The 1969 Book that Killed AI Research

The neural nets at the core of AI were invented in 1943, and popularized in 1957-60 as the Perceptron. It had the basic idea that was later developed into today's AI LLMs.

But not everyone recognized it. Prominent MIT professors wrote a 1969 book, Perceptrons: An Introduction to Computational Geometry, trashing the concept.

Reports by the New York Times and statements by Rosenblatt claimed that neural nets would soon be able to see images, beat humans at chess, and reproduce. ...

This book is the center of a long-standing controversy in the study of artificial intelligence. It is claimed that pessimistic predictions made by the authors were responsible for a change in the direction of research in AI, concentrating efforts on so-called "symbolic" systems, a line of research that petered out and contributed to the so-called AI winter of the 1980s, when AI's promise was not realized.

The book claimed to prove that perceptrons were hopelessly limited, as they could not even approximate a simple xor function.

The authors doubled down in 1988 updated edition, saying "little of significance [has] changed since 1969". At the time there was a revival of interest in neural nets, but the authors said they would fail to scale up.

We now know that those neural nets of the 1940s, 50s, and 80s do scale up, and perform spectacularly well. The bold predictions of the perceptron inventor must have been seen as wildly optimistic at the time, but they turned out to be correct.

I post this as an example of how smart people can be wrong about what is possible, and how brilliant research could be sidetracked.

Rosenblatt made a hardware computer perceptron in 1960, and that is now in the Smithsonian Museum. What no one knew until about 1912 was that video game chips would be the key to scaling up neural nets.

Here is a recent video that tries to relate these old ideas about neural nets to current research.

Schrödinger was an Indeterminist

Conventional wisdom is that classical physics is determinist, while quantum physics is not.

New paper:

Was Schrödinger ever a determinist?
Flavio Del Santo, Nicolas Gisin

Erwin Schrödinger is often portrayed as a reactionary who resisted the indeterminism introduced by quantum mechanics, but on closer inspection his views on determinism prove to be more complex and more radical.

Schrödinger argued, correctly in my opinion, that neither classical nor quantum physics is deterministic.

(i) The first argument challenges the mathematical idealization of determinacy in the initial conditions. ...

(ii) The second argument for indeterminism concerns discontinuity and irreversibility. ...

(iii) In his final argument, Schrödinger links quantum indeterminacy to the abandonment of the concept of trajectory and to the essential role of discontinuity. Remarkably, he can be seen here as anticipating the notion of “information” in physics (he does not use the term “information”, which was formalized only in 1948 by C. Shannon), suggesting that nature may be understood as a discrete sequence of yes–no answers, representable as strings of 0s and 1s.

None of this should be surprising or controversial, except for the notoriety of those with contrary views.

Einstein was a determinist.

The many-worlds advocates, like Sean M. Carroll, are determinists.

Sabine Hossenfelder is a superdeterminist.

The Bohm pilot wave theorists, like Tim Maudlin, are determinists.

It is philosophically possible to be a determinist, just like the rest of the existential crisis iceberg is possible, but silly. If you believe in free will, or in scientific and experimental choices, then determinism is wrong.

Here is a new comment with a typical modern scoff about free will, rejecting any suggestion that it could be possible:

OMG

“It is our free will that enables it to be so.”

1) your brain is nothing more than a collection of atoms

2) the atoms in your brain are no different from the ones in any other clump, and they just move around in the only way that’s compatible with the state of all the other atoms in the universe and the laws of physics.

3) No matter the details of the physics, there is only cause and effect, with some amount of randomness (assuming it’s even a thing), and there’s absolutely no room for your ego to magically unshackle itself from this simple ground truth.

4) the concepts of freedom and will you’re invoking stem from the notion of counterfactuals – because alternative versions of cause and effect chains are conjured in our mind (like an AI hallucination), and you believe those could have happened, therefore, at every moment, we must have the ability to pick among those possible future paths arbitrarily, before even knowing what those possibilities are, or could have been. As if, when imagining the counterfactuals after the fact, you’re able to signal the past version of yourself to alter its “choice”, and this would repeat in a loop until it stabilizes. But even in this far fetched theory of retroactive influence, everything that happens is still a chain of inescapable causes and effects, going in a time loop. And the stable version is left with no freedom of will whatsoever.

The earlier commenter replied:

I simply don’t agree. If you can explain what causes an electron spin measurement to be either up or down, and thereby abolish superposition, then maybe I am mistaken. Otherwise I can base my decisions on the outcome of the measurement and not possible for any entity to predict my resulting actions a priori. The cause of my actions is the measurement result that is random with no discernible cause. The cause and effect chain starts only after measurement and not before.

I agree my brain is made from atoms but don’t agree it is just atoms as I look at collections of atoms that form my environment. My two dogs are collections of atoms but I realize one of them is not even bright enough to collapse the wave function. 🙂

How would one prove unequivocally there is no free will. Maybe if there were some digital simulation of me that perfectly predicted my actions then I would have to accept my sense of free will is illusory.. The future is perfectly predictable.

My point is that I can take steps to render a simulation inaccurate by consciously taking actions dependent on some measurement that cannot be known before the fact.

In lieu of a perfect simulation of me I will continue to believe that I have free will and that my future isn’t fully determined in a mechanistic way. If you have some predictive model of me we can conduct the experiment now to see if in fact I am perfectly predictable. I am ready and eager to participate. If you don’t then the claim that there is no free will is an unsupported belief while my belief I do have free will is supported by the failed experiment.

Free will is mostly a philosophical issue, but a few things can be said.

Everyone, except possibly for schizophrenics and stubborn philosophers, experiences free will and lives his life accordingly.

The above argument that science precludes free will is entirely fallacious.

Rejecting free will leads to nihilism and dispair.

Making a freely willed decision is somewhat like measuring an electron spin. For all the successes of quantum mechanics, we do not have a good theory for how such a measurement is determined. At best we can give probabilities. Some have even suggested that human consciousness is involved, it is so mysterious.

I am not saying that I have a physical explanation for how free will works. I do not. But it is also not the case that anyone has a deterministic explanation for how quantum measurements work.

The Bogus Dunning-Kruger Effect

Dr. Bee has a new video questioning the Dunning–Kruger effect. She gives papers on the subject 10/10 on the BS meter.

The dumber you are, the more certain you are that you aren't. That's the Dunning-Kruger effect, and it's been used to explain everything from bad politicians to the state of social media. Now psychologists have gone back to the biggest dataset we have and come out with the opposite conclusion. But there's a reason both sides can look at the same dots and see the opposite thing.

I agree with her. Somehow this obscure has entered public consciousness, and people commonly cite confidence as evidence of incompetence, according to this effect.

The whole subject appears to be an inside joke. Anyone citing the Dunning-Kruger effect is likely an idiot.

I have criticized the effect as bogus here, here, and here. The published effect is really just a misunderstanding of regression to the mean.

No Born Rule in Many Worlds Theory

There are dozens of papers on trying to reconcile the Born rule with the many-worlds interpretation of quantum mechanics. At first this seem bizarre. The Born rule is the main empirical prediction of QM, and any interpretation would have to obey it. How else could it be an interpretation?

The position I have repeatedly taken on this blog is that many-worlds is not an interpretation, and is wholly incompatible with the concept of probability, whether expressed by the Born rule or anything else.

Now a new paper agrees with me, and says it well:

Against Many Worlds
Emily Adlam, Jacob A. Barandes

Any viable interpretation of quantum theory needs to account for the Born rule, from which the theory gets its probabilistic empirical predictions. In this paper, we give an overview of possible approaches to this problem in the context of the Many Worlds interpretation. We argue that, for structural reasons, none of them can possibly succeed. More precisely, we argue that the Many Worlds interpretation must obtain the Born rule by proceeding either axiomatically, deductively, or inductively, and that all three of these approaches run into general, fundamental obstructions.

I previously posted a video explanation by Barandes.

You cannot just postulate probabilities for the many-worlds branches of the wave function. It is essential to many-worlds that the branches emerge as a result of decoherence. The branches cannot be assumed to have probabilities, and in fact no one has been able to make sense out of some branches being more likely than others.

The issue goes right to the core of what we mean by probability. For this, the paper recites a textbook probability problem for comparison.

A more fundamental obstruction is that merely having multiple things is not enough, by itself, to imply a conception of probability. To see why, consider, for example, a jar containing 43 red marbles among a total of 100 marbles. On the one hand, it would be entirely reasonable to say that the fraction of red marbles was 43%. On the other hand, it would not yet be reasonable to call this fraction a probability, because, at this point in the thought experiment, the probability would not be a probability of anything. It would not make sense to say that there was a 43% probability simpliciter.

Of course, if we were to stipulate that an agent were picking a marble ‘at random’ from the jar, under some suitable notion of ‘at random’ and with the further assumption that the marbles were ‘well-mixed,’ then we would be in a reasonable position to assert that there was a probability of 43% of the agent picking a red marble. But absent such a picking process, or the introduction of some other selection process to serve a similar functional role in the argument, there would be no probability, because, again, there would simply be nothing for the probability to be a probability of.

It is worth dwelling for a moment more on what premises are needed for a ‘picking process’ to legitimate the move from 43% qua fraction to 43% qua probability. Assigning a picking probability of 43% amounts to adopting a ‘principle of indifference’ (Eva 2019) assuming that all the marbles are ‘equally likely’ to be picked. And the philosophical literature on indifference principles makes clear that they are justified and successful only when they reflect relevant features of the process by which an outcome is selected, such as the symmetries of the picking process (van Fraassen 1989, Shackel 2007).

In particular, assigning a probability of 43% makes sense if the agent responsible for picking a marble is doing so by some process that is blind to color, meaning that there is a symmetry with respect to color – hence our invocation of terms like ‘at random’ and ‘well-mixed.’ Obviously if the agent is picking marbles in some way that is biased toward red marbles, then it would not make sense to assign a probability of 43%.

Just as you cannot abstractly assign probabilities to a jar of marbles, you cannot assign them to a wave function either. You have to have some notion of a measurement, and that is what the many-worlds folks refuse to do.

I doubt that any of these arguments are new. Many worlds theory is wholly incompatible with a scientific outlook, and is a crackpot belief.

The only way to understand many-worlds is as a complete rejection making scientific predictions, and a complete rejection of probability. That is, to accept it you have to say that no predictions can be made, and that probability theory does not make sense.

In Sean M. Carroll's recent podcast, he rambles in support of many-worlds theory:

I have heard you describe more unlikely worlds, thinner branches as not mattering as much. I don't know what your ontology of mattering is here, but it brought is it broadly speaking the fact that there are just more of the likely versions and therefore we should care about those more? If so, would you consider this a fundamentally utilitarian approach to mattering?

Well, uh it's not that there are more likely versions. ...

That's that's the entire trickiness of many worlds is that you can't get the counting that you need to get the Born rule right just from counting the numbers of worlds. The worlds don't count equally. They count as much as the amplitude squared. So 99% for this one, 1% for that one.

So I use the word mattering just to be indicative of whatever matters to you. So it might be the probability that you find yourself in such a world. It might be how much energy such a world has. It might be the utilitarian uh utility function that you have for such a world. Whatever you think is assigned to the world mattering goes along with the amplitude squared in many worlds. I think that's a good motto to keep in mind to get you through figuring out how to think about this scenario.

In quantum mechanics, the amplitude squared gives the probability. In many worlds, you cannot count the worlds or get more likely worlds. What you get is that some worlds subjectively matter more than others. You can think of the mattering as being based on probability or energy or utility function.

Really? This is all nonsense. There is nothing scientific about some worlds mattering more than others.

Superdeterminism makes Science Impossible

Superdeterminism has very few adherents, like Sabine Hossenfelder. I would ignore it completely, as too nutty to discuss, except that she has become one of the most prominent expositors of Physics.

Here is a new paper, discussing one of the main problems:

This is one of the longest-standing objections to (naïve) superdeterminist theories, often referred to as the Tobacco Company Syndrome (Hossenfelder and Palmer, 2020). Fundamentally, it is an epistemological concern. In Maudlin’s formulation (2019), the objection draws an analogy with a tobacco industry apologist who first invokes a common cause to deny that smoking produces cancer, and then, when confronted with randomised animal trials, insists that the randomisation procedure itself was somehow biased in favour of placing cancer-prone subjects in the experimental group, a move Maudlin regards as wholly unscientific. It is not difficult to see the risks to scientific practice if this objection were to remain unresolved within superdeterminism. Such a denial of Statistical Independence would imply, for instance, the impossibility of discussing isolated systems or experimental random errors. Moreover, it would pose a significant risk to the empirical coherence of the theory, as defined by Barrett (1996).

That is right, if you believe in superdeterminism then it is impossible to do an experiment to prove that tobacco smoking causes cancer. It is also impossible to do any other experiment, or reach any scientific conclusions.

Many-worlds theory has similar problems. It makes all scientific work impossible.

There is a Frame for Defining Now

From a recent video explaining relativity:

There is no way to draw a single 17:33 consistent line across spacetime that all observers agree separates has happened from hasn't happened yet. The failure parade. Some physicists argued this was a mathematical convenience with no ontological weight. But no experiment has ever found a preferred frame that would let now become absolute again.

Not correct. Such a frame is given by the Cosmic microwave background.

This video is AI generated, so should not be taken too seriously. It is a common error, so I point it out.

Relativitism Corrupted Humanities and Social Sciences

This blog focuses on the hard sciences and objective reality, mainly Physics. I expect physicists to be hard-headed scientists, and I am disappointed when they are not.

Outside the hard sciences, the academic world is a wasteland. They are concerned with relativitism, not relativity.

A bunch of experts were asked to write a report on the decline of humanities and social sciences. Objectivity has been abandoned. Political considerations have overwhelmed serious work.

Here is a summary by a biology professor. One example:

In 2023, the AAA canceled an accepted panel at its annual conference, titled “Let’s Talk About Sex, Baby: Why Biological Sex Remains a Necessary Analytic Category in Anthropology.” The AAA said at the time that the panel would have harmed members’ “safety and dignity,” and that its premise contradicted “settled science.” The Vanderbilt report cited this incident to argue that when politically charged questions “are treated as settled by the scholarly community, the result is an illegitimate suppression of dissent.”

Greene on the Most Mind-Boggling Result

Brian Greene answers questions:

And I'd say the most mind-boggling result of quantum 1:44:38 mechanics is the result of John Bell, John Stewart Bell, ...

But he was not willing to accept quantum entanglement at face value. He wanted to know if a version of quantum mechanics might be constructed that would be that would be local that wouldn't have these spooky non-local influences.

and he was trying to realize a dream really of Albert Einstein who thought that might be possible to come up with a new theory that embraces the same predictions as quantum mechanics but doesn't have this weird spooky action and he worked very hard thought he was coming very close to formulating that theory but every time he got to the near to the end it fell apart.

so finally he said to himself maybe this is a fool's errand maybe it's impossible to have a theory that agrees with the predictions of quantum mechanics and yet somehow excises spooky action non-local influences and he proved that that is the case a beautiful theorem John Bell proved a theorem which then because of the experimental data being such as to confirm the version that I just described that any quantum mechanically uh viable theory that agrees with the predictions of the standard approach is necessarily non-local. And so that to me is the most stunning implication. We necessarily live in a nonlocal universe. Yeah.

In fact, as a small teaser, I'll mention I'm actually writing a new book on quantum mechanics.

It is going to be a horrible book.

Quantum mechanics does not have any spooky non-local influences. Bell only proved that it differs from a type of local classical theories. It left open the possibility that a spooky classical theory might agree with QM, but that is not what Einstein was looking for.

Greene is famous for promoting string theory and multiverse theories. He now admits that these are speculative and unproven. Next he is going to promote quantum spookiness.

Meanwhile, Dr. Bee is writing a similar book and ventures into spookiness in her latest video on You probably misunderstand the double slit experiment. She is mostly correct, but goes off the rails when she says it proves either superdeterminism or spookiness:

4:37 But if the particle went through the other slit, you never interacted with it. You didn't touch the particle. It just knew you were trying to. The only way that you can avoid this faster than light update is that the particle knows whether you'll measure it before you've made the decision whether to measure it or not. This is what's called super determinism. And this is what's so weird about the double slit.

The error is that she keeps talking about particles, as if each has a definite position and momentum even when not observed. They do not. They are really waves, and waves always make interference patterns when going throught a double-slit, or even a single slit. There is no need to invoke spookiness or superdeterminism.

She refers to another quantum oddity. If you fire an electron at a double slit, it is really a wave that goes through both slits. If you put detectors on the slits, you only find one electron. If you observe the electron in one slit, you can be sure that the other detector will not see it. Conversely, if a detector says no electron went through one slit, then the other detector will see it in the other slit.

She describes this as very strange:

You make a measurement at one slit and the wave function on the other slit 4:22 changes immediately. And note that this is the case even if the particle does not trigger the detector. If it doesn't go into the detector, you know it must have gone through the other slit. But if the particle went through the other slit, you never interacted with it. You didn't touch the particle. It just knew you were trying to.

It is not so odd. Suppose I put my car key in my pocket, and I forget which of my two pockets. I can reach into one pocket, find it empty, and immediately deduce that my key is in the other pocket. I am finding my key, without interacting with it. It is not spooky or superdeterministic. The key did not know that I was looking for it.

It is peculiar to QM that measurements give eigenvalues. The theory says that the (unmeasured) electron goes partially through each slit. It could even tunnel through a barrier. But you never observe half an electron or an electron being in two places at once. Those are the real quantum mysteries. But the spookiness and superdeterminism are just nonsense.

Update: Greene has posted a long interview of Sean M. Carroll.

New Paper Calls Quantum Computing a Big Failure

Dr. Quantum Supremacy responds to a new paper:

A week ago, a philosopher named Amit Hagar put out a preprint entitled The NISQ Trap: Eight Years of Demonstrations the Hardware was Built to Lose. Here’s the abstract:

With a single clear exception, every NISQ-era flagship demonstration of ‘quantum advantage’ has, within eighteen months of its announcement, been classically reproduced, shown to rest on classically tractable structure, or closed by a simulability theorem.

Nevertheless, the basic thesis — that quantum supremacy in the NISQ (Noisy Intermediate Scale Quantum computing) era has been a failure, or even an example of pathological science — seems surprisingly widely shared, along with the opposite thesis that quantum computing already gives oodles of useful advantages for optimization and finance. ...

You’re right that, if you just look at the commercial value right now versus money spent to date, quantum computing looks like a colossal waste — as did classical computing before it scaled, as did nuclear power and heavier-than-air flight and every other nontrivial new technology.

Nobody invested a nickel in those other technologies, until someone had a proof of concept. Quantum computing is different in that it has attracted many billions of investment dollars, without a proof of concept, or a business plan on how it would ever make money.

This response is very weak. Go ahead and read it and decide for yourself. Someone publishes a paper saying the whole field has been a failure for 20 years, he admits that the view is widely shared, but says new technologies take time.

And of course he falls back on the argument that quantum computers have not been proved impossible:

Yes, the “real” goal remains, as it’s been since the 1990s, to build a scalable fault-tolerant quantum computer — and I’m glad that Hagar (unlike, say, Gil Kalai) never suggests that we’ve learned anything to rule that goal out. In the meantime, an intermediate goal would be to use NISQ devices to do physics and chemistry simulations that are commercially useful, or that help solve important scientific problems. The point of quantum supremacy experiments, you might say, is that by demonstrating the reality of quantum speedup about as clearly as it can be demonstrated with current hardware, they let us cleanly turn our attention to those more ambitious goals.

Hagar does not even talk about whether quantum computers are possible. His point is that they are empirical failures.

Update: Aaronson says:

For me, though, the real game always has been to learn the truth or falsehood of the Extended Church-Turing Thesis: that is, can all of nature be simulated on a standard Turing machine with polynomial overhead, or is quantum mechanics a counterexample to this? And from that standpoint, it seems obvious to me that the burden of proof lies with those who think the Extended Church-Turing Thesis still holds in a quantum-mechanical universe. Why does it hold? What astonishing new principle causes it to hold, despite the exponentially many canceling terms that generically arise in a path integral?

I get enraged when (some) QC skeptics seem not even to notice these profound questions, and content themselves to poke holes in this or that specific quantum fault-tolerance proposal and then call it a day. That seems to me to reflect a staggering incuriosity about the way the world actually is.

I am probably one of those enraging him. I agree that the big question is whether quantum computers can outdo Turing machines. Since he is a leading expert and he regards this as still an open question, then I accept that it is unsettled.

I disagree about the burden of proof. I say that a new technology should show that it works. He says the burden of proof is to show that it does not work.

The profound question is to understand why it does not work. I am somewhat incurious about this, because I am not sure any explanation will be fully satisfactory. Do we understand why F=ma, or how gravity workds, or why nothing goes faster than light, or why perpetual motion machines are impossible? Yes, I know the textbooks have explanations, but they do not stop people from theorizing about them, so apparently the explanations are not totally convincing.

That Extended Church-Turing Thesis was widely accepted for 50 years, including by very smart physicists who understood quantum mechanics. If quantum supremacy were possible, I would think that someone would have demonstrated it by now.

Update: Gil Kalai Says:

I also tend to agree with Amit’s conclusion that the quantum supremacy claims are either false or inconclusive. ...

For more than two decades, Scott has argued that the burden of proof lies on the shoulders of the skeptics. I suggest that we split the burden of proof as follows:

The burden of showing that the engineering obstacles to high-quality quantum error correction and quantum fault tolerance are not far beyond reach lies with the proponents of quantum computing. They still have a long way to go before providing convincing evidence that these goals are feasible and can be achieved in the foreseeable future.

Foundations of General Relativity

The foundational principles of general relativity are said to be the equivalence principle, and general covariance. This new video disagrees.

Here is how I see it. The equivalence principle could also said to be foundational to Newtonian classical gravity, and it is only associated with general relativity for historical reasons.

General covariance is a mathematical principle, and it is not clear that any mathematical principle, by itself, has any physical consequences. As Einstein is quoted:

As far as the laws of mathematics refer to reality, they are not certain; and as far as they are certain, they do not refer to reality.

The laws of math do have consequences, but they require interpretation.

We normally say that Euclidean geometry has physical consequences, but that is only because of some assumptions about how idealized points, lines, planes, etc. correspond to physical entities.

Historically, covariance under Lorentz transformation was crucial to the presentations of relativity given by Poincare in 1905 and Minkowski in 1907. Einstein had a weaker version that was more similar to what Lorentz published in 1895.

Einstein was unconvinced, and wrote a paper denouncing the concept in 1913. Eventually, Grossmann, Levi-Civita, Ricci, and Hilbert convinced him to accept it.

The main ideas behind relativistic covariance had already been published by Poincare and Minkowski, but they were dead by the time Einstein worked on general relativity

I believe that general covariance is a crucially important concept. It is a way of saying that the laws do not depend on coordinates. That different frames can be used to the same effect. These ideas go right to the core of why the theory is called relativity. And historically, covariance was a large part of why relativity was accepted. By 1908 it was widely accepted, even though Einstein did not.

Aaronson Starts a Venture Investing Fund

Dr. Quantum Supremacy is tired of watching his friends and associates get rich with half-baked ideas about bitcoin, quantum computing, and AI, so he is joing the gold rush. He announces:

So then, screw it! In partnership with my brother David Aaronson, who’s long done investing professionally, and on David’s guidance and encouragement, I’m hereby embarking on a new policy. ...

So far, David and I are investing in:

• Oratomic (the new neutral-atom quantum computing startup out of Pasadena started by Dolev Bluvstein, John Preskill, and Hsin-Yuan Huang among others, which I previously blogged about here). See here for their Series A announcement this week. ...

So let me educate you then. I played a central role in introducing the whole program of sampling-based quantum supremacy back in 2009-2012. That what was allowed first Google, then Quantinuum and others, to do such dramatic demonstrations starting in 2019. Almost anyone in quantum computing will tell you that that was essential to various QC startups being able to go public recently at ~$15B valuations.

Now, what’s true is that I didn’t succeed at capturing essentially any of the value that I helped create this way. What’s true is that all the billions of dollars went to other people, who optimized for getting equity early on, rather than to me or my students, who optimized for learning the truth about quantum speedups and writing the key papers that other people then managed to profit from.

This is another sign we are in a bubble.

Oratomic says that it will not try to demonstrate quantum supremacy. It will go straight for a commercial quantum computer.

The Tycho Solar System, Revisited

I always thought Tycho Brahe was the greatest pure astronomer, so I was surprised to find this new paper trashing him:

The Tychonic system, an implausible theory
Gabriele Vanin

Several scholars have accorded the Tychonic System a prominent place, considering it a credible alternative to the Ptolemaic and Copernican systems. In recent years, there have even been some enthusiastic contributions in support of Brahe, which have sought to show how his solutions were the most suitable for providing scientifics support to the new observations that were accumulating between the late 16th century and the mid-17th century.

Tycho had a model where the Sun went around the Earth, and Mercury and Venus went around the Sun.

At a time when Copernicus had already published his hypothesis, it could not but be considered a step backwards, a return to the past, a naive and useless compromise. That is, as an ad hoc theory (the kind that sci- entists gladly do without if they can, and, in this case, they could) contrived on purpose to have the best of both worlds, i.e. the immobility of the Earth and the concentric motion of the planets around the Sun, as a geometrically and physically poorly founded expedient. After Copernicus finally have the courage to declare that in the homocentric spheres, in epicycles and deferents lies hidden the orbit of the Earth, does Brahe return to the starting point, swapping the orbit of the Earth with that of the Sun again? It would be a bit as if, in the aftermath of the publication of the On the Origin of Species, someone had said that yes, evolution was acceptable, but that the evolution of man had occurred on a parallel track, not descending from the first common life-form, but from another one, completely different, and had developed quite independently.

This is ridiculous. There were good arguments for and against the motion of the Earth.

The paper later complains that Kepler's third law, discovered in 1619, could be applied to Saturns moons, discovered in 1656-84, and that would seem to falsify the Tychonic system. Tycho died in 1601.

I have seen others trash Tycho for being too stupid to accept Copernicanism. I wonder about what their idea of science is.

Tycho had the best pre-telescope astronomy observations of anyone. By far. The stars had been tracked for thousands of years, but Tycho's data was so good that it made all preceding data worthless.

A geocentric model is not necessarily wrong. Motion is relative. We say the Earth goes around the Sun, not because it is a fact, but because dynamic models make more sense in a center-of-mass frame. It makes more sense to say the smaller object is orbiting the larger object.

Tycho was creating a more accurate model of the solar system. He did not know about gravity. But without him, it is hard to see how Kepler and Newton could have accomplished what they did.

To me, science is all about collecting data, and constructing explanatory and predictive models. Tycho did this, possibly more brilliantly that any other astronomer in history. And yet people want to trash him as unscientific.

A scientific theory can be extremely good, even if the interpretation is not quite right. For example, Murray Gell-Mann is famous for his theory of quarks, but originally he did not even believe in quarks. As the quarks cannot be isolated, it is a philosophical issue whether they are real or not. Now everyone says they are real, but Gell-Mann was not really wrong to treat them as theoretical constructs.

IBM Plans to Scale Up Quantum Computing

The WSJ reports:

International Business Machines has spent a decade building, testing and improving in the once-theoretical realm of quantum computing at its glass-paneled laboratory complex in upstate New York.

This year, the company is laying the groundwork to turn that technology into a fully-fledged, scalable business from an expensive science project.

IBM said last month it plans to form a new independent subsidiary called Anderon, a foundry to produce the silicon wafers needed to make quantum-computing processors. The venture is seeded by a $1 billion investment from the Trump administration and another $1 billion of IBM’s own cash.

Anderon will give the company a new line of business in selling wafers to other quantum-computing companies. It will also provide a steady stream of wafers to continue developing its own quantum technology, positioning IBM to capture part of what the Boston Consulting Group projects will be a $90 billion to $170 billion market for quantum-computing providers by 2040.

The venture will begin with an expansion of the company’s in-house facility in Albany, N.Y., which makes wafers for IBM’s internal research. “Now we have to scale it,” IBM Director of Research Jay Gambetta said in an interview.

The company also plans to spend an additional $9 billion over five years to advance the final stages of its quest to build a quantum-mechanics-powered computer capable and reliable enough for widespread use, a goal known as fault tolerance. That computer, named Starling, is being targeted for 2029.

With Anderon, IBM is thinking beyond Starling, or even a more powerful quantum computer planned for 2033.

I don't see how IBM can be scaling up, when they do not have a working technology, and do not have applications for the scale up.

Dr. Bee has a new video trashing the prospects for quantum computing, and in particular, IBM for quietly shifting goals. It is still committed, but cannot achieve its previous goals.

Sean M. Carroll Pushes Many-worlds Again

New interview:

Quantum Mechanics Still Doesn't Make Sense | Sean Carroll
New Scientist

What really happens in the quantum world?

In this conversation, physicist Sean Carroll explores some of the deepest mysteries in quantum mechanics: the famous double-slit experiment, wave function collapse, the Many Worlds interpretation, entropy and the arrow of time.

Speaking to New Scientist reporter Jacklin Kwan, Carroll discusses why electrons appear to behave like waves, how observation seems to affect reality and whether the universe constantly branches into countless parallel worlds. Carroll also explains the measurement problem, the challenges of interpreting quantum theory and why physicists still debate what quantum mechanics is actually telling us about the nature of reality.

Carroll is a theoretical physicist, cosmologist and author whose work focuses on the foundations of physics, quantum mechanics, cosmology and the nature of time.

He gives a brief explanation of why he believes in many-worlds theory, if you want it straight from him.

He lays it out clearly. When you have a situation where several outcomes seem possible, you forget about making a prediction, and just assume that they all happen in parallel universes.

That's it. It does not really have anything to do with quantum mechanics or the Schroedinger equation. It is just a rule for replacing probabilities with unseen worlds.

You do not even get to say that some worlds are more likely than others. You just have to accept that all possibilities happen, and nothing can be predicted.

I'm a believer in the many worlds interpretation of quantum 16:58 mechanics. ... But really it's just a theory that says you want the answer to the reality problem. Wave functions are real. They represent reality precisely and completely. ...

When I observe the electron, I don't observe the wave function. I observe different possibilities. You can interpret that as saying that what the wave function really is is a superposition of all possible measurement outcomes. And what you're seeing when you measure it is you're picking out one of the set of possibilities with different probabilities. Right?

And so what many worlds ends up saying and it's not an assertion again it's it's comes out of the equations but it says that what you send up with is a superposition but not just of the electron the combined system of the electron and the scientist doing the measurement. There's a superposition of the electron was here and you observed it there plus the electron was there and you observed it there plus the electron was there and you observed it there and the many worlds boom they're there.

And so if you accept the picture of wave functions as superpositions of different measurement outcomes it's no work at all to accept them as superpositions of different possible worlds and then everything is just physics.

So you observe electrons, not wave functions. The wave function is just a way of representing all the possible observable outcomes. All that is correct. But then he makes the leap to say that "it's no work at all" to accept the possible outcomes as parallel worlds.

I spell all this out, because most readers here probably think I am attacking a straw man when I say how stupid many-worlds theory is. It really is stupid and nonsensical. Just listen to its most eloquent advocate.

It is no work at all to replace a scientific theory with one that makes no predictions. Just say that actual observables are not real, and that the real entities are the imaginary possibilities that no one ever sees. Just forget about making predictions, and brag about the explanatory power of being able to talk about unobservable parallel universes.

I continue to be amazed that anyone takes this nonsense seriously. It is impossible to reconcile this with a scientific worldview. It is the most anti-science philosophy.

Microsoft Qubits Questioned Again

Microsoft has been trying to build quantum computers with topological qubits, which would have durability advantages if they worked. But SciAm reports:

A top quantum computing expert assails Microsoft’s claims that it has a “topological qubit,” arguing in a new paper that the company has failed to demonstrate the technology.

University of St Andrews physicist Henry Legg argues that the “topological qubit,” a storer of quantum information that could theoretically maintain a higher fidelity than any in existence, might simply be noise.

The commentary was published today in Nature’s “Matters Arising,” the journal’s venue for formal criticism of its published papers. Legg’s response is aimed at Microsoft’s most recent Nature paper, which was published earlier this month—but it is just the latest in a string of criticism aimed at Microsoft’s Quantum division by other researchers in the field.

The company has been forced to retract certain previous peer-reviewed papers. And in the new commentary, Legg argues that their most recent Nature paper may be similarly flawed. In a response also published today by Nature, a member of Microsoft’s Quantum team argues that their measurements do justify the claim that they’ve produced a topological qubit.

The Microsoft paper has about 150 authors. I did not understand it. Maybe Microsoft has done some Nobel Prize quality research, but I doubt it. I wonder how the scientists convince Microsoft management to keep funding it. There is no possibility of a monetary payoff.

I did not know that Nature had a journal for formal criticism of its papers. That seems like a good idea.

Gil Kalai defends his Quantum Skepticism

Dr. Quantum Supremacy has running feuds with several people, including Israeli mathematician Gil Kalai who says in response:

When it comes to quantum computing, I regard the choice made by many colleagues and friends to pursue this direction of research as entirely (3000%) justified, and it has led to beautiful mathematics and science. (I hope that my own choice to pursue the skeptical direction was justified as well, though this is a more difficult call.)

For the record, I expect that scalable quantum computation — and even important milestones toward this goal — are inherently impossible. ...

One very clear and important point of disagreement between Scott and me concerns the following simple factual question:

Is it currently possible to produce (without classical-computing interference) samples of size 500K for depth-14 random circuits with 20 qubits and an XEB fidelity above 0.2? (This is something Google claimed in 2019.)

I tend to think that the answer to this question is negative; Scott strongly believes that this and much more have already been achieved experimentally. Thus these are sharp factual disagreements, and I hope they will be resolved within the next few years.

I do not know who is right about that factual question, but I am included to believe that Google did not really prove it, if Kalai is till unconvinced seven years later.

Aaronson is also wound up about some Middle East war issues. I understand why he is so pro-Israel, but I do not understand why he is so anti-Trump. At any rate, it is out of my expertise.

The Modern Crypto Primitives

The quantum crypto folks are threatening the cryptographic infrastructure that runs the world. I summarize the main functions.

Hash. Any digital message can be transformed to a 32-byte hash. It is believed that no one will ever find two messages with the same hash.

Public key pairs. A 32-byte private key can be converted to a 32-byte public key. No one can invert the operation.

Signature. A private key and a message hash can be signed to a 64-byte signature. The public key can verify the signature. No one can forge the signature without the private key.

Encryption. A 32-byte secret key can encrypt an arbitrary digital message. No one can decrypt without the secret key.

Key agreement. Two public key pairs can be combined to give a 32-byte shared secret, using the private key of one and the public key of the other. Both ways give you the same shared secret. No one can get it from just the public keys.

These functions are quite efficient and underlie the security of nearly everything on the internet today, notably https and ssh. They also underlie Bitcoin. There is broad agreement that they cannot be cracked with conventional Turing computers for the foreseeable future. There are widely available free libraries to do all these things.

Quantum computers threaten the public keys, making it possible to deduce the private key. Google predicts this will happen on what they call Q-Day, maybe as early as 2029. I doubt it will ever happen.

There are post-quantum public key pairs that are resistent to quantum algorithms, but they are much bigger and not as well tested.

Microsoft and Google are hard at work replacing their public key pairs with post-quantum ones. Some high-value targets want to switch now, because of the possibility that transmissions are being recorded now so that the Chinese can decrypt ten years from now. For a typical user ordering an online product, this is not a concern.

This reminds me a little about how IPv6 was designed to replace IPv4 internet protocol in 1998. Most of you reading this are probably still using IPv4. The reasons for change are different, but it shows how the networks are slow to change, if the old system is working just fine.

Einstein Objecting to using Observables

Einstein's unhappiness with quantum mechanics was already clear in 1926, within a year of the theory being formulated.

Heisenberg told this story:

In the spring of 1926, I was invited to address this distinguished body [University of Berlin] on the new quantum mechanics, ...

[Einstein said] "What you have told us sounds extremely strange. You assume the existence of electrons inside the atom, and you are probably quite right to do so. But you refuse to consider their orbits, even though we can observe electron tracks in a cloud chamber. I should very much like to hear more about your reasons for making such strange assumptions."

“We cannot observe electron orbits inside the atom," I must have replied, "but the radiation which an atom emits during discharges enables us to deduce the frequencies and corresponding amplitudes of its electrons. After all, even in the older physics wave numbers and amplitudes could be considered substitutes for electron orbits. Now, since a good theory must be based on directly observable magnitudes, I thought it more fitting to restrict myself to these, treating them, as it were, as representatives of the electron orbits."

"But you don't seriously believe," Einstein protested, "that none but observable magnitudes must go into a physical theory?"

[Heisenberg] "Isn't that precisely what you have done with relativity?" I asked in some surprise. "After all, you did stress the fact that it is impermissible to speak of absolute time, simply because absolute time cannot be observed; that only clock readings, be it in the moving reference system or the system at rest, are relevant to the determination of time."

"Possibly I did use this kind of reasoning," Einstein admitted, "but it is nonsense all the same. Perhaps I could put it more diplomatically by saying that it may be heuristically useful to keep in mind what one has actually observed. But on principle, it is quite wrong to try founding a theory on observable magnitudes alone. In reality the very opposite happens. It is the theory which decides what we can observe.

Einstein is credited with abolishing the aether, absolute space, absolute time, etc., but maybe that is not how he thought about it at all. He went on to try to develop unified field theories that were completely detached from observation.

Generating Quantum Randomness

I posted about someone claiming to prove quantum randomness, and now the journal Nature has published an article on Experimental randomness amplification. The article is paywalled, but here is the 2024 preprint.

In this context, randomness is defined as being fundamentally unpredictable, which means that the laws of physics forbid the prediction of its values. ...

Conventional random number generators, rooted in classical physical processes, grapple with a foundational concern — the potential for adversaries to predict their outputs by scrutinizing the microscopic degrees of freedom, thereby eroding their essential unpredictability.

Quantum-mechanical processes, on the other hand, feature innate randomness and therefore offer a natural ground to build such devices.

This is foolishness. There is no law of physics forbidding prediction.

When you measure X-component of electron spin, then the wave function collapses, and the Y-component has 50-50 chances. Heisenberg uncertainty prohibits measuring the X and Y components at the same time. So the theory is sometimes not able to predict spin. But that is not quite saying that prediction is forbidden. Maybe there is some way to predict, and we do not know how yet.

This paper does not even talk about spin. It merely assumes that you are doing Bell test experiments. Under some assumptions, you can make some random choices, and get even more random outputs.

This method could be used to generate random numbers for practical purposes like cryptography, but I do not think it is any better than tossing coins, or pointing a webcam at a lava lamp.

The Evidence for Quantum Supremacy

A newly revised paper reviews the outstanding evidence for quantum supremacy:

A brief history of quantum vs classical computational advantage
Ryan LaRose

In this review article we summarize all experiments claiming quantum computational advantage to date. Our review highlights challenges, loopholes, and refutations appearing in subsequent work to provide a complete picture of the current statuses of these experiments. In addition, we also discuss theoretical computational advantage in example problems such as approximate optimization and recommendation systems. Finally, we review recent experiments in quantum error correction -- the biggest frontier to reach experimental quantum advantage in Shor's algorithm.

So has it been proved, or not? No, not really.

It seems at this moment in history we are just on the boundary between quantum and classical computational advantage, and in the near future we expect the status of computational advantage to continue shifting between quantum and classical. We hope that this brief history helps to propel readers to the research frontier and develop new ideas which advance both classical and quantum computation.

There is still no convincing experiment that quantum computers are possible.

I would have predicted that by 2026, quantum computing would be proved possible, or investor money would dry up. I was wrong. Investor enthusiasm for QC is stronger than ever. The Trump administration just instested $2 billion into QC, mostly in startup companies. There are about five of these companies worth about $10B apiece. Most of them went public in sneaky financial maneuvers where they did not have to disclose all their risks.

New Scientist magazine just posted a video on Quantum Computers Are More Dangerous Than You Think:

On Q-Day, your privacy will be at stake. This is the moment when quantum computers break the encryption protecting the modern world, bank transactions become readable, private messages get exposed and even state secrets become vulnerable.

For years it sounded like sci-fi, something that was decades away from happening, if it happened at all. But now, research suggests that we may be hurtling towards Q-Day at a rapid speed.

In this video, New Scientist uncovers why many experts think the countdown to Q-Day may already have begun, and explains how quantum computers work and why these machines could both threaten the security of the modern world and unlock breakthroughs that could change our lives. Special thanks to Quantum Motion for letting us film at its facilities.

If and when the countdown to Q-day starts, I think that we will have about ten years. Only the computer security companies, NIST, and the US Dept of War need to prepare now.

For most people, SSL/TLS/SSH is just a way of getting assurance that a web site is real, and that no one is stealing a password or credit card number. If Q-day hits, you will just update your browser and not notice the difference.

The Bitcoin blockchain would have to be restructured, and that is feasible as long as there is a consensus on how to do it. A consensus could take a couple of years.

New Carroll Podcast to Defend Many-Worlds

Sean M. Carroll is frustrated by physicists preferring Copenhagen over many-worlds, so he posted a podcast with his views:

Solo: Looking Quantum Mechanics in the Eyeball | Mindscape 355

One of the major obstacles to understanding quantum mechanics is the difficulty we have in simply accepting what the theory itself is telling us. The problem is that we know what the everyday world looks like -- stuff, arranged in space, evolving through time. So we can't resist the temptation to impose that picture on the quantum description, even if it's not actually there. In this solo episode I talk about what it means to take quantum mechanics at face value, and the difficult work involved in understanding how the everyday world of our experience fits into the picture.

For a more technical description, he refers to his paper, Reality as a Vector in Hilbert Space.

This podcast is a defense of many-worlds theory. He says people believe in textbook QM because they are stupid:

And I suspect that most of the people in the physics survey who said that they're Copenhagenists don't actually have that view themselves, but mostly because they haven't thought about it very carefully.

He has learned that he loses his audience if he talks about the parallel universes too much.

[24:35] Many worlds by contrast and we're not going to get into the worlds aspect of many worlds that much because it's actually not relevant for what we're talking about today. I've often said that many worlds is not mostly about the worlds. The worlds come along. They're there, no doubt. But what many worlds is really about is saying there's no such thing as collapse.

I think the problem here is that he makes several serious conceptual errors.

First, he misunderstands probability. Probability is not a real physical thing, like energy or temperature, but it is essential to all scientific theories. Let me repeat. All scientific theories are based on probability predictions.

You might say: No, when Nasa sent the Artemis II rocket around the Moon, it was all deterministic Newtonian physics.

I say: Not correct. Nasa calculated the splashdown location as a probability distribution.

Next, a probability is a prediction that something will happen, with other things not happening. If I say a coin toss has a 50% chance of heads, I am referring to getting heads, and not getting tails. If I get heads, I can then disregard any possibility of that toss being tails. For example, if I toss a coin three times, and the first is heads, then I can immediately eliminate the possibility of three straight tails.

The core concept behind many worlds does not even have anything to do with QM. It is: Throw away probabilities. Throw away definite outcomes. Assume all possibilities happen, in parallel universes.

If you toss a coin and get heads, it just means that you are in the universe with heads, and a parallel universe has tails. You cannot even say that the universes are equally likely.

When Carroll says (above) that "there's no such thing as collapse", he means that if you toss a coin and get heads, you cannot say that tails did not happen.

Carroll has no evidence for anything he says. His big argument is that it somehow simplifies QM to never predict any probabilities, never do any experiments, never accept any outcomes, and never reject any possibilities.

By eliminating QM's relations to the real world, he says we get more a pure and abstract theory where we do not have to worry about issues like what defines a measurement.

At 1:42:20, he denies that there should be any experimental test for many-worlds theory. He says "that's just not how science works."

On YouTube, a comment points out:

A very good introduction to the subject, but you dodge all the important critiques:

- Probability (Born Rule) is a huge problem for Many Worlds, but I know you are working on that.

- The Schrodinger equation is obviously wrong because spacetime is an assumed background. You dealt with space/locality a few times. There is work on entanglement creating space, and I agree so far, but how can QM/QFT get time to be emergent when it is a parameter not observable? (momentum is just a thick slice of time for mass position? - someone needs to handle that. What is the calculus of position and momentum in background-free local (infinitesimal) proper time?

- It seems that Hilbert Space changes dimension with the number of 'entities' in the system. But we know that 'entities' are variable (n particle quantum number is not defined). So we must immediately leap to the HS of the whole universe: a total matter/field equation that must be fixed, which defines everything (Wheeler/DeWitt). Anything less becomes some strange variable dimensional complex space - I don't buy that. [My personal world is computation and variable size matrices are just ... mathematically impossible. ] What happens with particle creation/destruction and second quantization - hint total hack!

- I don't see any contact with gravity - yes, AdS/CFT and holography, maybe perhaps with actual finite dS boundary defined by SR light cones, etc. but that has to emerge from large (in)finite HS, I don't buy it (yet). My tendency is to trust/believe the criticism from GR gravitationalists (Sir Roger, et al). I don't think you have addressed any of those criticisms (yet). You have to admit that background-free GR is much much more beautiful than background-dependent QFT (yeah, okay, subjective).

- Of all the symmetries in the SM, matter/antimatter and chirality asymmetry are the frontier. Yes, RH/LH neutrinos are vague/undecided/outside SM, and (maybe heavy dark) RH neutrinos could be the answer (cue Neil Turok) but GR people also have tentative answers for the background (Twistors, PIN decompositions, etc) I don't hear solutions coming from pure QFT formalisms.

I could go on, but enough already.

Carroll's excuse for ignoring all these points is "I have to confess that the whole idea of steelmanning is not really my vibe". [10:00]

Einstein's Friend Wrote a Relativity Paper

Hector Giacomini writes a new paper:

This note examines an apparently unpublished manuscript on special relativity written by Conrad Habicht in 1914 and made available online by the ETH-Bibliothek Zürich in December 2024. To the best of my knowledge, no study of its content has yet been published. Habicht was one of Einstein's closest companions during the Bern years. ...

The manuscript offers a clear and pedagogical presentation of special relativity.

Habicht was a friend of Einstein before 1905, so this might have given insights on where Einstein got his relativity ideas. Unfortunately it does not.

Habicht’s manuscript is much closer to the 1907 exposition than to the original 1905 paper. The 1905 article is remarkably austere in its historical framing: it begins with the magnet-and-conductor asymmetry, formulates the two postulates, defines simultaneity, and proceeds directly to the transformation laws and their consequences. It contains no historical reconstruction of Lorentz’s theory, no explicit discussion of the Michelson–Morley experiment, and no narrative account of the ether problem. The 1907 review article, by contrast, restores much of this background. It places special relativity within the history of the conflict between the Galilean principle of relativity, Maxwell–Lorentz electrodynamics, the stationary ether, negative optical experiments, and Lorentz’s work....

Habicht’s manuscript shows that, within a circle personally close to Einstein, special relativity could be presented through the more historical, Lorentz-centered, and electrodynamical framework that Einstein himself had adopted only two years after the 1905 paper. This manuscript reminds us that the history of special relativity was not transmitted only through the paper of 1905, but also through later acts of exposition, recollection, simplification, and reorganization.

When Lorentz's work is presented, it is really hard to understand what Einstein did that was original. In 1907, the theory was known as Lorentz-Einstein theory.

There is no reference to Poincare. This is bizarre, as it is known that Habicht and Einstein read and discussed some of Poincare's papers before 1905. Poincare's 1904-05 relativity papers were widely read in Europe. It is likely that Einstein knew about them in 1905. Even if he did not, he certainly knew in 1907, and Habicht knew in 1914.

The above paper says "the history of special relativity was not transmitted only through the [Einstein] paper of 1905". That 1905 paper was not even historically very significant. Much more important were Lorentz's 1895 and 1904 papers, Poincare's 1898-1905 papers, and Minkowski's 1907-08 papers. Special relativity as we know it came from those papers, and almost nothing came from Einstein's 1905 paper.

I think this Habicht paper is mostly interesting for what it does not say. If Einstein had some of the main ideas behind special relativity before Lorentz and Poincare published them, then Habicht would be the most likely witness. This Habicht paper could have credited Einstein with having some of the ideas before 1905. It does not. It is only known that Einstein told Habicht he was preparing a paper that “provides a new conception of time and space,” in May 1905.

Giacomini wrote other papers on relativity history, discussed here and here. The first paper was just updated.

Distinguishing Newton's First and Second Laws

Isaac Newton's three laws of motion are: (1) if F = 0 then a = 0; (2) F = ma; (3) actions have reactions.

The first law is a trivial consequence of the second. Why did Newton bother to state it as a separate law?

A new paper answers the question:

On the independence problem of Newton's first law Ido Yavetz, Ehud Aharoni Newton's laws of motion pose an apparent problem, sometimes referred to as "the independence problem": the first law seems to be a simple consequence of the second law, raising the question of why it was included as a separate law. Numerous answers to this question have been proposed in the literature. The main contribution of this paper is a novel answer which we call "the formal explanation." Unlike previous accounts it relies on mathematical formalism and argues that the definitions of Euclidean geometry necessitate the inclusion of the first law. We provide evidence in support of this claim. A second contribution is a comprehensive review of previously suggested explanations, which so far have often been treated in a fragmented manner, and a discussion of the plausibility of the various answers.

It turns out that many people have addressed this, and this paper has a new explanation.

Roughly, mathematicians did not always treat zero as a number. There are many examples in Euclid's Elements where an argument gets restated for the zero and nonzero cases, when a modern textbook would unify them. This paper shows that Newton's usage is consistent with Euclid's.

Newton did not have vectors either, so he had to state the directional components of F = ma separately. One version of his second law was:

”A change in motion is proportional to the motive force impressed and takes place along the straight line in which that force is impressed.”

This is similar to Euclid, as he only wrote about nonzero quantities being proportional.

Randomness Cannot be Proved

There are scientists who believe in strict determinism, and say it is necessary to a scientific view. Albert Einstein was such a believer. In particular, they do not believe in free will.

I have criticized this as wrong, and I also point out the opposite view is also wrong.

Roman Schnabel writes in a new paper:

Experiments testing Bell inequalities [12, 13, 15] have proven that there are physical events that occur without a causal reason, i.e., that are truly random. The proof does not make any assumptions, in particular, the proof does not use QT but only pure mathematics and reproducible experimental observation. Regardless of the validity of QT, random numbers can be certified as truly random using an experiment to test Bell’s inequality [27]. It is remarkable that a truly random process can be proven as such. The only assumption made here was the reasonable but, in principle, unprovable one that there is no entity that has predetermined every single microscopic event in the universe in advance [28].

Others also say this -- that randomness has been proved, except for the silly possibility of superdeterminism.

The whole idea is absurd. No experiment or theorem could possibly prove true randomness. At best they could only prove that something is not predictable with the data and methods available.

Since QT is complete, as proven by Bell’s tests, radioactive decay is a truly random process. There is no causal reason, no cause that leads to decay before or after the half-life. If you start observing a single atom at any point in time, there is a genuine, equally weighted quantum mechanical randomness as to whether the atom decays before the half-life expires or only afterwards. ... Radioactive alpha decay occurs in a truly random manner, i.e., without cause, which is referred to as ‘spontaneous’ in quantum physics.

Yes, radioactivity seems about as random as anything could be. K-40 has a half-life of about a billion years. Decay statistics are predictable. But no one can predict when a particular potassium atom will decay.

Maybe the problem is that we do not the physics and technology to measure the K-40 state well enough. If we knew exactly how those quarks were bouncing around the nucleus a billion years ago, maybe we could predict when one will bust out.

I know that seems absurdly impractical, but the Bell argument says something else. It seems to say that some things are unpredictable if we assume a local hidden variable theory. Maybe that K-40 nucleus has a state that can predict the decay without using classical variables. Bell's theorem does not say anything about what can be done in a non-classical theory.

The paper has also been criticized:

The paper repeatedly suggests that Bell-test violations show that quantum theory is complete, that hidden variables cannot exist “in general,” and that some events happen without causal reason and are therefore truly random. We think these claims go too far. Bell’s theorem rules out local hidden-variable models only under a definite set of assumptions, including locality or factorizability, measurement independence, and an ordinary Kolmogorov probabilistic framework in which the relevant joint probabilities and expectation values are defined. It therefore does not, by itself, settle every broader question about causality, ontology, or completeness.

PBS TV just dropped a video on atomic clocks, and Sean M. Carroll makes his usual dopey appearance:

[0:00] [Narrator] Albert Einstein famously said that he didn't believe God plays dice with the universe. ...

[Carroll] [0:42] The basic idea of quantum mechanics, the thing that we really struggle with to get our heads around even as professional physicists, is that unlike any other version of physics, quantum mechanics separates what happens in a system when we're not observing it from what we see when we measure it. ...

[1:50] That opens up a whole world of questions. You know, what happens to the observational outcomes that are not observed? What picks out which outcome is going to happen? This is still what we're thinking about today.

He would argue that what happens is that the observational outcomes that are not observed slip off into a parallel universe.

I do not see how Carroll's struggles with QM relate to atomic clocks, the subject of the video. Yes, atomic clocks depend on the quantum mechanics of atoms having discrete energy levels, but Carroll's philosophy rants are not relevant.

Update: Carroll also has a new podcast on Free Will and Levels of Reality. Both Carroll and his guest are determinists who believe that free will is an illusion. Philosophers call this compatibilism.

Aaronson's Latest Lecture on Quantum Computers

Dr. Quantum Supremacy has posted a new lecture:

Title: The TRUTH About Quantum Computing
Date: 2026-05-13 @5:00PM

Abstract: Yes, scalable quantum computing should actually work! Sooner than many expect, which will create a huge headache when it breaks the encryption currently used to protect the Internet. But no, we don't think quantum computing can do most of what the popular articles promise in AI and optimization and so forth. Come to this talk to learn about why!

Somehow he has gotten to be the academic authority on whether quantum supremacy has been achieved, so I always check his latest opinion. He seems to be getting more confident, but it is still just a prediction. He says it may or may not be achieved. And that not to trust your RSA keys.

And another one:

Title: Scott Aaronson - Theoretical Computer Science and AI Alignment
Date: 2026-05-14 @1:00PM

Abstract: I'll survey some areas where I think theoretical computer science, math, and statistics can potentially contribute to the urgent quest to align powerful AI with humane values. These areas include: the watermarking of AI outputs, mechanistic interpretability (including Paul Christiano's "No-Coincidence Principle," and succinct digests of the training process to aid interpretability), and theoretical guarantees for out-of-distribution generalization.

Also on the subject, Dr. Bee dropped a new one today that starts:

In Einstein’s theory of general relativity, the time an object experiences depends on its acceleration. But time in quantum physics works more intuitively – it’s a universal parameter experienced by every object in the same way. In a new paper, physicists say they want to use a special type of clock to test that difference. Let’s take a look.

0:00 Quantum physics says that objects can be in two places at the same time. A group of physicists now says that they can also be at two times at the same time, ...

No, quantum physics does not say that objects can be in two places at the same time. It does not say that cats can be alive and dead at the same time.

Quantum computers are often explained this way, but QM really just estimates probabilities for different place. Once you observe the object, it is in just one place. Her statement is like saying: probability theory says that a tossed coin can be heads and tails at the same time.

Aaronson is an AI enthusiast: [11:40] AI is what "I would regard as, you know, maybe the most consequential technology, uh, that humans are ever going to be building."

Update: At the other extreme, Brian Greene released a new interview with an AI skeptic. The guy says that the AI LLMs are just a stupid trick of little value.

Sean M. Carroll Lectures on Many Worlds

NewScientist just posted a marathon series of math lectures, but almost half of it is from a 5-year-old lecture, Sean Carroll: The many worlds of quantum mechanics, that is not even on math. He gave a similar lecture last year, The Many Worlds of Quantum Mechanics | Dr. Sean Carroll.

My disagreements start early. Carroll says:

Why does 7:14 quantum mechanics have this reputation of being so hard? Here's the answer. There's an amazing feature of quantum mechanics that was nowhere to be found in classical mechanics which says that what you observe when you look at a system is not what you see.

What you see and what is really there are two different things. There's a difference between what a thing is when you're looking at it and when you're not looking at it. What you can possibly see is much less than what really exists. This sounds weird. This sounds bizarre. like it's very very different than what we had in classical mechanics.

He is trying to say that reality is the wave function, which is not directly observable.

The word see and observe are synonyms. I do not get his explanation. Even in classical physics, there are things we cannot see. We cannot see the center of the Earth.

You might think that in that 14:03 circumstance, since quantum mechanics is the foundational theory for all of modern physics, you might think that the quest to understand quantum mechanics at a deep level would be recognized as one of the most important things we could possibly do in physics. The people who devoted their lives to these would be academic superstars. You would have different universities trying to steal them away with highpriced packages and salaries and so forth. and it would be the highest prestige occupation you could have in physics.

Sadly, no, that is not what we do. It is the opposite of that. We have adopted a strategy of denial where if you're a physicist and you think hard about answering these questions, you are labeled not a physicist or a physicist who is too old to do important work anymore and you're sent off to retirement.

That is because the important issues were settled in 1930. People like Carroll want to be paid to do research on many-worlds, and real physicists consider it a complete waste of time.

36:37 Another what I think incorrect objection is that this idea cannot be tested. Can it be tested Right? It's important in science that we not just have good ideas, but that we compare these important ideas to data. Right? that we experimentally probe our ideas and people say you've invented all these new worlds. How do you ever test that idea?

The response to that is, I didn't invent any new worlds. I just took quantum mechanics seriously. The entirety of the assumptions that go into the many worlds theory is there are wave functions and they obey the Schroinger equation. That's it. Everything else is a consequence, a prediction and implication of those assumptions.

And are those assumptions testable? Hell yes they are. Of course they are. Whenever we do a quantum mechanical experiment, we're implicitly testing the many worlds interpretation.

If you want to falsify the many worlds interpretation, remember that the prediction of many worlds is wave functions don't collapse. They never do. They appear to collapse because of decoherence.

So there is no way to test many-worlds. He says it is logically correct, so everything you see confirms it. That's all.

As he admits, we see wave functions collapses. They would contradict many-worlds, except that he does not believe that they are really collapsing. He calls that a test.

This is all nonsense. I don't know why anyone takes him seriously.

Carroll also has to answer this new survey:

World's largest ever survey of physicists, results & reaction

May 12, 2026
What do physicists really think about the biggest mysteries in the universe?

In this video, leading voices in theoretical physics come together to unpack the results of the Big Mystery Survey—the largest survey ever conducted of professional physicists, prepared in collaboration with the American Physical Society's Physics Magazine." Featuring reactions from experts like Sean Carroll, Niayesh Afshordi and ,Ghazal Geshnizjani we find out what physicists really think about topics such as: What should we think about fine tuning ? What is dark matter? What is dark energy? How should we truly understand the Big Bang?What the right approach to quantum gravity ?

Whether you’re a physics enthusiast or just curious about the universe, this conversation offers a rare glimpse into how experts think about the unknown.

It reports 11% for many-worlds, 6% for Bohm pilot wave. These are the two crackpot interpretations. It is amusing to see how the leading popularizers of theoretical physics have views that are rejected by most physicists.

Carroll also has a lecture in this newly-released video on the physics of time and the history of relativity.

[4:50] All you have to do is entirely jigger your thoughts about what space and time are. And in fact, it wasn't until two years later when Hermann Minkowsky, who was a mathematician who had been one of Einstein's professors, said, "You know, the right way to think about Einstein's theory is to say that space and time aren't separate anymore. To imagine there's one thing called spacetime, and different people, different observers moving in different ways through the universe will divide it up into space and time differently.

There's no objective true fact about when I snap my fingers now what's happening light years away. That's going to depend on who's doing the observing and who is doing the measuring. ...

It can all be explained very beautifully by imagining a single four-dimensional spacetime instead of separate space and time. Einstein himself was not impressed by this move. ...

And when Minkowsky says, "I have some new math that unifies space and time based on Einstein's theories," Einstein himself is like, "I don't need that. That's like extra mathematical nonsense."

Minkowski never referred to "Einstein's theory", but got that 4D spacetime from Poincare. That 1907 Minkowski paper cited Poincare's 1905 paper with the 4D spacetime.

Yes, people like to credit Einstein for spacetime, but that was published by Poincare and Minkowski, and Einstein rejected it.

Carroll wrote a whole book on relativity, so he surely knows the history. It is weird that he gets it so wrong.

Update: The survey has just been posted in a paper.

We present results from the Big Mysteries Survey, a large-scale survey conducted through the American Physical Society's Physics Magazine on foundational and controversial topics in contemporary physics. The survey provides a snapshot of physicists' views on issues in cosmology, black-hole physics, quantum mechanics, quantum gravity, and anthropic coincidences. A central finding is that several positions often described publicly as field-wide ``consensus'' views are, in practice, supported by much narrower majorities or by pluralities rather than majorities.

19% believe string theory is the best hope for quantum gravity. 30% believe that information dropped in a black hole is preserved in Hawking radiation. 24% believe in the dark energy cosmological constant. No consensus on dark matter. 51% believe in cosmological inflation.

The Most Cited Physicist Alive

Brian Keating interviews Juan Maldacena, The Most Cited Physicist Alive.

Juan Maldacena is a theoretical physicist at the Institute for Advanced Study whose 1997 paper remains the most cited in the history of theoretical physics.

We cover:
-why wormholes and quantum entanglement may be the same thing
-what actually happens to information when you throw something into a black hole
-the reason Hawking radiation accidentally gave cosmologists the equation that explains why the universe has structure
-whether science-fiction wormholes are ruled out by the laws of physics
-the one unsolved problem Juan says matters more than black holes.

The most important problem in quantum gravity is understanding the beginning of the Big Bang — not black holes.

And also the Holographic Principle, AI, complex numbers, etc.

Someone from 50 or 100 years ago would really be disappointed at what a dead-end theoretical physics. He rambles from one absurdity to another, with barely any relation to the real world.

The universe is not locally real

AI has risen Feynman from the dead, and explained Bell's theorem:

The Universe May Not Be Locally Real – Feynman Explains Bell’s Paradox

What if everything you believe about reality is wrong? In this mind-bending 23-minute exploration, Richard Feynman walks you through one of the most unsettling discoveries in the history of science — Bell's Theorem and the Nobel Prize-winning experiments that proved the universe is not locally real. From Einstein's stubborn belief in hidden variables to the groundbreaking work of Aspect, Clauser, and Zeilinger, this video breaks down quantum entanglement, the EPR paradox, and what it truly means when particles separated by miles somehow "know" what the other is doing — with no signal, no instructions, and no classical explanation.

It concludes:

[20:05] Let me bring it all together. Now we started with a simple idea that the world is locally real that things have definite properties that distant objects can't affect each other instantly. John Bell took this idea and turned it into a testable prediction. He said if local realism is true then the correlations between entangled particles must obey a mathematical limit. experiments culminating in the Nobel Prize winning work of aspect, Clauser, and Zelinger showed that the correlations exceed that limit. Local realism is violated. The universe is not locally real.

This means one of two things, or possibly both. Either particles don't have definite properties before they're measured, meaning reality isn't real in the way we thought, or distant particles can somehow affect each other instantly, meaning the universe isn't local in the way we thought. We don't know which. We don't have a complete agreed upon interpretation of quantum mechanics that resolves this.

I think we do have an agreed QM interpretation. It is the first option. Nobody thinks distant particles can somehow affect each other instantly.

The QM textbooks say that under Heisenberg uncertainty, it makes no sense for a particle to have a definite position and momentum at the same time. Particles only get definite values for position and momentum when they are measured, and you cannot measure both at the same time.

So yes, QM says particles don't have definite properties before they're measured. And I think that is what Feynman would have said, and not that this is an unsettled question.

Not everyone agrees with this, as maybe 10% of physicists subscribe to goofy interpretations like Bohm pilot wave or many worlds. But the mainstream textbook QM view is pretty clear on this point.

Hassabis likes a Wacky Book

Demis Hassabis says:

[25:41] well, my favorite book is The Fabric of Reality by David Deutsch. So, I think that still holds. I'd hope to answer the questions in that book with with the AGI. That's my post AGI work.

Hassabis is a genius, and certainly one of the world's leading AI researchers. But I am lowering my opinion of him, if this is his favorite book.

At first glance, the book is a celebration of modern science. It is enthusiastic about science, theory, the scientific method, explanations, and testability. It explains what science is all about.

Except that it is really anti-science. For all his talk about the merits to testing theories, he mostly promotes theories that cannot be tested at all. For all his talk about reality, he mainly believes in things that are not real.

His favorite ideas are Popper's falsification, Darwin's evolution, and many-worlds quantum mechanics. He raves about their explanatory power.

These are so bizarre, because many-worlds is just an unscientific belief that cannot be falsified.

Darwin evolution has some explanatory power, but the main principle -- survival of the fittest -- is a tautology. Fitness is defined as what survives.

Deutsch starts the book by attacking Steven Weinberg for saying an instrumentalist remark: "The important thing is to be able to make predictions". Weinberg's point was anti-geometry, not instrumentalism, and he later changed his mind when geometry proved crucial for theoretical physics.

For a more up-to-date description of Deutsch's opinions, see this new interview of him. He says that he now believes in free will, because we have the ability to create novelty.

[1:18:36] By the way, Bryce DeWitt said that he was sure that if Einstein had lived another few years, he would have come to like the Everett interpretation, because it's fully realistic.

I am pretty sure that Einstein would have hated many-worlds. Regardless, it is bizarre to say that many-worlds is "fully realistic". It is the opposite of realistic. Nothing is more unrealistic than postulating an infinity of parallel universes that have no interaction with ours.

Crypto Doom in 2029

Dr. Quantum Supremacy announces:

Will you heed my warnings NOW?

Holy crap … yesterday I was elected to the US National Academy of Sciences! ...

See, some of the most reputable people in quantum hardware and quantum error-correction — people whose judgment I trust more than my own on those topics — are now telling me that a fault-tolerant quantum computer able to break deployed cryptosystems ought to be possible by around 2029. ...

And I’d say that that makes my own moral duty right now ironically simple and clear: namely, to use my unique soapbox, as the writer of The Internet’s Most Trusted Quantum Computing Blog Since 2005^TM, to sound the alarm.

So, here it is: if quantum computers start breaking cryptography a few years from now, don’t you dare come to this blog and tell me that I failed to warn you. This post is your warning. Please start switching to quantum-resistant encryption, and urge your company or organization or blockchain or standards body to do the same.

He co-authored a paper on the threat to Bitcoin.

I say the opposite. There is no need to switch, and there is no chance of a break by 2029. [corrected] But I am not a member of NAS. You have been warned. Check back in four years to see who is right.

Update: I looked at the consensus white paper, to see if it endorses the idea that quantum supremacy is inevitable or has already been proved. It does not:

A few distinguished mathematicians, computer scientists, and physicists, including Gil Kalai, Leonid Levin, Michel Dyakonov, and Gerard ‘t Hooft, have taken the position that quantum computing is impossible in principle—that what seem to others like engineering difficulties will in fact never be surmounted for some inherent reason. This could be true, for example, if quantum mechanics itself were to fail as we tried to scale up QC, or if we were to discover some new principle on top of quantum mechanics, which guarantees the existence of noise that violates the assumptions of the Threshold Theorem and therefore can’t be corrected by FTQC.

It’s important to understand that this was never a mainstream view among experts, at least since the discovery of FTQC in 1996. The mainstream view could be summarized by saying that, if quantum mechanics were to fail, or some new principle were to be discovered that “screens off” or “censors” QC, that would constitute a once-per-century revolution in physics, and would be vastly more surprising and exciting than “mere success” in building a QC that worked as the theory has long predicted. From a physics perspective, QC working as predicted is the conservative option.

So believing in QC is the conservative option, but it might not be possible.