[ExI] Why do the language model and the vision model align?

[Ben Zaiboc]

Apologies for not changing the subject line (again!), as this is now so far off the original topic. Probably not worth it now, though, as I think this has run its course.

On 19/02/2026 18:12, Jason Resch wrote:
> On Thu, Feb 19, 2026, 7:05 AM Ben Zaiboc via extropy-chat <extropy-chat at lists.extropy.org> wrote:
>
>     On 17/02/2026 12:05, Jason Resch wrote:
>
>     >
>
>     >
>
>     > On Mon, Feb 16, 2026, 1:33 PM Ben Zaiboc via extropy-chat <extropy-chat at lists.extropy.org> wrote:
>
>     >
>
>     >     On 16/02/2026 16:34, Jason Resch wrote:

> ...

>
>      Maths is an expression of the properties of the universe, a consequence of the particular laws this universe uses, or at least a consequence of the way we see them. 
>
>
> That is one way to look at it, among many.
> For example see this section:
>
> https://alwaysasking.com/why-does-anything-exist/#Math_Matter_Mind
>
> Which shows even among three physicists, they each hold different positions.
>
> For example, you say that "Math is an expression of properties of the universe." But I think it is just as possible that "The universe is an expression of properties of mathematics."
>
>
>     We don't have to abandon maths just because it might not be 'true', or might not actually exist in some hypothetical mystical plane of existence.
>
>
> Those who subscribe to mathematical realism hold mathematical objects to exist as concretely as any existing physical universe does. There's nothing mystical about it.

If someone tells me that the square root of -1 exists as concretely as the monitor sitting in front of me, I'm going to call that mysticism. I don't know what else to call it.

>
>      We use maths because it's useful, not because it's true, or actually exists.
>
>
> But note that to be useful, a mathematical theory must accurately differentiate true from false. So when one of our useful mathematical theories says it is true that "$1000 - $995 = $5" also tells us that 9 is non-prime because an integer factor of 9 (besides 1 and 9) exists, are we not right to say "3 exists"? 
You've lost me there.

> What about when the theory says there are primes so large we will never be able to compute them? This is an inevitable conclusion if we take our mathematical theories seriously.

Well that's simple enough. The theory says that these prime numbers will never exist. If there can be such a theory.
>
> It is no different from the physicists who takes general relativity serious and who concludes, based on the measured curvature of the universe, that there exist regions space far beyond the cosmological horizon. They are so far away that we will never be able to see them. But these regions must exist if our theory of GR is true.
So these regions of space /have/ been calculated. Which is a different thing.
>
> In both cases, we are taking established useful theories at their word, and using them to predict the existence of things we may never see.
I think you are confusing things which can be shown to exist and things which can't.
We can show that there are regions of space beyond what we can see, but we can't show that there are (or are not) infinitely many primes.
>
>
>     ...

>
>     Do those certain things exist independently of being worked out? That's kind of a non-question. Does a falling tree make a noise if nobody hears it?
>
>
> Did this physical universe not exist before life arose in it?
Yes, it did. We can figure that out.

>
> Would it not then still exist even if no life ever evolved in it?
That's not a question that can be answered.

>
> ...
>
>
>
>     > You believe there are more than 52 Mersenne primes, don't you?
>
>
>
>     I have no beliefs concerning Mersenne primes, mainly because I don't understand what they are. I did look up the definition, but that didn't help.
>
>
> A Mersenne prime is any prime number that's one less than a power of 2. In other words, a prime that when expressed in base 2, consists of all 1s. For example:
>
> 3: 11
> 7: 111
> 31: 11111
>
> As of today, only 52 Mersenne primes are known. But it is believed more (and possibly infinitely many) exist. Let's assume there are more.
>
> Then consider the following statement:
>
> "A 53rd Mersenne prime exists."
>
> Is such a statement true?
You just said 'it is believed...', so the answer to your question is "some people believe so".
>
> Or does it only become true after someone finds it?
After someone finds it, it is certainly true. As to  whether it's true before then, well, "some people believe so" is the most you can say.
> ...
>
>
>      I don't think that someone in ancient mesopotamia who said "there are no buildings half-a-mile tall" could reasonably be said to have been wrong, despite the fact that, given the right circumstances, it's possible to create buildings half-a-mile tall.
>
>
>
> I am not talking about possibilities which may or may not exist, but rather, conclusions we must accept if the theories we use and rely on happen to reflect the underlying reality.
Which applies to tall buildings as much as to mathematics. We know for a fact that half-mile-tall buildings can exist. Nevertheless, we don't conclude that the mesopotamian was wrong.
>
>
>     > There are different forms of existence.
>
>     >
>
>     > There is existence defined by being mutually causally interactive (what we normally think of as physical existence, or existing within this universe). 
>
>     >
>
>     > But then there is also existence for things which are acausal. For example, two bubble universes in eternal inflation that will never interact, or two decohered branches in many worlds, or even just other universes with different laws, which we presume must exist to explain the fine tuning of the laws of our own universe. In what sense do these other universes exist?
In the 'hypothetical' sense, unless they are proven to be factual.
>
>
>     >
>
>     > Are they still worth of the full "concrete physical existence" when we can't see them and can't interact with them? Or should their existence be demoted to inferred/abstract/theoretical?
>
>     >
>
>     > If the latter, isn't that the same sort of existence that mathematical object have? Other physical universes can be studied via simulation, we can analyze their properties, what structures exist as a result of their different laws, etc.
>
>     >
>
>     > The abstract sort of existence that other possible universes have seems to be, to be the same sort mathematical objects have.
>
>
>
>     Ok, we can agree on that.
>
>     It could be called 'imaginary, but with rules'. Which is a subclass of 'imaginary'.
>
>
> I think other physical universes deserve a category slightly higher than imaginary.
Don't think I mean anything derogatory by the word 'Imaginary'. I have a lot of respect for the imaginary. It could even be said to be the key factor that makes us human. And allows us to devise things like maths and science.

We don't know that there are other physical universes, we only actually know about this one. We can theorise about them, though ('imaginary with rules').

>
> Certainly the physicists who postulate the actual existence of other universes (to explain cosmological fine tuning observed in our universe) are doing something a little more serious than contemplating things in the same category as Santa.
Absolutely, hence my simplified classification system.
>
> Likewise, I think mathematicians who devote their lives to thinking about objects in math are doing more than playing imaginary games.
That is exactly what they're doing. We have countless examples showing this, from Einstein through William Hamilton to Kekule (yes, chemistry rather than maths, but it's the same process: playing imaginary games (with the relevant rules)).
>
> Quite often in history, mathematicians had already laid the groundwork for physical theories not yet conceived.
>
>
>
>
>     So I'd propose a simple classification system:
>
>
>
>     A) Stuff that physically exists (ducks, people, sofas, stars, magnetic fields, etc.)
>
>
>
>     B) Things that are imaginary (exist as information patterns in minds: Santa, Jealousy, Immoveable objects, Other minds, etc.)
>
>         B1)  Imaginary things that conform to specific rules (Maths, Cricket scores, Cutlery etiquette, etc.)
>
>
> A gold start, but I don't think there is a clear spot for:
>
> - Regions of space so far away we can't see them or interact with them?
> - Other branches of the wave function?
> - Actually existing alternate universes with different laws?
>
> Are all these things physical?
Ok, to expand it:

A) Physical stuff
    A1) Stuff that can be demonstrated to physically exist (ducks, people, sofas, stars, magnetic fields, etc.)
    A2) Stuff that can be shown by theory to physically exist (space-time beyond our light-cone, black holes, quarks, etc.)

These two categories are closely related, with experiments sometimes proving theories, and sometimes giving birth to new ones

B) Things that are imaginary (exist as information patterns in minds: Santa, Jealousy, Immoveable objects, Other minds, etc.) 
    B1) Imaginary things that conform to specific rules (Maths, Cricket scores, Cutlery etiquette, etc.)

I don't know how to categorise 'other branches of the wave function' because I don't exactly know what it means. Sounds like quantum stuff, though, and might mean the same thing as 'other universes'?

>
> If so, consider that string theory suggests there are at least 10^500 different sets of physical laws. All these different universes existing as different laws as a result of one mathematical foundation of string theory equations.
>
> According to strong theory, all these universes physically exist.
According to religious theory, so does the holy ghost.
>
> But what makes the equations of string theory special? Why shouldn't there be universes that follow other equations besides those of strings? If other equations defining other universes, are no less valid than string theory, then the line between physical existence and mathematical existence dissolves.
>
> Physical existence is nothing more than mathematical existence. And we are back to Platonism. Or as Tegmark describes it, the mathematical universe hypothesis:
>
> https://en.wikipedia.org/wiki/Mathematical_universe_hypothesis

Well, I think it's the other way around.

In the end, as John Clark is fond of reminding us, there's no disputing matters of taste.

-- 
Ben