Oct 8th, 2019 - Category: Miscellaneous
We were just getting over the fear of AI (or were we?) and now headlines are full of Google’s claims of Quantum Supremacy. 
The Googlians must be so pleased with themselves and as a bonus, many articles include this photo of a device that even looks like something supremely sinister.
However, as with many headlines lately, the truth is much more mundane. According to Wikipedia, Quantum Supremacy just means that Google’s quantum computer “has the potential ability to solve problems that classical computers practically cannot.” Now Google has built one and specifically chose a trivial problem for it that is solvable in 3 minutes and 20 seconds but would take the world’s fastest traditional supercomputer around 10,000 years to finish. Even crazier were the facts that the paper was only briefly published to a NASA website before being removed, but luckily The Financial Times retrieved a copy just in time and published it. Nice way to announce a breakthrough Google!
The term Quantum Supremacy itself was coined in 2012 by John Preskill. He explains why and his misgivings about his word choice in this article in Quanta Magazine. He was certainly qualified to create it since he’s a professor of theoretical physics at the California Institute of Technology (CalTech) and his main research area is quantum information science. His article goes on to clearly explain what Google achieved, why it’s a breakthrough, but also why it is still just the first baby-steps towards a general purpose quantum computer. “This [Google’s] quantum computation has very little structure, which makes it harder for the classical computer to keep up, but also means that the answer is not very informative.”
However, the sheer volume of headlines that this discovery generated piqued my curiosity to learn more. Wikipedia’s trio of articles were a good start: “Quantum computing,” “Quantum algorithm,” and “Quantum supremacy.” Even better was the ExtremeTech article,”How Does Quantum Computing Work?” which starts with the basic concepts and adds what makes these concepts the “special sauce” of quantum computing. For example, qubits (quantum bits) are the basic units of computing, but their properties such as superposition and entanglement is what make them powerful. Superposition simply means that a qubit can have a range of values between zero and one that only collapses into a definite value when it is measured. This creates a kind of trinary “yes-no-maybe” system where traditional computers only can give binary “yes-no” answers. Entanglement is even more strange. When qubits are entangled, they affect each other instantly when measured no matter how far they are away. “Quantum Entanglement” is a property of quantum systems originally described by Einstein and others in 1935 but only experimentally verified this year. Einstein referred to it as “spooky action at a distance.” The article also explains why building a quantum computer is so incredibly difficult.
So if Google has achieved Quantum Supremacy by building their 53-qubit quantum computer why haven’t they built the ultimate AI that will take over the world? Maybe first and foremost programming these machines is even more bizarre than the theory behind it. Also, Google’s machine is not nearly the most powerful quantum computer. D-Wave has a quantum microprocessor with 2000-qubits but like everything else quantum, traditional concepts of power don’t necessarily translate. The ArsTechnica article, “D-Wave 2000Q hands-on: Steep learning curve for quantum computing” provides some of the gory details programmers face when trying to solve real-world problems even when equipped with the latest and greatest tools, in this case D-Wave’s “quantum optimizer.” In short, the author “found it quite a challenge to change the way I conceive of a physical problem so that I could even begin to write the code.”
Finally, getting back to that ultimate, humanity crushing, quantum computer based AI, Quanta Magazine comes to the rescue again with the article, “Job One for Quantum Computers: Boost Artificial Intelligence” which covers many of the major challenges in this field. Somewhere in all that quantum-ness there is a potential for application to AI. But I’m not fearful of a revolution, maybe just an evolution over the coming decades. There is one thing that is 100% sure, there will be more sensationalistic headlines featuring this topic!
P.S. If you’re interested in why I think a human level AI (quantum or not) is improbable in the near future, you might want to read this amazing explanation of the complexity of the human brain. It’s very understandable and a fun read as well. Each one of us has around 100 billion neurons in our brain that make-up our unthinkably vast neural networks and yet we feed it potato chips and program it with cat videos!
