Are electrons too big to simulate? with Jonathan Gorard
HTML-код
- Опубликовано: 16 дек 2023
- How big are electrons compared to the hypergraph?
Is one electron formed of 10 nodes, or 10100 nodes?
And if it’s 10100 nodes, might it prove impossible to simulate an electron on any computer we can possibly imagine?
When I asked Jonathan Gorard this question, he took us on a tour of the scales of the universe, from the Planck scale to the Hubble scale.
He revealed how the Wolfram Physics Project’s early estimate of the scale of the hypergraph was based on a tower of rickety assumptions.
And he explained how the Wolfram model might connect with particle physics regardless of the disparities of scale.
-
Jonathan Gorard
• Jonathan Gorard at The Wolfram Physics Project www.wolframphysics.org/people...
• Jonathan Gorard at Cardiff University www.cardiff.ac.uk/people/view...
• Jonathan Gorard on Twitter / getjonwithit
• The Centre for Applied Compositionality www.appliedcompositionality.com/
• The Wolfram Physics Project www.wolframphysics.org/
Concepts mentioned by Jonathan
• Planck scale en.wikipedia.org/wiki/Planck_...
• Hubble scale en.wikipedia.org/wiki/Hubble_...
• General relativity en.wikipedia.org/wiki/General...
• Fluid mechanics en.wikipedia.org/wiki/Fluid_m...
• Quantum mechanics en.wikipedia.org/wiki/Quantum...
• Quantum Field Theory en.wikipedia.org/wiki/Quantum...
• Scattering amplitudes en.wikipedia.org/wiki/Scatter...
-
The Last Theory lasttheory.com/ is hosted by Mark Jeffery markjeffery.com/ founder of the Open Web Mind www.openwebmind.com/
Prefer to listen to the audio? Search for The Last Theory in your podcast player, or listen at lasttheory.com/podcast/053-ar...
Kootenay Village Ventures Inc.
Any chance you could release this whole interview? It’s killing me in tiny chunks.
Yes, sorry, it's taking me a long time to edit and release, but I'm nearly there! I'll be releasing the last excerpt or two in the next couple of weeks, then I'll put the whole thing together and release it as one long video. Thanks for your patience!
A thought just occured to me, could it be that there's a limit to the number of graph changes that can be done in each update, and that's why time slows down near massive objects, because there's literally more stuff to process, just like a game slows down when it has to render and simulate more?
Ah, yes, thanks Karl. That's an interesting idea.
The Wolfram Physics Project initially focused on applying as many updates as possible across the entire hypergraph at each step. I confess I never found that a convincing way to think of the evolution of the hypergraph (though it did generate beautiful images!) I prefer to think of updates as happening one at a time.
That feeds into the idea of the multiway graph, where we don't ask how many updates are happening simultaneously, but trace different possible paths through the multiway graph, depending on which updates happen in which order.
This sort of reverses - and also validates - your question. It means that time _emerges_ from these updates. Humans, as perceivers of time, are also in the hypergraph, so time is how fast we see the _universe_ updating relative to how fast our _brains_ are updating.
I don't know the answer to your question, but for sure, it's an intriguing possibility!
@@lasttheory not sure I would say time emerges, possibly better to say that as beings encoded in the graph we have an experience of things changing? When discussing graph changes, it's tempting to say, this piece of graph changes, then that piece, etc, so there's an implicit assumption of time being baked in, it's so difficult to think about these things as temporal beings! I tend to think of some computing device that updates the graph one step at a time, but perhaps it's better to think of the graph as a structure that's just there, and it changes itself? But even that assumes some sense of time...
@@karlwest437 Yes, I think time _is_ baked in: sequential updates to the hypergraph do seem like they're the basis of a very simple concept of time. And yes, we can get into trouble when we think of an actual computing device (see my video _Where's the computer that runs the universe?_ ruclips.net/video/m6pI9ndsEK8/видео.html ) so better, I agree, to think of the applications of the rules just happening.
This is my point about free neutron decay creating both dark matter and dark energy.
The charge free neutron exerts no Coulomb forces. It is hermetic. But when the charges separate they become radial, dark energy, extending a far as they can.
In neutron decay cosmology The universe has shock absorbers since the neutron decay has the soft play of a gas.
I don't understand where the rules in this (or any theory) are encoded. For example, how does an electron "KNOW" it's an electron (and how an electron behaves etc)?
That's an interesting philosophical question. You can ask it of _any_ theory. In Newton's theory of gravity, for example, how does the Earth _know_ that it's supposed to be gravitationally attracted to the sun, and how to follow an elliptical orbit around the sun as a consequence?
Because it's a philosophy question rather than a physics question, there's no precise answer. Physics can tell you _how_ the universe behaves. It can even give you lower-level explanations, for example of how that elliptical orbit arises from the geometry of space-time. But it can't tell you _why_ the universe behaves the way it does.
For more on this in the context of computational rules, take a look at my _computer that runs the universe trilogy_ ruclips.net/p/PLVwcxwu8hWKng3gsKKzmSw56ehIjENyh9
Thanks for the question!
If I compare an atom @ 10^-10m to a hurricane @ about 5.3 * 10^5m, I get one hurricane to be about 5.3 * 10^15 atoms to cross the distance. Again, I am going to make a large assumption and use the Ether Vortex model that places two vortexes spinning around each other with the similarities of atoms in a hurricane, I get the size of a single Ether particle to be around 2(size of an electron @ 10^-18 times the number of atoms it takes to span a hurricane @ 5.3 *10^15) or 2(5.3 * 10^-33) or (1.06 * 10^-32).
So near about 10^-32m range is where I would ballpark it at. Any other bets.
Ah, that's an interesting analogy, and an interesting figure, thanks Martin! Maybe we should start a sweepstake here?
Why would the Planck length represent an upper limit? As I understand it the Planck represents the lowest end of the length scale where normal notions of length cannot be supported by current theory. AFAIK, current theory are not *required* to work down to that length. Could the graph edge length not be much larger? (eg
Good question, thanks, Paul. Yes, the Planck length represents the limit where our current theories kind of give up, because general relativity and quantum mechanics don't work well together at that scale. But that doesn't mean that we _have_ to give up there.
The reason it's an _upper_ limit for the Wolfram model is that we know that general relativity and quantum mechanics _do_ work well at scales _above_ the Planck length, with their assumptions of continuous space and time. That suggests that space and time _aren't_ discrete at these larger scales, so a hypergraph-based model can't work unless the scale of the nodes and edges is at least a little smaller than the Planck scale.
If planck scale is the maxima for the size of atoms of space what is the minima? How do you get to the 10^-100? What is the constraint that leads to the latter?
Good question. And the answer is... we just don't know. I didn't get from Jonathan the full stack of rickety assumptions they made to arrive at the hypergraph scale 10⁻¹⁰⁰ m, but it sounds like they're _so_ rickety that the number is more or less meaningless. So there's no minimum.
@@lasttheory I suppose since we are thinking about computation, there really could (must?) be any number of stacked machines simulating the next. And I suppose the question, then, is how deep one can dig, in the sense of at what level observations start having predictive power or at what scale interventions start having the potential of influence on future events.
@@tusenkluster I don't think we're talking about _real_ machines here. The Wolfram model is a _computational_ model, but it's just a _model:_ the universe evolves _as if_ it were being computed, but that doesn't mean that there's a real _computer._ See my series on _the computer that runs the universe_ ruclips.net/p/PLVwcxwu8hWKng3gsKKzmSw56ehIjENyh9 for more on this.
You're spot on to ask about how deep you can go and still see influence and predictive power. It seems likely that most of what happens at the level of individual nodes and edges of the hypergraph has no influence whatsoever on what we observe on a human scale.
Thanks for the comment!
@@lasttheory What I meant is that any graph in this most abstract sense could in turn be emergent from some underlying graph. It could go on forever. I know Wolfram et al. have ideas of the totality of graphs. To my mind, that would include a totality in depth too. Just some wild thoughts though, thanks for the reply.
@@lasttheory The a0 of MOND betrays that the level is 1.5e-95 m (as predicted by the direct product of sporadic groups). Alternatively, Hubble tension points to the same number. A single electron should contain 5.8e37 parts (or 2pi the number). Just at the edge of possible simulation capacity.
Quantum particles are just a name for those processes that occur in quantum fields. Why is it not about reproducing the behavior of quantum fields? Maybe the number of hyperedges does not matter for an electron, but only their behavior? That is, first you need to prove or disprove this statement .
Yes, thanks Yarov. Of all the areas of physics that the Wolfram model needs to connect with, particles seem the furthest away. There's much work to do to move from speculation to solid statements, as you say.
@@lasttheoryHow about quantum computer?
@@PPP-on3vl Quantum computers are well explained (as far as they're well explained at all!) by conventional quantum mechanics. It's possible that the Wolfram model might some day change the way we think about quantum computing, but first it'll need to more fully explain quantum mechanics itself!
A response to Scott Aaronson's recent interview in the RUclips channel Theory of Everything might be interesting.
Thanks, I'll take a look at that! I've not heard anything from Scott Aaronson that's more than an off-hand dismissal of the Wolfram model based on technicalities that may not hold in the new model, but if there's something new in the TOE interview, I'll be sure to put it to Jonathan! I appreciate the heads-up.
Just to follow up quickly here, I've listened to Scott Aaronson's critique on the Wolfram model on Curt Jaimungal's Theories of Everything ruclips.net/video/1ZpGCQoL2Rk/видео.html
Scott's critique seems unchanged here, and I confess I don't see how it makes contact with the work Jonathan Gorard has done on the Wolfram model.
On quantum mechanics, Scott seems closed new approaches to the century-old conundrums of the theory. It's odd. I don't think _anyone_ thinks that quantum mechanics is _solved_. The Copenhagen interpretation where the "observer" (whatever that is) collapses the wavefunction with an "observation" (whatever _that_ is) just doesn't make any sense (as Schrödinger so clearly illustrated with his cat thought experiment). So sure, we could just give up and say la la la la nothing to see here... _or_ we could try different interpretations, e.g. using the multiway graph. It's certainly true that the Wolfram model is a very long way from a complete explanation of what's going on, but the accepted theory is, too. _Not_ keeping an open mind to new approaches is a certain way to never make any progress.
On general relativity, Scott mischaracterizes Jonathan's work. He seems to criticize Jonathan for deliberately working to derive Einstein's equations from the hypergraph. But, well, Einstein's equations are clearly correct on a large scale, so it would be crazy _not_ to work to derive them from the hypergraph! What does he expect Jonathan to do, try to derive the _wrong_ equations? Scott's criticism is that you "have to actually mathematically derive these things from some simple starting postulate." But that, if I understand correctly, is precisely what Jonathan has done, starting from the hypergraph and just three simple assumptions.
Honestly, I find Scott Anderson's position rigid and glib. Maybe I'm missing some fundamental flaw that he has seen in Jonathan Gorard's papers. I don't know. But I _do_ know that the first steps towards finding a way forward in physics are _admitting_ that we're stuck and _allowing_ radical new ideas, even if they're only half-formed.
Thanks again to pointing me to this new interview @tantzer6113
Thanks for the response. I had similar views on the critique which I also though was unfair and applicable to any attempt at a ToE.
But then again, I am probably quite biased. Since I begun following this channel and WF physics I can no longer see the universe being anything other than computational on a fundamental level. @lasttheory
@@dexterdyall3425 Yes, thanks Dexter. Me too: once you start thinking of the universe in these terms, it's difficult to go back to the old paradigm!
At what point will we see Wolfram physics making a difference in our technology, how we interact with the world?
Ah, that's an interesting question!
My inclination is to say: maybe never. General relativity governs our world at a large scale and quantum mechanics at a small scale, so these existing theories pretty much have it covered as far as our interactions with the world are concerned.
But... I'm conscious that I should never say never! If you asked the physicists of the early twentieth century how general relativity and quantum mechanics would make a difference in our technology, they might have given the same answer: maybe never. But here we are with GPS and quantum computers. It's astonishing how much these twentieth century theories have contributed to how we interact with the world.
All of which is to say: I don't know, but you _never_ know!
Thanks for the question!
This is exactly the problem of this model. There is no way to link to physical representation nor can derive something like an electron with its properties. It looks nice and i see as useful to show networks under particular rules but it stops there
Hi Jay, thanks for the comment. You're absolutely right that there's no connection between the Wolfram model and particles right now, as Jonathan admits. But there _is_ a connection between the model and other aspects of physics. Take a look at my episodes with Jonathan on how to derive General Relativity ruclips.net/video/1tjhE0U-mgc/видео.html and Quantum Mechanics ruclips.net/video/YZhCYLZanEE/видео.html from the Wolfram model. The former is, for me, especially convincing that there's a path from the model to physics as we know it.
@@lasttheory thanks for the input.
Will take a look. In terms of gr, indeed it's not surprising because you can derive continuity from discrete systems so you can ideally recognise GR. In terms of quantum mechanics I will look up your video as to what aspects. I did some models based on fibre bundles and built lattice models, but these has close connection to physics by geometric quantisation. Furthermore, with fibre bundles you can realise spin systems, topological order, particle physics and many more. I can't see how Stephen can achieve any connection to close this gap. The hypergraph is not self consistent with these physical representations. I watched his last workshop, and concluded that it could be a promising substrate for modular systems if he can integrate higher level instructions to integrate all nodes to a self interacting system similar to AI but not based on specific language model but modular self recognition algorithms
@@JAYMOAP Interesting, thanks Jay. It's great that these different approaches are all getting people's attention. It'll be fascinating to see whether and how they'll converge.
@@JAYMOAP "I watched his last workshop, and concluded that it could be a promising substrate for modular systems if he can integrate higher level instructions to integrate all nodes to a self interacting system similar to AI but not based on specific language model but modular self recognition algorithms"
This comes across as word-salad to me. Most systems (that we care about) that exist, are modular (more appropriate to say, complex), so why would they only be useful for this particular niche computer science use you are talking about? The whole point of the Wolfram Model is to use that framework, to build up to the higher level instructions that physics has already established (physical equations), which by and large they have done, and to use it to construct complex systems, which most physics models can't do, where this one, excels. I mean... there is a reason it's called a "neural network" ...because LLM's use network models, like the Wolfram Physics Model. Why would you have any doubts that a fundamentally a network model like the wolfram model wouldn't be able to do something like make a neural network? lol question mark?
Your conclusions seem short sighted to me. There's plenty of connections, and personally, I've already made a number of systems based on the Wolfram Model (because like mentioned you can use it to construct any complex system, and by proxy "modular systems.") You just need to spend more time learning and understanding how it works, and let go a little bit of the mainstream physics your attached to.
you really don't have to cut to every single 'yeah' and 'yup' that you say, it's so incredibly distracting
Yup. Oops, sorry, cut that "yup". I hear you, but I need to do a bunch of cuts to tighten up the conversation, e.g. to cut out hesitations and false starts. Cutting to myself is often the only way to do that seamlessly. I do try to keep the cuts to me to a minimum!
Weird. I didn't even notice it until you mentioned it.