Hacker News Reader

Comment by shoggouth | Hacker News Reader

This is a good introduction to it by Stephen Wolfram himself[0].

[0]: https://writings.stephenwolfram.com/2020/04/finally-we-may-h...

Does anyone understand this theory? When I read through it and listened to his talk, it sounded like a bunch of different ideas stitched together without an underlying explanation.

It's like string theory - a framework with tons of free parameters which will never produce a coherent physical theory, but you can always write a paper that contains a promise of explaining any observation with the right fine-tuning of some parameters (the promise will never be fulfilled, of course).

It is utterly unlike string theory.

String theory is a very rigid coherent theory. It has produced plenty of deep mathematical insights. I personally don't believe that it describes our universe, but it is possible to calculate its properties.

Wolframs "Theory" is a bunch of relatively conventional ideas (by high energy physics theory standards), tied together by wishful thinking and speculation. In parts it seems almost possible to show that the ideas are actually contradictory. It is only saved by being to vague to fail coherence checks.

It is, to use the old cliche, not even wrong.

"Not even wrong" is a title of a book about String Theory by Peter Woit.

I've noticed a common tactic in online disinformation campaigns: taking a common term associated with critique of some concept and spamming it in a different (sometimes opposite) context, to break the semantic link.

Not even wrong is a quip attributed to Wolfgang Pauli. This is where Woit got his blog and book title from:

https://en.wikipedia.org/wiki/Not_even_wrong

Oof. Lol. If I ever start rambling about They's malfeasance in Their campaigns, may I receive as humble and direct a reply as this, as soon as possible.

There are exactly zero free parameters in string theory [0]. The details of why string phenomenology is hard is a difficult subject, but the characterization you've given of it is not correct. If you have a proof that string theory is not self-consistent, you should publish it, because there is no such proof in the scientific literature today. (Source: my PhD in physics.)

Unfortunately, there is a ton of misinformation about this topic on the web. For example, people love to say that string theory predicts anything and everything. But it predicts (and rejects) a lot; it’s just that all of the known predictions happen to fall into the categories of (1) predicting things that are very hard for humans to measure (behavior of black holes at long time scales, graviton scattering, etc) or (2) retrodicting things we already know are true (e.g. gravity, Lorentz invariance, etc.). This state of things isn’t by design of nefarious string theorists designing their theory to be untestable, it’s just cruel fate of what comes out of the math. Hopefully someday we can find some other type of prediction, but string theory isn’t easy.

[0] See e.g. https://indico.cern.ch/event/630393/contributions/2890113/at...

How is string theory useful?

To whom? To other branches of physics? Look up AdS/CFT. To the general public? Dunno, I guess the pursuit of understanding the universe is its own reward for some.

How is it understanding the universe if it’s not applicable/useful in the general case?

You mean, if string theory does not turn out to describe the universe, how could it be useful? Well, by giving extraordinarily powerful tools for understanding things that are well established to be useful, like quantum field theory. AdS/CFT gives us the only tool we have to analytically understand quantum field theories in the so-called “strong coupling regime”. This is useful for discovering new properties of quantum matter in systems where you would otherwise need simulations. You can think of it intuitively as string theory providing a glue between two descriptions of the quantum matter, like a “type cast” in programming where you start with one kind of object but reinterpret it as another. The thing that is incomprehensible in one representation is simple in the other. This was discovered by studying limits of string theory in interesting geometries.

I thought lots of variants of string theory do predict things inside human means. But they've all failed, leaving only variants that predict things outside of it.

If you have something specific in mind, I’m happy to address it! But I’m not quite sure what you’re referencing.

Isn't SUSY one of them? That's the first thing I can remember

You would probably learn more by listening to Cumrun Vafa [0] than anything I could say. It’s hard to say much about string theory without space time supersymmetry not because it doesn’t exist (we know it does) but because it’s so hard to calculate anything…physicists are very reliant on a few tools, supersymmetry is a big one, and without it it’s really hard to say anything concrete, yet.

[0] https://www.youtube.com/watch?v=yppqz12ngbM&t=654s

Correct me if I'm wrong, but it seems that it's a mathematical structure so open-ended that you can write most relevant mathematics in terms of it (as you can with e.g. set theory). And of course, if you can write math you can write physics.

Yes. The problem is that you can write any physics in it, so you cannot use it to figure out anything about our universe.

This is a crucial part that many intelligent people somehow utterly fail to understand. If you can explain everything (including the things that are not true), you can explain nothing.

A theory that explains, is a theory that says "...therefore, X can happen, but Y can't happen". Like a mathematician who says that 2+2 is 4, but also says that 2+2 is not 5. Or a physicist who says "apples fall down from the tree, they don't fall up".

Compare that to a "genius" mathematician or physicist who simply says "yes" to everything. Is 2+2=4? Yes. Is 2+2=5? Yes. Do apples fall down? Yes. Do apples fall up? Yes. And then people on internet are deeply impressed that he can answer everything. Such an amazing skill! Ask him about gravitons, he has a clear answer. Ask him about dark matter, he has a clear answer. Ask him about time travel, he also has a clear answer. The only problem is that he can both write a physics that contains gravitons, and a physics that does not contain gravitons. Etc.

Ultimately, we want to know what is real about our universe. (Or multiverse, or whatever.) A model that says "yes" to both the things that are true and the things that are false, is useless. After you figure out what is true, you get "yeah, my model explains that". Problem is, the model explains the opposite just as well.

...then you take a step back, and remember that "can write anything" is simply a different way to say "Turing-complete". Yes, if you invent something that is Turing-complete, you can simulate a universe in it. Any universe. Both the ones that exist, and the ones that don't.

(And the idea of Turing-completeness was discovered a few decades before Wolfram was born. So we can't even credit him with inventing the concept. He just uses the concept to impress people who either never heard about it, or never connected the dots.)

The mathematical structure is called a hyper graph, a graph where nodes and edges come from the same set. Sounds weird at first but think about it like little sentences of the form subject predicate object. The word google can be all three places, as in: Google is a company, I googled his name or I work at Google. Here Google and its conjugations is an socio-economic object, but the set can also be the different grades within a geometric algebra where the different kind of reflections can be both operator and operand.

What makes it "good"?

Well, I can say I made more headway with this than with his book on the same topic that I saw at the library. In particular, this article seemed to justify the steps toward the "ruliad" a little better. Or maybe it's juuuust short enough that I didn't lose patience first.

Reply on news.ycombinator.com