Google has officially laid claim to quantum supremacy. The quantum computer Sycamore reportedly performed a calculation that even the most powerful supercomputers available couldn’t reproduce.
In the latest infomercial out of Gizmodo, a new company called Caffe2 is teaming up with Google to develop, and test, a quantum computer. Sycamore is not super important because it is a tiny, nonexistent quantum computer based on the way atoms in the lab are stripped away, or left in their natural state, namely a crystal. (Nature calls it a “Neutrino,” one you have to stop and get a cold drink of.)
Today, Sycamore did a 1.3 exaflop calculation that is 10 times the power of the world’s best supercomputer. So powerful that it would trump its next closest theoretical rival, Berkeley’s “Maverick,” which reportedly maxes out at 7.5 exaflops. So powerful that even the world’s best supercomputers couldn’t do a calculation like that.
Google? Maybe. Either way, this is one gigantic bonafide victory for Caffe2. And if you take their statement at face value, that means that they figured something was wrong with Sycamore, right? In that case, they hit back hard. Namely, Caffe2 has a friend.
Bryan Walsh, a professor of quantum chemistry at Stanford University who specializes in the unification of particles in the nuclear binary state, has noticed that many of Caffe2’s calculations make no sense. He also notes that, due to the mysterious nature of their method of computation, Caffe2 doesn’t seem to have made their calculations for all atomic entities–including ions, as the name suggests.
“How exactly do they have such a low success rate?” Walsh asks, pointing out that a whopping 91% of the Sycamore calculations using the parameters of the Astrolabe ended up being incorrect.
Those mathematical equations are relatively simple in nature, but they can obscure parts of the order in which the quantum states, as in quarks, have to be computed. They are also hard to explain because of the nature of the chemical reactions and physics involved, “like a beer koozie!” Walsh says.
But, perhaps, if quantum things were found to be working efficiently (instead of extremely well) around the world, then maybe a computer could find a way to replicate the basic structure of particles and not get into messy, experimental messes.
With the ability to perform calculations more cheaply than any other supercomputer, and would be able to turn that into real-world problems, Caffe2 is perfect for scientists and engineers. And we may just be witnessing the birth of a holy new age of scientific literacy.