Google AI is Rewriting History

Chinese Deflate Quantum Hype Again

Sabine Hossenfelder is now doing short daily physics news videos, and her latest is on Bad News for Quantum Computing: Another Advantage Gone.

In short, quantum computing researchers have been claiming quantum supremacy for years. Some call it quantum advantage. However, there has never been any convincing demonstration that quantum computers have any speedup at all over conventional computers.

The latest is that IBM claimed last year to do a quantum calculation on a "noisy" quantum computer. Some thought that they had outdone Google. But a Chinese group outdid them by doing the calculation faster and better on a classical computer.

The quantum enthusiasts will argue, as usual, that this does not disprove quantum computing, and maybe a more clever experiment would show an advantage. I am waiting.

The physicists philosophy of physics

Princeton astrophysicist PJE Peebles writes:

The starting idea of the natural sciences is that the world operates by rules that can be discovered by observations on scales large or small, depending on what interests you. In fundamental physics, the subject of this essay, the idea is narrowed to four starting assumptions.

A: The world operates by rules and the logic of their application that can be discovered, in successive approximations.

B: A useful approximation to the rules and logic, a theory, yields reliably computed quantitative predictions that agree with reliable and repeatable mea- surements, within the uncertainties of the predictions and measurements.

C: Fundamental physical science is growing more complete by advances in the quantity, variety, and precision of empirical fits to predictions, and by occa- sional unifications that demote well-tested fundamental physical theories to useful approximations to still better theories.

D: Research in fundamental physical science is advancing toward a unique mind-independent reality.

These sound reasonable, but they leave no room for many-worlds theory, string theory, simulation hypothesis, superdeterminism, or many of the ideas that are now fashionables.

The essay gives way too much attention to philosopher Thomas Kuhn.

It quotes Einstein:

The supreme task of the physicist is to arrive at those universal elemen- tary laws from which the cosmos can be built up by pure deduction.

This sounds a little like Weinberg's mythical Final Theory, also discussed.

No, trying to build the cosmos from pure deduction is foolishness.

Dissecting Einstein's Brain

The RadioLab podcast just rebroadcast this:

Albert Einstein asked that when he died, his body be cremated and his ashes be scattered in a secret location. He didn’t want his grave, or his body, becoming a shrine to his genius. When he passed away in the early morning hours of April, 18, 1955, his family knew his wishes. There was only one problem: the pathologist who did the autopsy had different plans.

In the third episode of “G”, Radiolab’s miniseries on intelligence, we go on one of the strangest scavenger hunts for genius the world has ever seen. We follow Einstein’s stolen brain from that Princeton autopsy table, to a cider box in Wichita, Kansas, to labs all across the country. And eventually, beyond the brain itself entirely. All the while wondering, where exactly is the genius of a man who changed the way we view the world?

Later in the show, it discussed theories for the origin of Einstein's most brillian idea -- special relativity. Besides his extra-smart brain, it mentioned his physicist wife and a philosopher. It even had professor Galison explaining how train schedules causes people to rethink time.

Okay, but there was no mention of Lorentz and Poincare, or the fact that they had published the entire theory ahead of Einstein.

Galison is unusual because he does not recite crazy stories about Einstein's originality, like other Einstein scholars. He read Lorentz and Poincare and obviously understands that they did it all first, but he refuses to comment on the priority dispute.

What's the difference, said Heisenberg

From a math site:

In the 1960s Friedrichs met Heisenberg and used the occasion to express to him the deep gratitude of mathematicians for having created quantum mechanics, which gave birth to the beautiful theory of operators on Hilbert space. Heisenberg allowed that this was so; Friedrichs then added that the mathematicians have, in some measure, returned the favor. Heisenberg looked noncommittal, so Friedrichs pointed out that it was a mathematician, von Neumann, who clarified the difference between a self-adjoint operator and one that is merely symmetric. "What's the difference," said Heisenberg.

- story from Peter Lax, Functional Analysis (slightly edited for length)

There is the difference between a physicist, and a mathematical physicist.

John von Neumann wrote a 1932 book on quantum mechanics, and turned it into a real theory.

To a physicist, an observable is a symmetric operator, because those are the ones that give real values, and only real values are observed. To von Neumann, an observable is a self-adjoint operator on a Hilbert space, where some additional technical requirements are needed in order to prove the spectral theorem.

I am not trying to say that Heisenberg was stupid. But it is striking that a world-famous physicist could get a Nobel Prize for using operators as observables, and still be oblivious to the formal mathematical definition found in textbooks. We cannot expect physicists to understand mathematical subtleties.

The World is not Discrete

Some people like to say that Quantum Mechanics makes the world discrete. That is not true. But I always assumed that QM models could be approximated by lattice models.

Apparently this is not true. We know that the weak force is chiral, ie, it violates mirror reflection symmetry. Neutrinos are left-handed in the Standard Model.

From the Scott Aaronson blog:

“There is currently no fully satisfactory way of evading the Nielsen-Ninomiya theorem. This means that there is no way to put the Standard Model on a lattice. On a practical level, this is not a particularly pressing problem. It is the weak sector of the Standard Model which is chiral, and here perturbative methods work perfectly well. In contrast, the strong coupling sector of QCD is a vector-like theory and this is where most effort on the lattice has gone. However, on a philosophical level, the lack of lattice regularisation is rather disturbing. People will bang on endlessly about whether or not we live “the matrix’”, seemingly unaware that there are serious obstacles to writing down a discrete version of the known laws of physics, obstacles which, to date, no one has overcome.”

There is a whole industry of physicists doing lattice approximations to the SM, but the SM is chiral and the approximations are not, so there is no hope that the approximations converge to the SM.

Aaronson is commenting on the silly idea that we live in a computer simulation. If we did, it would raise another silly idea that we could overwork the simulator by doing certain experiments.

Quantum Computer Revolution may be Further off

IEEE Spectrum reports:

The quantum computer revolution may be further off and more limited than many have been led to believe. That’s the message coming from a small but vocal set of prominent skeptics in and around the emerging quantum computing industry.

The problem isn’t just one of timescales. In May, Matthias Troyer, a technical fellow at Microsoft who leads the company’s quantum computing efforts, co-authored a paper in Communications of the ACM suggesting that the number of applications where quantum computers could provide a meaningful advantage was more limited than some might have you believe.

“We found out over the last 10 years that many things that people have proposed don’t work,” he says. “And then we found some very simple reasons for that.” ...

Even in the areas where quantum computers look most promising, the applications could be narrower than initially hoped. In recent years, papers from researchers at scientific software company Schrödinger and a multi-institutional team have suggested that only a limited number of problems in quantum chemistry are likely to benefit from quantum speedups. ...

“In the public, the quantum computer was portrayed as if it would enable something not currently achievable, which is inaccurate,” he says. “Primarily, it will accelerate existing processes rather than introducing a completely disruptive new application area. So we are evaluating a difference here.” ... “Most problems in quantum chemistry do not scale exponentially, and approximations are sufficient,” he says. “They are well behaved problems, you just need to make them faster with increased system size.”

Compare to the hype surrounding Artificial Intelligence (AI). It is also over-hyped by its enthusiasts, but it has also delievered a lot of very impressive demonstrable results. Quantum computing has delivered nothing, and may never deliver anything.

Someday we really will have personal robots and self-driving cars, but we may never have a useful quantum computer.

Google Research just released a video:

Quantum Computing - Hype vs. reality | Field Notes

Google Research
35.7K subscribers

25,188 views Jan 22, 2024 #GoogleAI #GoogleResearch
As the race to build the world's first truly useful quantum computer intensifies, so too does the need for clear-eyed assessment. This Field Notes episode brings in the Google Quantum AI team to help answer a few fundamental questions to drive understanding of its impact now and in the future.

It says quantum computers could become useful by 2030, or maybe a few years later.

the group is called "Quantum AI", but the video said nothing about AI. Just combining buzzwords, I guess.

The most touted application was fusion simulations, in order to help bring fusion power plants to market. Othere were discuvering drugs, and making the planet greener with chemistry for better batteries and fertilizer.

No mention of breaking everyone's cryptosystems. That is the only think quantum computer enthusiasts are sure about.

I am amazed that Google keeps funding this pipe dream. It has canceled hundreds of really useful products. It has developed some really good AI, but is checken to market it like OpenAI and Microsoft. Elsewhere Google touts quantum machine learning, but I doubt this will ever be practical. The non-quantum methods are progressing rapidly, and there is no sign that quantum computers would be useful.

Self-driving cares are over-hyyped, but I believe we are making progress and will get there. I do not think that that we are getting any closer to quantum computing.