In 2019, a Google-led quantum computing team used a 53-qubit processor to perform a sampling task, sampling from a probability distribution that a quantum computer produces, in a demonstration described as 'quantum supremacy': solving a problem no classical computer could feasibly solve. The task was intentionally designed to showcase this gap rather than to solve a practical problem.
- 01
IBM responded to the 2019 result by arguing a classical algorithm could reproduce the task in about two days instead of seconds, disputing the supremacy claim.
- 02
The hosts note a later chip with 100 or more qubits pushed the classical comparison to millions of years, making the classical catch-up argument much harder to sustain.
377 words · auto-generated from the episode video
2:11:16absolutely incredible. We literally are looking at a black hole. >> Unbelievable. >> Yeah. This this one when science breaks through in the mainstream this one definitely >> I mean this one was insane. Everybody was like that's crazy. >> That's crazy. Uh that's crazy. We are still in 2019 and I remember talking to you about this experimental demonstration of quotequantum supremacy. >> Yeah, I think this is a pretty big deal. This was John Martinez and Google quantum AI team at UC Santa Barbara. They achieved quantum supremacy with their 53 cubit Syncore processor and they solved a complex sampling task. They were basically trying to sample from a probability distribution that a quantum computer would produce. kind of
2:11:57a roundabout way of declaring quantum supremacy, which is this idea of solving a problem that no classical computer can solve. Right after this paper was published, IBM came out and said that actually there is a classical algorithm that'll do it. Um, it'll take you 2 days instead of like the seconds that it took you, but it'll still do it. And these guys weren't that worried because this is a 53 cubit Syncore processor. They figured I mean that classical algorithm again, it's not going to be able to scale. They recently released I think a 100 plus or 200 cubit um algorithm and that one definitively a classical computer couldn't do for millions of years. Um it's it's a roundabout
2:12:38problem. It's kind of a problem that's made to show quantum supremacy but I think it shows just how far >> um technology has gone >> to create actually like a workable quantum computer that can that can take from a quantum distribution. Right. That's something that a classical computer can't do. Um, I thought it was still a really cool paper, >> meaningful. I remember we were talking, I texted you immediately after I saw the headline, like what does this all mean and where do we go from here? We are almost at the conclusion of our 250ear list. Thank you all for those who are still with us. Two and a half hours in 2021 launch of the James Web Space
2:13:20Telescope. >> This is huge. The James Web Space