How quantum computers beat supercomputers at real tasks

HostWe have been hearing for a long time that quantum computers are the next big thing, like some dream from a movie that's always ten years away. But lately, the people who build these machines say the wait is over and they're starting to win races against the biggest supercomputers on earth. How did we get from science fiction to these machines actually doing work that a normal computer just can't handle?

GuestIt helps to think of this as a very specific kind of win. A few years ago, a team at Google ran a test where they asked their quantum chip to solve a math problem about random numbers. It was a very strange, messy problem that doesn't really help you or me in our daily lives. But the quantum chip finished it in about three minutes. At the time, they said the world’s fastest supercomputer would need ten thousand years to finish that same task. Even if that number was a bit high, the point was clear. For that one specific job, the quantum machine wasn't just faster. It was playing a totally different game.

HostBut if the task is just a strange math trick that doesn't help anyone, does it really count? I mean, I could claim I'm the fastest in the world at a game I made up myself, but that doesn't mean I'm a better athlete.

GuestThat's a fair point, and it's exactly what the field has been fighting over. But in the last year or two, things changed. We moved from those toy math problems to things like physics and chemistry. IBM took one of their chips and used it to model how atoms in a magnetic material behave. This is a real problem that scientists care about. They found that their quantum machine could handle the complexity of all those tiny parts interacting at once, while a normal supercomputer started to choke. The reason is that nature itself is quantum. When you try to use a normal computer to model a molecule, you're trying to use a machine that speaks in zeros and ones to describe a world that doesn't work that way. As the molecule gets bigger, the normal computer has to work twice as hard for every single atom you add. Very quickly, you run out of memory and time.

HostSo a normal computer just hits a wall because it has to keep track of too many things at once?

GuestExactly. Think of a normal computer like a librarian who can only look at one book at a time. To find an answer, he has to check every page of every book in order. If you have a billion books, that takes a long time. A quantum computer is more like a librarian who can look at every page of every book in the library all at once. It uses the fact that tiny bits of matter can exist in many states at the same time. By doing that, it finds the answer by looking at the whole picture instead of checking every single path.

HostIf they're that much better, why aren't they everywhere yet? If I have a machine that can look at every book at once, why am I still using my old laptop to do my taxes?

GuestBecause these machines are incredibly fussy. They're like a high-end sports car that breaks down if a single piece of dust hits the engine. To work, they have to be kept at temperatures colder than outer space. If a tiny bit of heat or light or even a vibration from the street gets inside, the machine loses its train of thought. It makes a mistake, and the whole math problem falls apart. For a long time, we thought we might never be able to fix those errors. But just recently, researchers found a way to group many of these fussy parts together to act as one solid, reliable unit. They call these logical units. By doing this, they can spot an error and fix it while the math is still happening. This was the big wall we had to climb over.

HostSo it sounds like we're moving from machines that are just fast at tricks to machines that can actually stay focused long enough to be useful.

GuestThat's the big shift. We're seeing the first cases where these machines are used to find new shapes for medicines or better ways to make batteries. We're not at the point where they can crack every code or fix the world’s weather yet, but they're no longer just a dream. They're starting to do the heavy lifting in areas where our best supercomputers have been stuck for decades.

HostIt's wild to think that the race is already happening in these deep-freeze labs while we're all just waiting for the future to show up.

GuestThe real test now is seeing if we can build them big enough to solve the problems that are still way too complex even for these early winners.

HostThe librarian is finally starting to read the whole room at once, and it changes everything we thought we knew about the limits of a machine.