## Sunday, September 02, 2018 ... //

### A recent dissatisfied Weinberg's talk on QM

On June 3rd, Steven Weinberg gave this talk at Case Western University. The video is 46 minutes long, you may speed it up (up to 2x).

He's introduced as a hero. Weinberg is a hero who doesn't need an introduction. He's done amazing physics, he's been an important public intellectual. He's still doing physics.

Weinberg says that from the beginning, he knew that important physicists like de Broglie, Einstein, and Schrödinger were grumpy about quantum mechanics. It was tragic that they skipped the development of this exciting framework and its application on atoms, molecules, and elementary particles, among others. But then Weinberg tried to explain the essence of quantum mechanics and he found out that he couldn't do it to be personally satisfied. So he became another anti-quantum zealot, we're told.

Wait a minute. Why wasn't he satisfied? One may have scientific or intelligent reasons to be dissatisfied and one may have unscientific, irrational, or stupid reasons to be dissatisfied. Which kind of dissatisfaction is Weinberg's?

I think that the event is framed so that it's "polite" to think or at least say that it's the former. Too bad because it's clearly the latter, as I will discuss.

Fine. How does he formulate his complaints? In his terminology, there are three main views:

1. instrumentalist
2. realist
3. quantum mechanics isn't quite true
Here he inserts the first irrational intermezzo:
Isn't it a little disturbing that the very theorists who are most satisfied with quantum mechanics don't agree with each other what it means?
This is supposed to be half-serious, half-funny. There is a general problem with this rhetorical format inside lectures that are supposed to discuss serious matters. The speaker may always retroactively reinterpret or resegregate what was meant seriously and what was meant as a joke.

We should take the statement seriously. First, it is not really true that people who agree that the foundations of quantum mechanics are settled disagree with each other – they surely agree with each other much more than the critics of quantum mechanics. The latter group is collectively confused about really everything. Are there many worlds? Do the worlds split during observations? During reasoning? All the time? Is Born's rule fundamental? Derived? Psychological? Are there extra objective positions of particles? Do they disappear during annihilation? During observation? Is there an objective collapse? Does it occur frequently? Does it occur inside protons? Does it cause extra proton decay? Invent any question like that – you will get contradictory answers from big groups of critics of quantum mechanics because their dissatisfaction is really a result of their ignorance of modern physics and there are very many ways to be ignorant because you may guess the answers to all the questions that you don't really understand.

Second, even if you found many people or groups who say that quantum mechanics is settled but they would disagree about the details of the settled principles, no, it isn't disturbing. More precisely, it surely doesn't imply that all these physicists are wrong. Almost every important discovery was first made by one scientist or a group that was a minority. That discoverer became certain about the "new truth" while there were many other groups that still claimed to be "certain" about their "old truth".

Heliocentrism became settled while lots of others thought that astronomy was settled in the geocentric way. And similar comments apply pretty much to every important question in the history of science. Was that disagreement disturbing? I don't know what "disturbing" means. "To disturb" is a description of subjective feelings, not some objective scientific facts. What's certain is that the bulk of the scientific discoveries looked like that and had to look like that. People disagree about things. That doesn't mean that all of them are wrong.

So the existence of people who disagree with me about some matters – although they also say that quantum mechanics is settled – isn't disturbing to me. They're similarly wrong as the people who say that quantum mechanics isn't settled. What's disturbing to me is that the percentage of the people, even physics PhDs, who understand foundations of modern physics is going down. But it's not my fault or the fault of quantum mechanics so you just can't rationally use it as an argument against quantum mechanics or my explanations of that.
Then you begin to suspect something.
It's too bad if Weinberg really builds on similar šitty, sociological arguments and methodologies that have never worked. It's not science to draw far-reaching conclusions out of the existence of people who disagree with each other. If all people's opinions had to be "equally true", then the disagreements would be "disturbing". Well, they would be logical contradictions! But in science, people's opinions are obviously not equally true. Some statements are true and their negations are false! There is no contradiction.

It's interesting that the apparent Weinberg's assumption "all the people who disagree are equally correct" sort of mimics the anti-quantum zealots' wrong understanding of the wave function. They imagine that all the terms in a superposition "exist equally". But in reality, they just describe probability amplitudes for alternatives and only one alternative exists – we just don't know which one. The cat is alive or dead.

Instrumentalist approach

OK, he associates the instrumentalist approach with Bohr, Wigner, currently Mermin, and the Born rule. Weinberg says that "we want to know what's happening out there". But this is just the standard anti-quantum zealots' misunderstanding of the essence of quantum mechanics. Quantum mechanics says that all knowable things about the physical system must be obtained through a measurement whose outcomes are predicted via the Born rule. There is nothing else happening out there, at least nothing else knowable that is happening out there! So quantum mechanics does answer the question what is happening out there and the claim that quantum mechanics is incomplete in this sense is simply a lie.

Mr Weinberg could prefer science that says that something else is happening out there, or more things are happening out there. But this is just about his prejudices. Not even Weinberg has the power to dictate Nature how She should work. Quantum mechanics speaks a clear language, it says that observations and the corresponding collapses are the only things that are really happening out there, and the theory is therefore complete in this sense. Weinberg's psychological dissatisfaction simply isn't an argument against the validity or completeness of quantum mechanics. To find a problem with a well-defined theory, you need to falsify it. For example, Weinberg may try to falsify the claim by quantum mechanics "that only results of observations exist" by observing something that isn't a result of an observation! Good luck with that! ;-)

His second complaint is that these laws of quantum mechanics aren't impersonal and they should be. We wanted such impersonal laws ever since Newton. Now, what is the correct response to this complaint? It depends on the meaning of "impersonal". Quantum mechanics surely has no dependence on humans – on the homo sapiens species with the 4 limbs and 23 pairs of chromosomes. So to say that it is "human" or "anthropomorphic" in this sense would be a lie.

On the other hand, quantum mechanics requires a perspective, usually identified with an observer. A perspective that allows to define what the observations are and what they results were or will be. In the normal worlds, the application of quantum mechanics requires an observer. If this is enough for you to say quantum mechanics "isn't impersonal", well, then quantum mechanics "isn't impersonal"! It is "personal", if you wish, because it needs an observer to be applied. The observer may be a smart animal or an extraterrestrial robot – and less usefully, it could even be a big molecule (it's less useful because the molecule usually doesn't know how to deal with the information intelligently). But an observer is needed.

Is that a "problem" of quantum mechanics that it is "personal", and therefore "less impersonal" than what Isaac Newton hoped for? You gotta be kidding, Dr Weinberg, right? Isaac Newton was just wrong – his theory wasn't the final word – when it came to these questions. Is that so shocking that Isaac Newton's words weren't the final ones? Isaac Newton also wanted laws with the absolute time and absolute space – and in fact, he described those equally religiously as the "impersonal nature" of the laws. The absolute space had the spirit flying everywhere. Relativity showed that simultaneity was relative. Can we conclude that relativity is wrong because it contradicts some "ideal" that Isaac Newton had? Are you serious? Do you really want to define science as the eternal worshiping of Isaac Newton's views as the final word? Because that's exactly what you seem to be saying.

Quantum mechanics simply is "personal" in the sense that it requires observers – in the role of axioms, it depends on pre-existing information about the nature and outcomes of observations that are extracted through subjective perception. This "personal" is an equally established adjective as "heliocentric" or "evolutionary". People who say that the Solar System should be "geocentric" or the origin of species should be "creationist" or the laws of physics should be "impersonal" are simply wrong. To declare that the laws must be "impersonal" in the sense of independent of observers is an indefensible – and for more than 90 years disproven – dogma that is totally analogous to the claim that the Solar System has to be geocentric and the species were created in a week. You can say and demand geocentric, creationist, and impersonal laws but it just shows your stupidity and bigotry and it doesn't matter whether you got Nobel prizes. Science knows that the right adjectives are "heliocentric", "evolutionary", and "personal" with these definitions.

David Mermin tells Weinberg that these features don't bother him – but they bother Weinberg, we heard. They don't bother me, either, because there is simply no tension between the predictions of quantum mechanics (logical, internal tension or tension with observable facts). I think it's obvious that the reason behind the "bothering" is nothing else than an irrational prejudice. There's no difference between the "bothered Weinberg" and the "bothered Inquisition" that just disliked the heliocentric models or the Christians who hated evolution. There's no logical and scientifically valid way to justify either of these dissatisfied sentiments.

Realist approach

In this section, Weinberg repeats some delusions held by the many-worlds ideologues. The realist interpretation automatically leads to "many worlds". Well, that's rubbish. The mathematics of Schrödinger's equation implies one wave function that is evolving in time. There is absolutely no natural, unique, or canonical way to divide the wave function to many terms (there is a hugely infinite number of ways to decompose a state vector into a sum) and that's why it's completely wrong to say that the single wave function describes "many worlds".

It's still a single wave function. In proper "instrumentalist" quantum mechanics, it's a single wave function remembering all probability amplitudes for all possible outcomes of measurements in the single world. When an observation is made, the wave function abruptly changes – collapses to an eigenstate of the just measured observable. Certain folks want to deny this discontinuity associated with the measurement. These "many worlds" people throw the only physically viable interpretation of the wave function to the trash bin – so what they're left with is a mathematical equation from quantum mechanics, but without any usable connection to the world around us.

It's completely idiotic to deny that the wave function has to change in the wake of an observation.

OK, Weinberg totally uncritically repeats lots of delusions by the many-worlds ideologues, e.g. that the "many worlds" unavoidably follow from Schrödinger's equation – which is complete nonsense, as I have mentioned. He also remains completely silent about dozens of other problems with this ideology. For example, he does mention that Schrödinger's equation should be linear but the wave-function-like-another-classical-wave pictures don't have the slightest justification for the linearity of the equation. He isn't bothered. In this case, a scientist must be bothered because the nonlinearity of the equation is actually a generic prediction of that "realist" picture that is easy to be falsified experimentally. (There's no known way to falsify the claim by quantum mechanics that the theory is "personal".)

A positive thing is that Weinberg realizes at least one thing – that there is no way to derive the Born rule from that picture. In fact, he seems to understand my favorite observation that there can never be such a derivation because the proposition saying "the probability is given by a specific formula" must appear at a certain moment. In a derivation, it can't appear spontaneously – so there has to be an axiom that says something about the value of the probability. There has to be something like Born's axiom somewhere or another, equivalent or more complex, axioms of the same kind. All the people who believe that Born's rule may be derived noncircularly from nothing can't add two plus two.

Weinberg is also bothered by "too many worlds", an aesthetic problem. Sure, "too many worlds" would be ugly but as I have mentioned, it's just wrong to say that a single wave function "splits" to many worlds. There's no (and there can be no) well-defined rule to dictate what such a split should be. There's also no problem that would be "cured" by such a conclusion and no other evidence that such a splitting takes place.

Objective collapse

At 13:35, he reveals his first proposed "kind of modification of quantum mechanics" that he would like to see. A superposition $\alpha\ket{\uparrow}+\beta\ket{\downarrow}$ should spontaneously change to $\ket\uparrow$ or $\ket\downarrow$ after some time. It's some generalized "objective collapse" or Ghirardi-Rimini-Weber picture.

Now, why would he believe such a thing? Why would he consider such a modified theory superior? Or "beyond quantum mechanics"? The complex superpositions are clearly an important new concept in quantum mechanics – just like the relativity of simultaneity in the special theory of relativity. What's going on is that Weinberg clearly has a deep psychological problem with the very existence of superpositions. That's shown by the fact that he would consider a theory without superpositions – and even a theory that gradually erases superpositions – to be superior.

This view is absolutely irrational. It's the geocentrism once again. Indeed, a postulate of quantum mechanics is that all superpositions of a collection of states are equally allowed as the individual states in the collection themselves. The space of allowed states is a linear complex Hilbert space – a vector space closed under addition and linear combinations. I would claim that this isn't even something that the misinterpreters of quantum mechanics could have started to obfuscate. The linearity of the Hilbert space of allowed pure states is a totally elementary property of the whole formalism. You can't have anything similar to quantum mechanics without that linearity of the Hilbert space.

Imagine that a superposition of "up" and "down" gradually drifts either to "up" or "down". Why would he consider it better? A superposition of "up" and "down" for an electron's spin simply describes the spin polarized along a general axis in three dimensions. Take this axis $\hat a$. Why should the axis of electron's spin spontaneously approach $\hat z^+$ or $\hat z^-$? What is so special about the $z$-axis? Clearly, such a drift would break the rotational symmetry of the laws of physics. Noether's theorem makes it clear that it would break the angular momentum conservation law, too. And indeed, if a nonzero probability of having the angular momentum "down" could suddenly disappear, it would correspond to a violation of the conservation law, or at least a "nonzero probability of such a violation".

Such a modification of quantum mechanics would cause infinitely many similar problems. The real problem is that there is never any preferred basis of the Hilbert space describing a generic complex enough physical system. One can't really divide the Hilbert space to "superpositions" and "non-superpositions" (i.e. preferred basis vector) simply because there isn't any preferred basis of the Hilbert space. The whole Weinberg's way of reasoning is totally wrong – it's a full-blown anti-quantum zeal. To have a fundamental problem with superpositions means to be a full-blown anti-quantum zealot. Superpositions are clearly as vital in quantum mechanics as a frame-dependent time coordinate in special relativity. If you can't accept these rudimentary features, that's too bad but your real understanding of these theories is bound to be zero in that case. At most, you may pretend that you understand these theories. But because you don't really believe what you say, you don't actually understand them.

But imagine that there are some laws that push the superpositions towards some preferred basis, which you assume to exist (although it cannot). What would you gain out of it? Why would you consider such a theory better or more satisfying? It still makes absolutely no sense whatsoever. The resulting theory would be some strange hybrid of classical physics and quantum mechanics that would contain the puzzling features and inconsistencies of both frameworks.

You know, at least in the short run, when a molecule evolves, you clearly need to allow the general superpositions. We can verify that after some time, molecules may be found in all superpositions – the corresponding probabilities of outcomes exactly match the conjectured superposition state. If you (irrationally) associate superpositions with some "ugliness" or even "potential inconsistency", well, then Nature has this "ugliness" and "potential inconsistency", at least at short distance scales and time scales, because the relevance of superpositions may be directly experimentally observed. If you order superpositions to "fade away" after some time, great, but they are still not eliminated from the laws of physics.

Weinberg spends a lot of time in his talk by talking about some Ghirardi-Rimini-Weber collapses. He often stops by saying that he doesn't know whether some numerical value of an energy scale or another parameter is OK or not. He probably doesn't want to know! If he did proper science, like he managed to do hundreds of times in his life, he could quickly conclude that nothing can work well in these theories.

If some Ghirardi-Rimini-Weber "spontaneous collapses" of the wave functions are too frequent, too vigorous, affecting too many bound states etc., then the theory would predict new "flashes" and new types of a radioactive decay that don't exist. So all these parameters must be pushed to weak or otherwise unobservable corners. But when they're unobservable, they will lose the ability to change anything about the meaning of quantum mechanics, too. Small enough objects evolving for short enough times will evolve just like in quantum mechanics. You will need the same superpositions and the same need to interpret the superpositions probabilitically.

A very specific example: Imagine that you assume, like Weinberg probably does, that there is something wrong about quantum mechanics because it allows the brain to be in a superposition "I saw an explosion" and "I saw no explosion". A generic initial pure state evolves to a superposition. You either accept it and admit that the probabilistic interpretation of the superposition is needed to explain why you only had one of these feelings; or you introduce modifications that are strong enough so that they mess up with your brain before the realization and cause some explosion themselves. What are these new explosions? It's something that produces at least a bit of new uncertainty (or an erroneous bit) per a kilogram of the brain stuff and per second – otherwise it couldn't "cure" the superpositions in time. But that's clearly a fast enough new process that you simply could observe otherwise if it existed.

You just can't eat a cake and have it, too. Either you modify quantum mechanics so that your new theory produces new effects in the currently observable best experiments – in which case you're screwed because QM has worked perfectly in all of them which excludes all the modifications. Or your modifications of quantum mechanics will be so weak that the whole experimental situation will be described by laws that are effectively indistinguishable from quantum mechanics – so the modifications can't help to solve the (psychological) "problem" that you have with quantum mechanics.

What Weinberg and other "modifiers" of quantum mechanics doesn't seem to understand is that his psychological problems with quantum mechanics aren't about some details. They're about the bulk of quantum mechanics, about O(100%) of the theory. That's true for the value of the predictions as well as the states themselves – he has a problem with "superpositions" and almost all vectors in the Hilbert space are "superpositions". If he has a problem with O(100%) of the theory, he can't cure this problem by changing the theory by O(0.00000000001%). Small corrections just can't be enough to fix his problem – these two things have a totally different magnitude. He has to modify quantum mechanics by a huge amount, by O(100%), and then he gets a theory that is in a totally obvious disagreement with the observations, just like fundamentally non-relativistic theories contradict tests of relativity and fundamentally creationist theories contradicts tests of evolution. Your alternatives are totally hopeless.

All the anti-quantum zealots are constantly fooling themselves, trying to be persuaded that their totally qualitatively, game-changing modifications of quantum mechanics miraculously allow the theory to work just like before. But that's complete nonsense – a sign that this movement has turned into a religion that has abandoned all rationality. They're like a creationist who hates evolution. So this creationist may be shown some "micro-evolution" after a few generations. It's an experimental fact but he tries to "confine" this fact and argue that in the long term, the origin of species is actually governed by creationism (this creationist example is actually as real as the anti-quantum zealots). Similarly, a hater of relativity might say that after a little while, some effects adjust all the clocks and restore the universal time for everybody. A geocentrist could say that after a year, the illusion that Mars orbits the Sun will go away and Mars will start to orbit the Earth as it should.

How could the hybrid theory work? It cannot. And what would one gain from a theory that works like evolution in the short term but reduces to creationism in the long run? Nothing. If there is something conceptually ugly or bad about evolution, this hybrid theory would still have this ugliness! On top of that, it would be a mixture of totally different theories. Weinberg's proposed modifications of QM are equally stupid as my other examples. He's clearly prejudiced and prefers any modification of physics that makes it "less quantum" according to some sentiments. But this clearly proves that he is not driven by any rational or meritocratic criteria at all. He is on a mission that starts with the commandment "quantum mechanics is bad and you will always spit on it and invent arbitrary fog as long as it makes quantum mechanics look bad". There can exist no other justification for his efforts.

So his Lindblad equations that I have previously discussed are an example of this totally misguided, quasi-religiously motivated anti-quantum effort. You may generalize the evolution equation for the density matrix to a Lindblad equation by making some additional parameters nonzero. But proper quantum mechanics is still a special case of your new theory. It means that the amount of difficulties you have to clarify – and the amount of psychologically unpleasant things – couldn't have decreased by adding the new Lindblad terms.

Around 18:00, he discusses locality and entanglement and says that you can "change psi at the distance but not rho at the distance". This is a sloppy way to express a valid point but a very sloppy and misleading way, indeed. What he means is that the reduced density matrix for the subsystem B isn't changed by the measurement of subsystem A – assuming that you don't learn the result of the measurement of A. I have given the proof many times. This is a way to show that the laws of quantum mechanics are local.

He says that "you can change the wave function at a distance". These sentences make it sound as if the "density matrix" laws were more local than the "wave function laws". But that's ludicrous, of course. The evolution laws governing the density matrices and wave functions are exactly the same – they can be derived from one another. So the difference between the apparent "degrees of locality" isn't between pure and mixed states. Instead, it is between reduced and non-reduced states. And due to entanglement, the mixed states are needed to define a reduced state!

It means that mixed states are needed to discuss the properties of subsystems separately (we need to calculate the reduced density matrix, and there's nothing such as a reduced pure vector for a generic – entangled – state of the composite system), and therefore mixed states are needed to directly show that measurements don't act at a distance.

It doesn't mean that in terms of pure vectors, the measurements act at a distance. Instead, the "defect" of the pure states is that they don't allow us to discuss subsystems separately at all! (At one moment, Weinberg also makes the beginner's error and talks about the wave functions of individual particles. Sorry, there's only one wave function for the whole system – in principle, for the whole Universe. There aren't "many" wave functions for parts.) So pure states are a "wrong language" to discuss locality in general. They don't prove nonlocality in any sense.

Before 20:00, Weinberg says a lot of additional confusing things about the relationship between the pure states and mixed states. He discusses the possibility that the "density matrix could be something to tell us something about the pure state" and so on. But all these ideas make it clear that he still assumes that the wave function is a collection of classical degrees of freedom. But it is not. The wave function and the density matrix play a totally analogous role: they contain all the probability amplitudes (and therefore the probabilities) of outcomes of any measurements that are knowable about a physical system.

When the observer has the maximum precise knowledge allowed by quantum mechanics (including the uncertainty principle), e.g. when a complete set of commuting observables was just measured, the system is described by a pure state. When there's some additional uncertainty, the system is described by a mixed state i.e. a density matrix. Both possibilities are equally "real" and equally "possible". It is completely wrong to imagine that the density matrix is a generalized probability distribution while the pure state is more real or objective. They're equally probabilistic. Both psi and rho are complexified quantum generalizations of the concept of probabilistic distributions.

I said that even if the extra Lindblad coefficients generalizing Schrödinger's equation were found to be nonzero, they couldn't possibly "clarify" anything about the aspects of quantum mechanics that Weinberg calls dissatisfying. But I think it's really pointless to look for such coefficients. They can't be nonzero. One simple way to see it is that in principle, every physical system may be brought into a pure state. There's no way to derive an "equation for a pure state" from the Lindblad equation for the density matrix. So the Lindblad deformation would mean a pure-to-mixed evolution. That would amount to information loss. If that existed in principle, we could really observe it all the time. The probabilities would leak almost everywhere. An atom couldn't conserve its angular momentum for too long. But it does conserve it. So you really need the angular momentum to have a special relationship with the time evolution – it must commute with the Hamiltonian and the Hamiltonian has to tell the full story about the evolution.

Just like there's no Schrödinger's equation for a pure state corresponding to the Lindblad equation for the density matrix, there could be no Heisenberg picture, either. We couldn't say that the observables themselves are evolving. There would be no Feynman's formulation of such a theory, either. Even if you believed that those deformations aren't an inconsistency, why would you think that they are an improvement of anything?

You know, in theories we use, we have e.g. the algebra $SU(2)$ of rotations generated by the angular momentum. The three generators precisely commute with the Hamiltonian. The commutators are precisely zero, it's pretty and clean, it agrees with everything, the evolution and rotation transformations are exactly given by the generators. Weinberg wants to pollute these things by some extra deformations. Why would he think that the newer theory would be more satisfactory or "better" in any sense? It makes no sense. It would be polluted and less consistent. The number of problems (or perceived problems) could only increase.

At 24:00, Weinberg says that "atomic clocks aren't clocks – they are frequency standards". That's silly. Atomic clocks surely are clocks, "clocks" and "frequency standards" aren't mutually excluding in any way, and all clocks that have the best precision at a given moment of time have been used as "frequency standards". To some extent, "good clocks" and "frequency standards" are synonymous.

OK, Weinberg wants to use some atomic clock to find the nonzero Lindblad coefficients. Too bad, he doesn't want to complete the calculation to see whether the parameters may be both measurable by his method and relevant to "cure the alleged problems" of quantum mechanics. Clearly, the answer is No.

At 34:00, he promotes Roger Penrose's totally irrational claim that gravity collapses the wave function. You know, "pieces" of a wave function packet aren't two objects. They are alternatives for two possible locations of one object. So they cannot gravitationally attract each other because "their" existence is mutually exclusive. The gravitational field around a cat depends on (is entangled with) the cat's being alive or dead. So the gravitational field is found in a superposition of possibilities, too. Penrose's gravitational-psi-collapse are pure crackpottery.

More generally, Weinberg asks what distinguishes microscopic and macroscopic systems. You can see that it's the same frantic effort to "contain" quantum mechanics – just like the creationists want to "contain" evolution. The big picture has to be classical and creationist. Sorry, there is obviously no universal way to define a boundary between microscopic and macroscopic systems. Even if you found such a boundary, it would be irrelevant because quantum mechanics applies both to "microscopic" as well as "macroscopic" objects! This is really a statement totally analogous to the proposition that Darwin's theory works for "microevolution" and "macroevolution", too. Only bigots are obsessed with the efforts to reduce the applicability of evolution or quantum mechanics. Bigotry is the only justification of such efforts. There is no rational justification.

At 34:30, Weinberg makes a "groundbreaking" observation that macroscopic systems have exponentially close energy levels. It's been understood from Day One of quantum mechanics. In fact, the reason why the close levels have something to do with the classical limit was understood by Bohr 5 years before the birth of quantum mechanics – it was called the correspondence principle. Bohr clearly articulated it in 1920 but he had already used it in the old model of the atom in 1913. But the correspondence principle is a way to see why an approximate, classical theory works as a usable enough approximation. It is not a way to raise doubt about the validity of quantum mechanics.

I think it's terrible when people try to sound original by vaguely talking about totally standard things that have been known even before quantum mechanics was born, 100 years ago, and they were understood correctly pretty much right away while the vague babbling of Weinberg's only has the goal to spread some incorrect understanding of the nearby levels.

Needless to say, the exponentially tiny separation of energy levels is something that subfields of quantum mechanics use all the time today. People doing black hole information puzzle are extremely aware of the exponentially fine spectrum of black hole microstates – and what it means for the representation of various operators etc. It's been discussed in quite some detail on this blog, too. Why would Weinberg believe that some vague comments that the dense levels could perhaps have something to do with the emergence of the classical limit are insightful? They're not insightful at all and the amount of evidence (hiding in the fine levels) that there is something incomplete about quantum mechanics is zero. If you give a talk where A,B appear in the same sentence, it does not mean that A implies B.

His last sentence is that physics need a guiding principle how to go beyond QM – that would tell us more than the general principles. Well, we do have such a guiding principle, it's the general postulates of quantum mechanics. And everything he does simply violates this guiding principle.

In the first question, Weinberg says that neutrinos are not useful for his efforts. In the second question, Weinberg says that he's not Penrose and gravitons are useless for his efforts. The third question asks why the absence of superpositions doesn't rule out "realism". Weinberg makes some incorrect statements that we always have to calculate with "ensembles of pure states". It's just not true. We may calculate with mixed states directly. Or without states, in the Heisenberg picture.

He repeated the correct comments that the many-worlds have to add a Born-like axiom, anyway. Then he laughs when he asks: "Would there be no Nature without people or physicists?" Sorry but quantum mechanics is a universal theory of the Universe and everything in it. So in principle, observers are needed to meaningfully apply QM to any question, including those about the Universe and its long-term evolution. You misunderstand the big picture if you find this fact laughable. This laughing is the same stupid mocking as when creationists at our college laughed that Darwin predicted a bird to change to a squirrel while sitting on a tree. Well, not really. Darwin made predictions that agreed with all our observations, e.g. that we don't see a realtime transmutation of birds into squirrels. But in principle, changes like that do occur in the long term, whether you like it or mock it, and in the same way, quantum mechanics demands observers for a precise formulation of all questions, including questions about the long-term evolution of the Universe or the number of spiral arms of a galaxy. The information about the specific number of spiral arms doesn't "exist" in any scientifically meaningful way if there is no observer. Period. To laugh here means to show one's ignorance of modern physics.

At 40:50, 't Hooft asked a question that was really an incoherent statement. He wants to say that all the laws are the same they have always been, we just don't know the initial state. As I mentioned many times, he really ceased to understand quantum mechanics completely – e.g. the existence and meaning of the relative phases. His resolution (I mean the fineness of his analysis) is just so bad that he no longer distinguishes a diagonal density matrix from a pure superposition state. Weinberg says that he has no overlap with 't Hooft's thinking and repeats that Born's rule needs "people" which is nonsense. Weinberg agrees with 't Hooft that all problems with Born's rule could be explained by the ignorance of the initial state. That's great but this picture of theirs cannot make quantitative predictions of the probabilities while proper QM can – so theirs is clearly an inferior theory. Well, not a competing theory at all.

It's just frustrating to see these Nobel prize winners as they discuss some totally basic undergraduate material from quantum mechanics just like two confused, mediocre undergraduate students would after the third lecture of an undergraduate course.

I am afraid that we're entering a dark age and when it's over, perhaps in the 22nd century, the future historians will analyze what will have gone wrong in the early 21st century and whether it was possible to stop the collapse of the mankind's discourse into a bunch of medieval superstitions. They will know that the evolution was linked to the post-truth era, endlessly softening education and Academia, and politically correct taboos. Weinberg could have stopped and was expected but actually helped the mankind to sink deeper into this profoundly irrational antiscientific bigotry.