Sabine
patreon
New approach to quantum computing could soon beat IBM
🕑 Added 2024-03-03 16:00:05 +0000 UTC[This is a transcript with links to references.]
Last week we had bad news for quantum computing, this week we have the most excellent good news. Thanks for living with me through the ups and downs. The news is that physicists have reached a new record for building a quantum computer out of single atoms. They’ve now made it to more than 1000 atoms and they say their method can be scaled up quickly. Let’s have a look.
The story of quantum computing is a big drama. The major reason is that it’s incredibly difficult to judge the promise of this new technology.
In the past decade there have been hundreds of startups who tried their hands on getting quantum computing to work. The biggest players on the market have so far been Google, IBM, Honeywell and IonQ. They all bet on specific types of quantum bits, qubits for short. These qubits are the physical basis of the quantum computer, and the so far most widely used ones, which the big players have used, are superconducting circuits and ion traps.
The new paper is about an entirely different approach to create qubits called atoms in tweezers. The tweezers themselves are actually not made of atoms, they’re made of light. They basically trap atoms with lasers and then poke them with more lasers. Yes, most of current physics is basically poking things with lasers.Physicists really really like lasers.
Atoms in tweezers have been catching up very quickly with the approaches pursued by Google and IBM. While superconducting circuits and ion traps have a big starting advantage in being fairly easy to produce and operate, they’re difficult to scale up. Scaling up quantum computers doesn’t just mean you lump together more qubits. This isn’t all that difficult. The difficulty is that you still have to be able to control and read out all the qubits individually.
The nice thing about working with individual atoms is that they’re small and electrically neutral, so they don’t influence each other too much. You can encode information in the atoms for example by using spin states, this is how you get the quantum bits. But one of the challenges of putting atoms into tweezers has been to get sufficiently many of them into the same trap.
In the new work they solved this problem in a clever way by using a microlens array. That’s basically what the name says, it’s a tiny plate with even tinier lenses on it, the entire thing is less than a square millimeter. You shine with the laser on it and that creates a regular array of dots.
They use two of these arrays, shine two lasers on them, and then overlay them, so that the second layer pushes the atoms into the traps of the first. They call this “superchargning”. And look how lovely the atoms sit there in the array. Isn’t this amazing. These are images taken under fluorescent light, each dot is an atom. They managed to trap more than 1000 atoms in this array with regular clusters exceeding 400.
Now let me be clear that they didn’t actually calculate anything with these atoms, but in principle they know how to do it. Yes, it’s more lasers. If they manage to calculate with these atoms, then in terms of quantum bits, qubits for short, that’d be competitive with IBM’s biggest chip which just breached the 1000 qubits threshold a few months ago.
But wait this isn’t it with the good news. They say that with these microlens arrays they can in principle scale the setup to one hundred thousand atoms. That still falls somewhat short of the 1 million or so where the first commercially relevant applications for quantum computing could begin, but that’d be getting really close.
So, yes, the hype around quantum computers is thick, with some people claiming that they will bring a new industrial revolution and change the world. But the reason quantum computing news is such a roller coaster is that it isn’t all hype, there’s real promise there.
First and most importantly, quantum computers don’t rely on any speculative new theory, they’re just normal quantum physics. But what’s normal for quantum physics isn’t normal for us. Because quantum things have a very strong form of correlation, that’s the entanglement. You can’t do this on a conventional computer. And this entanglement is why a conventional computer can do some calculations faster than on conventional computers. Or rather, I should say, they don’t get slowed down that much when problems become larger.
You see, suppose you want to figure out how to optimize your financial assets. What to buy what to sell? Well, the more stocks you take into account and the more details you want to be in your model, the more difficult the calculation gets and the longer it takes. This is the case both for a conventional and a quantum computer. But on the conventional computer, the computation slows down much more. So for big enough problems, the quantum computer will win eventually.
And for banks that means money, a lot of money. This is why, the head of Bank of America said that quantum computing would be “bigger than fire”. It’d be a computational advantage that converts right into money.
But the issue is that this cross-over point where a quantum computer outperforms a conventional computers doesn’t just depend on the fundamental laws of physics. It depends on how good your conventional computer is and how good its algorithm is, and on the other side it also depends on the algorithm on the quantum computer, and how fast you can operate the quantum-bits. Also depends on how good your press department is in claiming you’ve done something interesting.
And this is why there is so much argument about whether quantum computing is really the game changer it’s been made out to be. Because artificial intelligence has much improved how quickly even conventional computers can solve some problems, and that’s an increasingly large competition to quantum computers.
Still, if a quantum computer would just be large enough it would eventually outperform the conventional computer, because that’s a question of this scaling law and that is indeed due to fundamental physics. There’s no doubt about it. No, you can’t argue with the laws of nature. So today I’m super optimistic about the prospects of atoms in tweezers, but maybe next week I’ll change my mind again, so don’t forget to subscribe.
