Today I’m delighted to announce Willow, our latest quantum chip. Willow has state-of-the-art performance across a number of metrics, enabling two major achievements.
↫ Hartmut Neven, Founder and Lead, Google Quantum AI
- The first is that Willow can reduce errors exponentially as we scale up using more qubits. This cracks a key challenge in quantum error correction that the field has pursued for almost 30 years.
- Second, Willow performed a standard benchmark computation in under five minutes that would take one of today’s fastest supercomputers 10 septillion (that is, 10^25) years — a number that vastly exceeds the age of the Universe.
The concensus seems to be that this is a major achievement and milestone in quantum computing, and that it’s come faster than everyone expected. This topic is obviously far more complicated than most people can handle, so we have to rely on the verdicts and opinions from independent experts to gain some sense of just how significant an announcement this really is. The paper’s published in Nature for those few of us possessing the right amount of skill and knowledge to disseminate this information.
>”Second, Willow performed a standard benchmark computation in under five minutes that would take one of today’s fastest supercomputers 10 septillion (that is, 10^25) years — a number that vastly exceeds the age of the Universe.”
Yeah, but can it run Crysis?
Also, we’ve seen many times where hardware is designed to beat the snot out of a specific benchmark and we get crazy benchmark numbers that seem impossible – but then we find that for any useful, productive computing the new device is rather poor.
Also also, this particular PR piece emanates from the people at Google AI, which I do not believe to be a trustworthy source of any information at all.
I gather the lack of commentary shows how jaded the general public has become with various claims about AI and Quantum processing, searching for investors the marketing people have grossly overhyped the capability and short term returns.
But this leads to a bigger problem, because while the GP capabilities of Quantum and AI are overhyped the task specific capabilities are very real. The gernal public might be jaded, but the megalomaniacs in the corporate world are buying into the task specific side of Quantum and AI big time, It’s going to be last man standing stuff with the winner taking it all, and we should all be very worried about it!
cpcf,
Interesting to mix quantum computing and AI together. These are quite different ideas and applications at least as things stand today. Who knows, maybe quantum computing could be the key to general AI in the future?
I wouldn’t say I’m jaded, just educated. I understand the capabilities of quantum, while neat wont directly affect my computing life for a while. Its cool that its progressing, it has a lot of challenges ahead. Its kind of like fusion energy production. They are making progress, but not there yet. I would bet on quantum computing taking off in some capacity before fusion reactors.