Recently, prominent physicists were asked whether a sufficiently powerful computer could come up with a Theory of Everything, by the sheer power of crunching numbers. As a recent New York Times article by Dennis Overbye shows, physicists were divided and uncertain: “It might be possible, physicists say, but not anytime soon. And there’s no guarantee that we humans will understand the result.”

But doubt, in the view of multiverse theorist Max Tegmark, means we are guilty of “carbon chauvinism”—the idea that humans could be smarter than computers. The late Stephen Hawking thought that computers would replace humans and was alarmed by the prospect. According to Overbye, Hawking had been warning that computers would start to replace physicists in particular since 1980. Advocates believe they have now found a tool for the job:

Their tool in this endeavor is a brand of artificial intelligence known as neural networking. Unlike so-called expert systems like IBM’s Watson, which are loaded with human and scientific knowledge, neural networks are designed to learn as they go, similarly to the way human brains do. By analyzing vast amounts of data for hidden patterns, they swiftly learn to distinguish dogs from cats, recognize faces, replicate human speech, flag financial misbehavior and more.

Dennis Overbye, “Can a Computer Devise a Theory of Everything?” at New York Times (November 23, 2020)

But now, does this add up to a Theory of Everything?

For now, he conceded, there are limits to what can be achieved by the algorithm’s recursive method of problem solving, a practice known as regression. Although the machine can retrieve from a pile of data the fundamental laws of physics, it cannot yet come up with the deep principles — like quantum uncertainty in quantum mechanics, or relativity — that underlie those formulae.

“By the times that A.I. comes back and tells you that, then we have reached artificial general intelligence, and you should be very scared or very excited, depending on your point of view,” Dr. Tegmark said. “The reason I’m working on this, honestly, is because what I find most menacing is, if we build super-powerful A.I. and have no clue how it works — right?”

Dennis Overbye, “Can a Computer Devise a Theory of Everything?” at New York Times (November 23, 2020)

Overbye took the trouble to ask other cosmologists and got more measured responses:

Steven Weinberg, a Nobel laureate and a professor at the University of Texas at Austin, called it “a troubling thought” that humans might not be smart enough to understand the final Theory of Everything. “But I suspect in that case,” he wrote in an email, “we will also not be smart enough to design a computer that can find a final theory.” …

Nima Arkani-Hamed, a theorist at the Institute for Advanced Study in Princeton, N.J., took issue with the idea that the computer would discover something too deep for humans to comprehend: “This does not reflect what we see in the character of the laws of nature, which we have come to see over the centuries are based on fewer, deeper, simpler if more abstract, mathematical ideas.”

Dennis Overbye, “Can a Computer Devise a Theory of Everything?” at New York Times (November 23, 2020)

But the critical question is, can computers really come up with better ideas, as opposed to calculating much more quickly? Eric Holloway isn’t so sure. He told Mind Matters News:

“The idea sounds promising because what physicists do looks very similar to what computers do. Particle physicists run experiments to pump out massive amounts of data and then look at charts to find patterns and extract equations. We can do the same thing with a computer. We can plot data points and have the computer find the best fit line, which gives us an equation. And even better, a computer can look at an enormous number of comparisons.

“Scientists have demonstrated some progress in this direction. They fed data for two fundamental particles into a neural network, and the network was able distinguish the two particles. With just the right algorithm, data and enough computer horse power, then AI may just, as the late Stephen Hawking predicts, put human physicists out of a job.

“However, there is a problem with this approach is known as the No Free Lunch theorem (NFLT). It basically says there is no best learning algorithm. In fact all algorithms are exactly identical in performance, when averaged across all possible problems.

“That means, in general, the scientists’ fancy neural network is just as good as lottery players picking numbers based on how many birds fly past their window. An astounding claim, but rigorously proven by physicists David Wolpert and William Macready.

“So, why does machine learning work? Because the human engineer brings problem domain knowledge to pick just the right algorithm for the problem. We cannot just run all possible algorithms against some problem set and hope for success.

“Coming back to the question at hand, let’s suppose scientists do discover the algorithm that cracks the fundamental theory of everything. To what or whom do we credit the discovery? Is it the scientists or the AI? Since the NFLT proves AI cannot in general solve problems without human guidance, then we must ultimately attribute the breakthrough to the scientists and engineers who created the algorithm in the first place.

“So, the possibility of an AI cracking the fundamental theory of everything is not out of the question but the proper attribution of the cracking goes to the humans involved. Contra the late Stephen Hawking, AI is not the “event horizon” for physicists. That’s precisely because AI in fact increases the potential for physicists to make new discoveries.”

In short, no “fix” will cause the computer to do ALL our thinking for us.

To some readers, the quest for the computer that answers our biggest questions may recall the “Ultimate Question” in Hitchhiker’s Guide to the Galaxy by Douglas Adams (1952–2001), a radio series, novel series, TV series, and film wherein, to hear Wikipedia tell it, things did not go smoothly:

When asked to produce the Ultimate Question, Deep Thought says that it cannot; however, it can help to design an even more powerful computer that can. This new computer will incorporate living beings into the “computational matrix” and will run for ten million years. The computer is revealed as being the planet Earth, with its pan-dimensional creators assuming the form of white lab mice to observe its running. The process is hindered after eight million years by the unexpected arrival on Earth of the Golgafrinchans, and is then ruined completely, five minutes prior to completion, when the Earth is destroyed by the Vogons to supposedly make way for a new hyperspace bypass. In The Restaurant at the End of the Universe, this reason is revealed to have been a ruse: the Vogons had been hired to destroy the Earth by a consortium of psychiatrists, led by Gag Halfrunt, who feared for the loss of their careers when the Ultimate Question became known.”

Ultimately, the question and the answer are recovered: “Six by nine. Forty two. That’s it. That’s all there is. I always thought something was fundamentally wrong with the universe.” We are given to understand that the question was “miscomputed”: 6 × 9 is actually 54 but the tumult on Earth has introduced errors into the computer. The number 42 and the phrase “life, the universe, and everything” later became part of popular culture.

Which raises the question, if cosmologists did discover a Theory of Everything, computer-aided or not, would they accept it if it made no sense or wasn’t to their liking? Or would they keep looking for a different answer from a yet bigger computer?

You may also enjoy:

No materialist theory of consciousness is plausible. All such theories either deny the very thing they are trying to explain, result in absurd scenarios, or end up requiring an immaterial intervention
(Eric Holloway)

and

A materialist gives up on determinism. Evolutionary biologist Jerry Coyne undercuts his own argument against free will by admitting that quantum phenomena are real.(Michael Egnor)