So… did they find anything where our current quantum computing is actually faster? Last time I read about this, it boiled down to: If you spend as much time on optimizing the regular calculation, then that is as fast/faster. So no actual, fundamental, benefit to our current state of quantum computing.
In theory, quantum computing should be faster once hardware that’s faster is available, and only if the problem you’re trying to solve is in BQP, which isn’t that much of what computers are used for. Progress has been slow, but continuous, so the gap between simulating a quantum computer and actually using one has been shrinking. In October last year, Google’s Willow chip was verified to have achieved quantum advantage, i.e. done something that could be checked externally faster than a classical computer could have. It was only 13,000x faster, and in one specific task, which isn’t really enough to change the world, but ten or twenty years ago it was still thought to be fairly plausible that the physics might not be right and even if the practical problems were solved, they still wouldn’t work.
Even if quantum computers get ludicrously fast, they’re still not going to be especially common, and they’ll be a piece of specialised equipment, more like an electron microscope than a home PC. Most people just don’t need to do any stuff that’s in BQP, so don’t care if they can do it faster. If you’re a company, university or government body that needs to do one of the very specific things that will be faster, though, they’ll be indispensable.
Edit: Of particular relevance to the article, at the moment, SHA256, the hashing algorithm underpinning Bitcoin, is considered to be quantum-resistant. Someone might discover some new maths that means a quantum computer can break it faster than a classical computer, but at the moment, even though people have looked into it, there’s no indication that it’s possible, so it should never become easier to break Bitcoin etc. with a quantum computer than a classical one.
It was only 13,000x faster, and in one specific task
Useless, contrived task.
at the moment, SHA256, the hashing algorithm underpinning Bitcoin, is considered to be quantum-resistant… so it should never become easier to break Bitcoin etc. with a quantum computer than a classical one.
It’s absolutely wild how grifters will hype articles like this without scientific basis.
Prime factorization is one. Another is an asymptotically faster Fourier transform.
There are several tricks in signal processing that have faster analogs in a quantum implementation. Those cool tricks are the foundation of much more complicated transformations and algorithms.
There is an entire complexity class you can read up in: Bounded Quantum Polynomial Time (BQP).
So… did they find anything where our current quantum computing is actually faster? Last time I read about this, it boiled down to: If you spend as much time on optimizing the regular calculation, then that is as fast/faster. So no actual, fundamental, benefit to our current state of quantum computing.
In theory, quantum computing should be faster once hardware that’s faster is available, and only if the problem you’re trying to solve is in BQP, which isn’t that much of what computers are used for. Progress has been slow, but continuous, so the gap between simulating a quantum computer and actually using one has been shrinking. In October last year, Google’s Willow chip was verified to have achieved quantum advantage, i.e. done something that could be checked externally faster than a classical computer could have. It was only 13,000x faster, and in one specific task, which isn’t really enough to change the world, but ten or twenty years ago it was still thought to be fairly plausible that the physics might not be right and even if the practical problems were solved, they still wouldn’t work.
Even if quantum computers get ludicrously fast, they’re still not going to be especially common, and they’ll be a piece of specialised equipment, more like an electron microscope than a home PC. Most people just don’t need to do any stuff that’s in BQP, so don’t care if they can do it faster. If you’re a company, university or government body that needs to do one of the very specific things that will be faster, though, they’ll be indispensable.
Edit: Of particular relevance to the article, at the moment, SHA256, the hashing algorithm underpinning Bitcoin, is considered to be quantum-resistant. Someone might discover some new maths that means a quantum computer can break it faster than a classical computer, but at the moment, even though people have looked into it, there’s no indication that it’s possible, so it should never become easier to break Bitcoin etc. with a quantum computer than a classical one.
Useless, contrived task.
It’s absolutely wild how grifters will hype articles like this without scientific basis.
But that is the thing: was it actually faster or did they just spend 10’000 hours optimizing that side and 0 hours optimizing the normal calculation?
Prime factorization is one. Another is an asymptotically faster Fourier transform.
There are several tricks in signal processing that have faster analogs in a quantum implementation. Those cool tricks are the foundation of much more complicated transformations and algorithms.
There is an entire complexity class you can read up in: Bounded Quantum Polynomial Time (BQP).