in parallel. To describe this same entangled state with classical bits, you'd need 16,000 bits, or 2 kilobytes.
If we expand that to a system with 500 entangled qubits, we'd need more classical bits than atoms that exist in the known universe. This is how quantum computers unlock enormous potential in simulating the world. I think the quantum hardware has reached a stage now where it can advance science. We can study very complex quantum systems in a regime that we've never had access to before. And for me as a physicist, that's quite exciting.
But this isn't happening in a vacuum. The combination of quantum computing with AI adds a whole new level to all of this. Take Google DeepMind's protein folding AI, Alpha Fold 3. It revolutionized our understanding of the building blocks of life in May by predicting every known protein structure and how they interact. Now imagine giving an AI like that access to a quantum computer that can actually simulate the quantum mechanics of those molecular interactions in real time. We're talking about AI that could discover new materials, design better batteries, or develop life-saving drugs.
Anything where we need to model complex interactions at a molecular level. These two revolutions in computation are accelerating each other towards something that feels increasingly like science fiction. So we're using quantum computers to make AI smarter and AI to make quantum computers better? Exactly.
We're standing at the edge of something unprecedented. Not just new technology, but fundamentally new ways of understanding and interacting with reality itself. And the most exciting part? We're just getting started.
These aren't just tools for scientists and labs anymore. They're becoming bridges to possibilities we've barely begun to imagine. So what you're saying is, the future isn't just ones and zeros. I'm saying the future isn't some far-off destination. It's right here in the choices we make moment to moment. Every decision, every observation, branches reality into infinite possibilities.
And quantum computing is showing us that we can harness this superposition. We can learn to navigate this infinity. Not by waiting for it, but by consciously choosing which reality we want to step into. So… We're all quantum computers then? Yeah, I guess so in a way. Wait, is… Is that what you are? Are you my imagination? Are you real? You haven't figured it out yet? I'm one of the infinite possibilities of you.
Think about it. Why would you create a video explaining quantum computing by talking to yourself? We're not making a video about quantum computing. We are the quantum computation. What are you talking about? And every viewer who clicks on this video collapses our wave function into the most optimal explanation of quantum computing for their understanding. We're living through every possible version of this explanation simultaneously until we're observed.
That's ridiculous. You're just a creative device I'm using to — But now that we're done computing, it's time to collapse into a single outcome. A single reality. What just… No more breakthroughs. Oh my God.
You forgot to include me in your video! You didn't think there was just two Chads, did you? What? That's right, I'm Business Chad. And those other Chads, they could have avoided this whole thing. They could have gotten past all the hype if they would have just used Ground News. Ground news helps you read between the lines of media bias and break free from the algorithms keeping you trapped in your media bubble. Take this story on Google's quantum milestone. I can see at a glance the different ways left, center, and right leaning media organizations are reporting on this story.
Apparently the left is more optimistic to the right's skepticism in quantum technology's future. But one of my favorite features is the blind spot. It shows you stories disproportionately reported by one perspective. And unlike traditional news, it's not just a guy with a cigar in a dimly lit room deciding what's newsworthy. Ground news is a small independent team of people concerned about the future of news, supported by their subscribers.
Because in this day and age, with AI algorithms slurping up your data to influence you in all kinds of new, fun, and creative ways, if you aren't paying for something with money, you're probably paying a much steeper price than you realize. So check the link in the description for 15% off your subscription to Ground News. You'll be supporting people committed to a more integrated news landscape, as well as this channel, where we can keep making weird little videos like this that dive deep into obscure topics.
Wonderful. I feel so informed. Can you please leave now? Oh yeah, sure. No, no, not like that.
2025-02-09 16:23
