The Wolfram Physics Project: A One-Year Update
writings.stephenwolfram.com
writings.stephenwolfram.com
Also this part definitely conjures up a certain series of sci-fi novels popular on HN (and elsewhere):
"So one of the unexpected implications of our models is that there can be dimension fluctuations in our universe. And in fact it seems likely that our universe started essentially infinite-dimensional, only gradually “cooling” to become basically three-dimensional. And though we haven’t yet worked it out, we expect there’ll be a “dimension-changing cosmology” that may well have definite predictions for the observed large-scale structure of our universe."
Dimensional warfare, baby!
Anyway Wolfram gets a lot of hate but honestly I'm glad that some smart dude is using his financially-independent time to go on wild theoretical excursions into what interests him. Living the dream.
For a given set of observations, the simplest set of posits that explain those observations are more likely to be true as there are fewer ways to "bake in" the observations into the model.
The surprise, and perhaps why it continues to be a pursuit, is just how far it's taken us. My reading of history suggests that in the 17th century, scholars assumed that the inquiry would either run into a dead end, or merge with theology. Neither has happened yet.
It certainly doesn't prohibit humanity from dropping any of those rules and exploring the universe in alternate ways, so it's not a great loss if physics sets some boundaries for itself.
I agree! I truly appreciate this man. And he seems kind and well-intentioned when I've watched his videos.
The silly mental imagery that I have of him, which I don't intend in a mean-spirited way: that his ego is such that he goes well beyond naval-gazing, beyond even sniffing his own rear, and then leans so far that he's looking up at the stars and the universe from a very different perspective than most everyone else ;)
As it turns out though, huge swaths of the universe's behavior can be produced without all that individual specification. There are massive holes in our understanding, but at pretty much every turn for the last few centuries we've gotten better at predicting more stuff with fewer equations. Does that mean it'll continue? No idea! But that's why there's a real sense that a simple universe is more 'probable' than any particular complicated one. Now maybe no simple universe can possibly fit ours, in which case we're just kinda fucked.
You're correct regarding the extreme unknowns, since of all possible universe-rules most are complex. However, should a simple universe-rule correspond with our own, it is more likely to be correct than any single complex universe-rule.
More importantly, the whole testable hypothesis thing? That relies heavily on the same intuition. Test results mean nothing if the universe can cheat, and in fact the universe can cheat. But we disregard the billions of possible cheating universes eating the same slice of the probability pie in favor of the non-cheating universe, should it exist.
Occam's razor does not apply to evolutionary systems because it would require analyzing every step of the chain like every mutation that increased the probability of survival in the environment it occurred in. The path from archae to amphibians, for example, was very convoluted, complex, and nonsensical, let alone specific examples like amphibians that became hooved mammals that crawled back into the ocean to become marine mammals. The simple explanation would be a fish that evolves lungs to outcompete gills, not an auroch that learns to swim!
Fact is, we don't even know if the universe is an evolutionary system or not. We don't know whether our universe just happened or if we exist in a universe that is the conclusion of a bunch of previous universes that fizzled out of existing or didn't support life and evolved into this one or if we exist in an infinity of universes with an infinite permutation of the laws of nature. We don't even know what "simple" really means in this context.
We just don't know. O(explanation) = ??? and P(explanation) = ??? for every value of `explanation`.
I'm confused about the evolution example; it seems to me that evolution is something of a triumph of Occam's razor. After all, for all it's myriad complexity the simple rule of natural selection explains pretty much all of it.
My gut says that the analogous component to 'natural selection' would be the laws of physics, but I might not be understanding that last bit. Are you going for more of an anthropological principle thing?
The theory of natural selection is the simplest explanation to our observations of the fossil record, genetics, etc., not an explanation for why marine mammals evolved from hooved land mammals instead of fish directly - except in the most trivial sense of "it is because it is" (a la anthropological principle). Why did nature select for that path instead of the more direct one? We can't even begin to formulate that question for the universe and its laws as a whole.
The laws of physics in this case are the mammals and the process of natural selection of universes (if there even is one) is a complete mystery because we don't have anything to compare it to establish probabilities. Whether that is even a valid question to ask is still in the realm of philosophy and Ivory tower theoretical physics.
In particular, Solomonoff Induction is the most information-efficient method of inducing computable rules from observations https://en.wikipedia.org/wiki/Solomonoff%27s_theory_of_induc...
It uses Bayes rule to update hypotheses based on observations; and its initial hypothesis is the "universal prior", which assigns probabilities to observations based on how likely they are to be output by a randomly-generated computer program (more specifically, an infinite sequence of uniformly-random bits, interpreted as a self-delimiting, Turing-complete language). The universal prior is uncomputable, but has been proven to dominate all computable probability distributions. Whilst computable approximations of the universal prior, like https://en.wikipedia.org/wiki/Speed_prior , aren't as information-efficient, they are more time-efficient, and approach the universal prior when given more time to run (they're identical in the limit of using infinite time).
This is relevant for Occam's Razor, since the universal prior itself is dominated by the shortest program that outputs an observation, i.e. the best prediction we can make about what any computable system will do next, is to look for the simplest description of (AKA shortest program outputting) everything we've seen it do so far, and make a prediction based on what that program does next. When we say the shortest program "dominates", this means that even if every longer program agreed on some different prediction, the shortest program's prediction is still more likely (it's related to the Zeno-like summation https://en.wikipedia.org/wiki/1/2_%2B_1/4_%2B_1/8_%2B_1/16_%... ).
Hence Occam's Razor is proven to be the optimal method of predicting the behaviour of any computable system.
Whether or not it's a "law of nature" depends on how we define that phrase philosophically; however, I would say the Church-Turing thesis certainly counts as one, and others agree; e.g. from https://existentialtype.wordpress.com/2012/08/09/churchs-law
> This is all just another instance of Church’s Law, the scientific law stating that any formalism for defining computable functions will turn out to be equivalent to, say, the λ-calculus when it comes to definability of number-theoretic functions. (Ordinarily Church’s Law is called Church’s Thesis, but for reasons given in my Practical Foundations book, I prefer to give it the full status of a scientific law.)
(I highly recommend that book BTW; preprints are available online)
The "physical Church-Turing thesis" is a physical law akin to relativity (it states that a particular mathematical formalism corresponds to "the real world"; despite the latter having no precise definition, and hence being mathematically unprovable, but empirically testable) https://en.wikipedia.org/wiki/Church%E2%80%93Turing_thesis#V...
Hence we have a physical law which tells us the universe is computable (or, if you prefer a slightly weaker statement, that no computable method can distinguish observations of a computable univese from those of an uncomputable universe, given a finite number of observations; hence even if "the real world" were uncomputable, that wouldn't have any impact on our observations)
Combining these two: Occam's Razor is a provable result of a rich and rigorous mathematical theory, and applies empirically to the universe via accepted physical law.
As far as testable hypotheses go, we can perform a grue/bleen experiment quite easily ( https://en.wikipedia.org/wiki/New_riddle_of_induction ). Here's one to try at home:
- Take any existing simulation S of a physical system P which we can compare against experiment, e.g. a computational fluid dynamics simulation of our kitchen sink.
- Ensure that a "prior run" of P agrees with a "prior run" of S, e.g. starting from the same initial conditions and lasting one minute. If they don't agree, go back to step 1 and alter the setup.
- Alter the code of S to get a new simulation S', which behaves identically to S during a prior run (e.g. the first minute), but differently when allowed to run longer.
- Try to shrink both S and S'. We can't do this exactly (since Kolmogorov Complexity is uncomputable), but we can use systematic approaches, e.g. refactorings, inlining, applying compression algorithms, etc. which should hopefully be unbiased.
- Our null hypothesis is that whichever ends up shorter (S or S'), its behaviour following the "prior run" is more likely to match the real behaviour of P than the longer program (of course neither might match; but if one of them does, we'd predict it to be the shorter)
Gas equations are emergent from statistical mechanics are emergent from quantum mechanics, Optics are emergent from quantum mechanics, gravitation is emergent from relativity. Now sure, relativity is a good deal more complicated than Newtonian mechanics, but saying it makes better predictions does it a disservice. It explains all the normal stuff, and also a bunch of weird stuff, while still being pretty simple and IMO way more elegant than what came before.
If you have two explanations for a phenomenon, one of complexity level 5 and one of complexity level 10, Occam’s razor would say choose the simplest.
But if you have one explanation of complexity level 10, Occam’s razor does not say that you should reject it because a simpler explanation must exist. It makes no statement that hypotheses have to be simple.
Last paragraph, last sentence. Now between you and me, I'm suspicious that we'll just keep describing more universe with simpler models forever, but you're right that Occam's razor doesn't guarantee that.
https://en.m.wikipedia.org/wiki/Kolmogorov_complexity
Edit: so the argument would be something along the lines of: the less we need to describe the universe, the more probable it is that it exists (due to the fact that, given that it does in fact exist, the more complex it is, the more different universes could exist in its place). That's of course assuming there is one universe! If we assume that there can be multiple universes then there is no easy way to posit that a simpler one is more probable.
Let's assume that the universe operates on one to an infinite number of rules. Could be one, could be 137, could be a Googleplex. Let us imagine that the number of rules when "rolling" a universe is equally distributed.
The question I pose to you is ... where on the that list do you think it would be more fruitful to start? You could throw your hands up and say "Could be anywhere!" but that isn't a great use of your time.
Instead, you posit a single rule, you try it out, and then there's an exception. This rule worked half the time! That is excellent. This is more likely to occur at the lower end of the rules-scale than the upper end of the scale, after all, simply because you have more things to try the more rules you have.
Could be that only a single rule exists, but it is not likely out of infinity, and so you start there because you can find out the exceptions faster than anywhere else.
Then you move on to two rules. Two rules seems to explain quite a lot! But not everything ...
You can witness this in how classical electromagnetism was constructed from first one rule with some exceptions, and then extended. In a similar manner, you can see how we need more sensitive and complex equipment to test for the existence of exceptions, like that muon magnetic moment requiring rather a lot to pick up.
Following this process will lead you to a lower bound on the number of rules, but never an upper bound, because eventually you are deconstructing galaxies to build particle accelerators and experiments take a billion years to run. We'll eventually say "Good enough for 99.9999999999999% of the cases" and leave it at that.
As an example, I had a database I had to investigate with just ... nasty data. Really filthy. Multiple things compounded into a single field, and so on and so forth. One field in particular I needed to decode for the process. So I tried a regular expression and it mostly worked on the first ten. But then it failed.
And I looked at the case that failed and scanned for other incidents, and now I had two expressions to try. And I looked at hundred instances, and it mostly worked, but there were some exceptions.
This went on until I had ten sets of "rules" to cover all four million of that field. It's not elegant, it probably isn't the fastest, but it starts at the bottom and works its way forward.
In fact, in mathematics, the opposite is usually true - there are many more complex objects than simple ones, so given a random object, you should expect it to be complex rather than simple. For example, given a real number, it is infinitely more likely that it will be irrational rather than natural.
1: https://www.scientificamerican.com/article/do-we-live-in-a-s...
2: https://academic.oup.com/pq/article-abstract/53/211/243/1610...
It would go like this: let's imagine that you were an omnipotent being. Then, one of three things must hold: either we will never become omnipotent beings, or omnipotent beings are not interested in creating a universe where creatures can become omnipotent beings, or we are living in the creation of an omnipotent being.
In fact, I find that the idea that the universe could be entirely simulated in a portion of itself is actually very similar to the idea of omnipotent beings existing. They are about as plausible to me (not very) and have similar implications.
In fact, if a human could run a simulation of the entire universe, they would be exactly equivalent to the idea of god for that universe, and could decide to enact whatever religious story they chose for their creations, including stuff like planting fossils that look millions of years old in a 6000 year old Earth, mutilating cows and causing crop circles, answering prayers for rain and floods, pulling the sun on a chariot across the sky etc. They could even decide to run beautiful simulations almost forever for people who do good after they die, and scary torturous simulations for those who did bad.
So, if you believe in a simulation, you should also seriously consider that religious texts could be real and passed down from the runner of the simulation, whose name might as well be YHWH or Brahman.
As to god, well, if our universe is simulated, probably there's a god or gods who made that simulation. But, unless this simulation is human-centric, we're just a speck of dust and I don't have any reasons to suspect that we're somehow special and those "gods" interacted with our ancestors. So no, from any kind of scientific PoV, our religions are not trustworthy for me and religious texts are just interesting historical records. And if our civilization is the point of interest for that entire simulation, those texts are still weird and not trustworthy. It's should be relatively easy for "god" to provide an unquestionable miracle every few centuries to keep our beliefs, but there were not any trustworthy ones. Also there are too many of religions and they don't exactly look similar, so that's another issue, as it's not possible to choose a correct religion.
The argument only works if you have an infinite or at least unbounded simulation tower. Otherwise you can't jump from the observation that there could be simulations to the idea that there must be simulations or a Great Filter.
My point with the religious stuff was to show that you can re-state any religious idea into a 'scientistic' interpretation if you accept the simulation argument. This doesn't mean you should believe in Thor any more than you do now, but to point out that if we accept the simulation argument we must also accept that all religious texts are valod objects of scientific inquiriy, as they could be presenting events that are non-physical in our simulated universe but physical in the upper universe. Even contradictory ones could be true, as the simulation could have been unwound and replayed differently.
To put it another way, a priori, without any predefined physical laws, anything is exactly as likely to exist. There's no reason, a priori, to say that a universe with an infinity of dimensions could not appear out of thin air, if you don't first postulate some physical laws. A universe where entropy always decreases is mathematically possible. If you lived in such a universe, your intuition may well be the reverse of ours - you may believe that more complex universes are more likely to appear than simpler ones, since objects with more states are more likely to appear out of thin air in such a universe.
How can you even know this? Do you have at least ONE model for the universe? If that model is complex, how can you show that it works? I think you're just operating on figments of your imagination.
Overall what I'm trying to say is that arguments that 'simpler universes are more likely to exist than complex ones' pre-suppose that there are some laws of physics that apply 'outside' the universe, apart from logic/mathematics.
By the way, I'm NOT trying to argue that we shouldn't be looking for a simple set of laws for our own universe - i believe thath is indeed the only option that makes sense, but only from a pragmatic stand-point - it is the only search strategy that guarantees we will arrive at the minimal possible explanation, regardless of how complex or simple that ends up being. Perhaps we'll find a single theory of everything with no physical constants, perhaps we'll end up with 100,000 different fundamental interactions, this part we can't predict.
There is a huge amount of complexity represented by very few distinct types of things.
An enormously complex set of rules is more likely to produce bizarre edge cases/inconsistencies/glitches. So far, we arguably we haven't observed any.
These rules we are addressing here weren't constructed by a human. We don't even know if they were constructed at all. It could be a perfect set of rules despite the complexity.
I'm more on the side of simplicity just because it seems that way, but technically and logically speaking... there is no evidence at all against what the OP said.
From the POV of 19th centure physics, the double-slit experiment was quite the glitch.
And one could argue that the consequence of that glitch did not exactly simplify physics.
This is similar to QM and the Standard Model. QM has a very elegant set of simple rules of motion that describe how particles interact. But there are infinitely many possibile universes that behave according to Schrodinger's equation. To see how our universe behaves, you also need the Standard Model, with its dozens of particles and relationships between them and constants and dimensions.
In contrast, String Theory only has one 'particle', the string, and a set of rules. You can derive the Standard Model entirely out of those rules apllied to strings, which some physicists find much more appealing. They even derive gravity the same way, though only for a universe that is not like our own. However, this may all be mathematical sleight of hand, as the rules precisely encode the constants of the standard model, and make no correct new predictions otherwise (they do predict supersimmetry, but that has been experimentally refuted so far, making the rules retreat to higher energies). There's a good chance Wolfram's models will turn up the same way - a new mathematical formalism that, if you plug in all of the constants of the Standard Model, predicts the Standard Model precisely.
> That's irrelevant, as the program can't be derived from the rules. To understand the behavior of the program (the universe) you need to know both the rules (physical laws) and the source code of the program (physical constants/quantities).
This does not remove the fact that the universe could be built from a very few simple laws. (my first and only point in my previous comment)
As with softwares, even if we do not know the 'source code' we can infer it. How? By designing experiments. There are then three possibilities:
1. Every experiment confirm our theory, in this case we believe with a certain degree of certitude that the model is a good approximation of the reality,
2. An experiment give results that do not correspond to our model, then we need to change or improve our model,
3. Two experiments on the same phenomenon give contradicting results. We finally found a bug, the behavior is undefined, maybe our universe collapse, maybe we are in a giant simulation and we found a limits of higher floating point numbers.
> This is similar to QM and the Standard Model. QM has a very elegant set of simple rules of motion that describe how particles interact. But there are infinitely many possibile universes that behave according to Schrodinger's equation. To see how our universe behaves, you also need the Standard Model, with its dozens of particles and relationships between them and constants and dimensions.
I do not see exactly why it is relevant but I would add that the standard model in not enough, lots of things exists in physics where we cannot define an experiment to decide what is the (our) reality. But we do not really care because the realities are somewhat equivalent.
> In contrast, String Theory only has one 'particle', the string,...
I will not address this comment. I think your point is that physics can be defined experimentally or mathematically and the mathematical version has more chances to be wrong because the minimum set or rules are more arbitrary?
> There's a good chance Wolfram's models will turn up the same way - a new mathematical formalism that, if you plug in all of the constants of the Standard Model, predicts the Standard Model precisely.
Sure there are chances that string theory and Wolfram's theory are a game of the mind and they are fundamentally flawed. But it is also true that if you chose well your first principles and they are easy enough, if both theories can be used to infer a wide variety of phenomenas, there is a good chance that there is an isomorphism between the first principles and they are equivalent in essence, so why bother, go for the system you believe in, hope to make great discoveries and one day, maybe, your theory will be merged with a better theory, but it is rarely lost. Theories are adapted and improved ;)
I have this sensation that people often hope to reduce physics to a few mathematical laws and hope to find they don't even need any constants to measure for those laws - that they'll just pop out of symmetries or other purely mathematical properties, only being left with experiments to confirm that the universe does indeed conform to this mathematical model.
While that would be beautiful, I don't see any reason to expect it to be that way - to be able to prove that the properties of the universe (e.g. The charge of the electron, the mass of the quark) are logically necessary and not mere observations.
(and what is a non mathematical object anyway)
Physics and mathematics have evolved a lot and current research questions cannot anymore be discovered through simple experiments, one of the reasons is because an experiment would be too costly. But not always, some things are intrinsically difficult to observe or even impossible. In comparison mathematics is a 'cheap' and extremely powerful tool.
When a theory is developed enough, we can start to design again experiments to test it. This is simply an example of the scientific method which is applied. In practice, even if a theory is purely mathematical (which is never the case as usually those maths are developed by trained physicists in classical theories) its power is assessed by how much of our reality we can infer with it.
People mix frequently how the science make discoveries with what is true, in the case of Wolfram's theory, if it is enough to prove and infer interesting things, it is already worth studying it. If not for the sake of physics, at least for the sake of mathematics :)
That is, the Standard Model interpretation is 'the universe just happens to have these particles with these properties and these interactions', while the hope behind String Theory is that we'll be able to say 'the universe consists of this single thing that obeys this mathematical law, and so because of such and such mathematical properties it can look like either an electron or a quark or a tau particle, and when its in the electron state it must have this mass, and when it's in the graviton state it will have that mass/spin/color and no other is possible'.
I used the term 'non-mathematical object' for any part of the theory whose properties can't be derived and must fundamentally come from experiment, such as the mass of the fundamental particles in the Standard Model.
> I have this sensation that people often hope to reduce physics to a few mathematical laws and hope to find they don't even need any constants to measure for those laws - that they'll just pop out of symmetries or other purely mathematical properties, only being left with experiments to confirm that the universe does indeed conform to this mathematical model.
Why we could not create an abstract system that when we plug in the constants we derive all the rules of our reality? And even better, this system would let us learn more on other universes by plugging different constants! If our abstract system does never contradict our reality when we plug the constants, then it is indiscernable from our reality, so we can accept it as a good model of our universe, even if it has very abstract foundations.
> While that would be beautiful, I don't see any reason to expect it to be that way - to be able to prove that the properties of the universe (e.g. The charge of the electron, the mass of the quark) are logically necessary and not mere observations.
You can draw the line where you want, some people take those as constants and develop theories from them, other people try to prove that those constants make sense in a greater theory. The same thing happens in maths, you can start with the real numbers, or even only with the naturals, while some other people will construct the numbers from first principles.
Why? Why not! And we also have real Incentives :
The beauty is when you discover a set of rules where everything we know works without contradiction. If we do not have this kind of system, then the logic would simply collapse and mathematics would have no validity whatsoever.
That said, it's interesting to me that the universe is actually much simpler than it appears, just based on the idea that the information contained in a region of space is proportional to the surface area, _not_ the volume [2]. The larger the area of space we look at, the more the discrepancy grows between how much information we think there is (using our human 3 dimensional intuition with x,y,z for each particle) vs how much there actually is (based on Bekenstein Bound) [3].
[1] https://www.youtube.com/watch?v=yjOV3xgydXE [2] https://en.wikipedia.org/wiki/Bekenstein_bound [3] https://en.wikipedia.org/wiki/Surface-area-to-volume_ratio#/...
"Therefore, as a cell increases in size, its surface area-to-volume ratio decreases. [...] If the cell grows too large, the plasma membrane will not have sufficient surface area to support the rate of diffusion required for the increased volume. In other words, as a cell grows, it becomes less efficient. One way to become more efficient is to divide; another way is to develop organelles that perform specific tasks. These adaptations lead to the development of more sophisticated cells called eukaryotic cells."
Source: https://bio.libretexts.org/Bookshelves/Introductory_and_Gene...
The latter I can understand intuitively as a function of scaling metabolic rates for different organism sizes. Specifically, there's some power-law relationship between growing volume and area[1], which means you require proportionally more metabolism (energy production) as you get smaller because your surface heat loss is smaller, to maintain the same heat rate as a larger animal.
I think there might be a rough analogy/relationship to space if we think of heat loss in terms of increasing entropy/loss-of-information, and the higher probability of low energy/low-information states in a volume over time? But, I don't have enough physics knowledge to know if this is the right way to think about it...
[1] This is the rough intuition, but the actual exponent is not consistent with square-cube rate: https://en.wikipedia.org/wiki/Kleiber%27s_law.
BTW, check out this video (especially starting at 10:15) for a better explanation than I gave for the Bekenstein bound [2]
[1] https://en.wikiquote.org/wiki/Andrew_S._Tanenbaum [2] https://www.youtube.com/watch?v=Ab8JIzckx_M
That type of code, in the hands of a gifted programmer (not java) using a decent programming language (not java) can usually be subsumed to less than a 100 lines.
There is therefore a good case for looking for simple rules :D
So it makes intuitive sense that their behavior would be incredibly simple. There is not room to store and execute 10,000 lines of Java in an electron.
Very far from a proof, but this argument makes sense to me.
If that's true, there is a third "rules" entity present when two particles interact. And that's the one we should detect and study.
Or maybe I'm talking nonsense. But that's how I think about these things.
The laws of the universe could just be abstract concepts that describe how these particles interact with each other.
But I don’t think that we should “take some of his claims with a grain of salt”. That implies reserving credit, which is appropriate if he is making claims about facts. In fact we should entirely discount his claims about how the universe works or the efficacy of his approach to physics until he can demonstrate that we should pay attention. The burden is entirely up to him, and the only yardstick is predictive power, of which his theories have zero.
The work Wolfram, Gorard, et. al have been doing on this is incredibly fascinating and mind bending. And I love that they are live streaming and posting videos of their working sessions so that non-academics can see how this kind of research is done.
Regardless of whether this approach leads to new experimental predictions, it's important in that it's a fundentally computational model of physics that reproduces general relativity and quantumechanics.
The concepts of multiway graphs and causal invariance are also valuable as ways to model and study nondeterministic computations; they turn such computations into objects with an underlying structure. And there are absolutely consequences for computer science and mathematics, in general, in studying such models.
1)Almost any mathematical abstraction can be used to match existing understanding of physics. Graphs, or groups or cellular automata. It is not wrong to try different kinds of those abstractions and there is nothing wrong with Wolfram's ideas in this department.
2) What matters is if you can deliver new insights from your model - be it general relativity, quantum mechanics or hypergraphs. Here his theory is extremely weak. Only two numerical values have been offered and they are experimentally non-testable as they are 50 orders of magnitude smaller than even corresponding Plank units (!) There are plenty of fringe theories with better predictive ability.
Edit: and in case it needs to be said, we now have had about 30 years of research into string theory and related theories without a single testable prediction. I agree that testability is critical, but every theory of physics beyond the standard model is lacking there.
In particular, the three documents linked at the top of https://www.wolframphysics.org/technical-documents/
The reason they suspect that quantum computers will not provide an inherent speedup has to do with the cost of "completion", i.e. measurement, which is explicit and finite in these models.
The bog standard response to anyone who attempts to do something new. Thankfully Wolfram has a thick skin.
To make a comparison that a lot of people use to justify fringe research, Einstein came up with general relativity mostly by himself, and it was a fringe concept, but it came with a set of predictions that could be verified with observation. Eventually those predictions were tested, and his idea “won”.
Until Wolfram’s ideas come bundled with refutable predictions that deviate from the standard model, it’s a smoke without a fire.
To restate the comparison, Einstein wasn’t writing pamphlets, going on speaking tours, or writing books about his “universal laws” for years before actually presenting the idea in a refutable form.
Einstein waited until he saw flames, to raise the smoke alarm, I’d that makes sense. Wolfram has been raising the smoke alarm for a while, but not providing any images of the fire.
It’s not a very good look, IMO, but reasonable people can disagree.
I thought his goal was a sort of parallel reconstruction (to unnecessarily abuse a shady law enforcement term), and so I'm not sure the above captures goals well...?
Disclosure: biochemist and technologist who happens to get a kick out of Wolfram's meanderings.
So far, I think (please someone correct me if I'm wrong) he's claimed the latter but not the former, and has it as a medium-term goal to claim the former but he is "not there, yet, but close".
I don't know if he's actually produced more efficient methods of computing results, if anybody knows of links, that'd be interesting to read.
I also think that he's heading down an interesting path (parsing the universe as a graph / cellular automata), it's an interesting choice of primitives. It does seem to be a Dual of fields, though, so I don't believe it can give us any more predictive power than modern Field Theory, but I guess we'll see!
Also remember it took decades before all of Einstein's 1905 predictions had been commonly accepted.
Does than mean we shouldn't judge Wolfram until 2031 at the earliest? I don't know...
Oh wait, he never actually did.
Of course one feels inadequate in an article supposedly about fire if all we ever get presented with are different shades of grey smoke and "as it turns out it's beautiful" and "I've researched this smoke since the 80s".
Hey, we give string theorists tenure! But for real, while it doesn't seem like this is going anywhere, I'm not sure I wouldn't be convinced by something similar that was. If I walked out of my room with a set of equations that made few additional assumptions and allowed for derivation of both relativistic and quantum equations, it kind of seems like that would be good evidence. Like even if all the predictions are the same as those two systems, the fact that it's much simpler would give it Occam's approval.
Now would such a single simple system fail to make new predictions? Probably not. There are holes in our understanding, and a unified theory couldn't really be simple and still tiptoe around those holes. But in the hypothetical scenario that there really was a simpler model with the same predictions, I wouldn't be opposed to it.
The CTO gave me his copy of a "New Kind of Science". I don't know why. It sat on my office shelf for years, hopefully convincing a few people that I was thoughtful and intelligent. It sure looked good, inside and out.
Yeah, there is: both are kooks.
So...Steven Hawking?
In the meantime, can this theory reproduce or derive something as simple a CKM-matrix?
To me it seems like his entire exercise is akin to re-inventing mathematical physics with, say, type theory instead of set theory, or to handle the surreal numbers rather than merely the reals, or likewise to restrict only to the computational reals.
Any one of these exercises might lead to newfound theoretical understanding. But that’s far from obvious at the onset, and my inclination is that these would be enormously complex undertakings that would only muddle the physics rather than clarifying anything[1]. Regardless, if you were halfway through this project and managed to pull something like the Einstein field equations out of your mess of math, it hardly proves anything. You started with a very general model that can express a great deal of general mathematics, and all you did was pull some specific mathematics out of it. It’s particularly egregious with Wolfram because nobody is impressed that cellular automata can express a great deal of mathematics which looks like physics.
[1] Perhaps some approach will make certain computations easier (e.g the type theory approach and certain problems with group representations) but that’s different from new theoretical understanding. The point is that none of these approaches seem likely to give a testable prediction they differs from the “standard” formulation.
He seems to have generated his own empirical playground whose output he can squint at, rather like a haruspex, and see parallels to ... pretty much anything he wants to see.
But the top down approach I don't think has ever worked. Revolutions usually start when a small experimental problem is solved in a novel way. And from that a new model is built.
I don't see that here, only delusions.
Does anyone know how this framework handles ‘spooky action at a distance’ ?
And it's also a deeply unscientific place. The option to be wrong and change course is at the heart of true scientific inquiry.
"A new kind of physics?" Not again, please no.
[1] https://writings.stephenwolfram.com/2020/04/finally-we-may-h...
I am just so put off by Wolfram's puffed up, bafflegab riddled and pompous writing "style" ... I feel like I must be missing something though?
It does make one wonder if "people" means "my employees, on pain of firing"
This is basically the definition of an ad hominem.
He may be an asshole, but so were many other great scientists (Newton comes to mind).
That does not have any correlation to the value of the work.
This article repeatedly mentions a "machine code" for the universe. What does that mean? Does he simply mean that computational power and advancements in machine learning allow us to model physical phenomenon better and to understand how complex ideas are actually emergent behaviors arising from some fundamental pieces we already know about? It seems like that's the type of exploration performed in other articles, such as the one on consciousness (https://writings.stephenwolfram.com/2021/03/what-is-consciou...). But I find it strange to see such explorations mixed in with what feels like marketing for the Wolfram language in here.
Regardless - ultimately, what Wolfram is promising when he uses phrases like "a truly fundamental theory of physics", is a Grand Unified Theory (https://en.wikipedia.org/wiki/Grand_Unified_Theory). I am not certain that he's found it, because I feel his past works suffer from delusions of grandeur, and also because I don't think humans have all the necessary inputs/data/sensors to actually characterize the universe fully (for example what is dark matter). Without those pieces, how do we know that what we have is "truly fundamental" versus just another emergent layer on top of some more fundamental piece?
Or maybe I am wrong and there is lurking genius in all these words and someone smarter than me or someone with more time will see it more readily.
He's is own worst enemy in many ways, as he undermines some actually interesting work with an outsider complex, a pathological need to self-aggrandize and a tendency to claim others work as his own.
Who's evaluated these ideas?
As some automatons are computationally complete, and some of the functions we use to describe physics are computable, I would not be surprised that some descriptions of the world coincide.
Wolfram's idea is, that all particle behaviour, etc., can be described this way, instead of the field equations we're currently looking for in physics.
I don't know whether Wolfram has a model of information propagation that is so much more useful than wave functions and field equations but they will agree for the simple cases.
(Shannon) information theory is also closely related to probability, and quantum theory is a form of probability theory (e.g. https://arxiv.org/abs/quant-ph/0101012 ).
Hence it's not entirely out-there to think that an informational/computational approach like Wolfram's might bear fruit. In fact, I was hoping that someone smart(er than me) would investigate such connections a bit deeper; if only to rule them out! Note that "causal set theory" seems to be going that direction too (although I'm not overly familiar with it) https://en.wikipedia.org/wiki/Causal_sets
I also find it promising that ideas from quantum computing (qbits, circuits, teleportation, protocols, etc.) have gone from a quirky engineering footnote (in the 90s), to become a core part of how many physicists now think about the world.
He's actually searching for an even loftier goal - a Theory of Everything [0] that is also a Grand Unified Theory (that is, a single mechanism underlying gravity, electromagnetism, the weak force and the strong force). Good luck to him, though I see no reason to imagine he will succeed.
A statement like "If A implies B, ~B implies ~A" is true whether or not there is a god, or a prime mover, or even an A or a B. Doesn't matter; the statement is still valid.
Now, you can expand that statement with all sorts of niceties that are equivalent: "If (A and B) implies B, ~B implies ~A". You're limited only by your imagination here, and the number of true equivalences is infinite.
The list of true statements might expand at a consistent rate across a series of complicated, embiggening rules...reminiscent of the Big Bang? And maybe within that expansion you could see a train of As moving to the left sometime, like a photon going the speed of light, which might sometimes beget a "reaction" when encountering an expression like (~B or A) when they meet, that sends a train of Bs heading out the other direction.
If such interactions are complex enough, eventually huge structures can be created, like self-reproducing entities that eventually develop consciousness. To those entities, there is nothing imaginary about what they perceive and experience. Their universe has been created, simply because it was possible.
Disclaimer: I could be all wrong as far as Wolfram's view, but this happens to be my own and I think Wolfram is getting at the same thing.
That makes me wonder when there's a differential equation, when solved numerically it's also solved in this very basic iterative way.