Hacker News Reader

Comment by 60secz | Hacker News Reader

Yes and the transistors are baryonic matter 1:1 with the actual universe. The map has become the territory. Seriously, the simulation argument has to be one of the absolute stupidest.

Even from a probability standpoint, it's not P(A) vs P(B), it's P(A and B) that we're in a simulation.

The amount of energy required to run a universe scale simulation is beyond our ability to even begin to calculate. It's an interesting and useful thought experiment, but a terrible argument. I suspect people like Musk popularize it because they want to highlight the real and terrifying risk that humanity will destroy itself.

As Occam stated: "Entities are not to be multiplied beyond necessity."

View on HN

The amount of energy argument is poor because it's based on the laws of physics and scale that we ourselves perceive. You have to think outside the box.

What if the simulating universe simply had 5 spatial dimensions? Simulating our 3d+time universe embedded in that universe would be trivial.

> The amount of energy required to run a universe scale simulation is beyond our ability to even begin to calculate.

Wouldn’t the simulation only have to run whatever is being perceived? No need to simulate things that are completely unattended.

The sim isn't 'computing consciousnesses' brains are emergent. However the reason time 'slows down' (Relativity Theory) could could be 'on purpose' because there is finite computing power in 'God's Computer', and the trick of Relativity length contraction and time dilation is a compensation to ensure calculations can be done in time, or even an 'emergent' phenomena due to the fact that calculations do take time, even in God's computer, and if the later interpretation is the case it implies "Gods Computer" does experience the same "spacetime" that we do in this universe. And if that's true it indicates Gravity is external to the Simulation while the Standard Model is part of the simulation.

Except that you need to ultimately compute the backwards light cone of every possible event that is part of a causal chain. I.e. this caused that, so we have to compute why this caused that, and what caused those things. Sure, if an event is never observed, i.e. it never influences any event which a consciousness perceives, either directly or indirectly, then you can leave it out. Problem is, we keep coming up with new ways of observing phenomena. We keep doing particle experiments and building telescopes, so the load on the computation of the universe is getting pretty big!

The simulator would just need to simulate your mind such that it believed that stuff existed. For example, if you looked in a telescope at a distant star the simulation wouldn't need to simulate the distant star and all the photons and wave length shifting and gravitational lensing and whatever, it would just need to draw a little picture for you.

You'll basically end up at solipsism if you follow this reasoning. There are good arguments against solipsism, e.g. that people smarter than you exist, and you forget where things are, only to rediscover them, then remind yourself you forgot, and so on. You have to play increasingly tortured games to keep believing that nothing exists outside your mind.

Another one against solipsism: Language evolves in groups. A true solipsist would be mute therefore. Source: Wittgenstein.

And you only need to simulate a single brain to be simulating it's entire perceived universe

Exactly right. The Sim is computing Physics itself. It's not computing brains, but only the atoms making up brains. Atoms do what they do, ergo brains form.

Okay, but you’re not running the simulation time 1:1, are you? What would be the point?

So very rapidly your simulation develops people, who develop algorithms and computers for which there are no physically possible (in the real universe) optimizations, so you’re computing for them 1:1. And they develop simulations of their own. And then their simulations develop simulations. And you’ve got to pay the energy cost for all that because some things can’t be faked; or you kill the simulation when it learns to develop its own simulation. And then... what would be the point?

There are some problems with “Okay, but then you just make things simpler,” where this would allow simulations to determine they were simulations. I don’t recall the details off the top of my head. (Its late and I’m typing this right before bed.)

So maybe there is infinite energy in the real world. Then why wouldn’t our universe look the same?

There are certainly kinds of simulations that are possible, but the most commonly discussed one is the ancestor simulation. And it just doesn’t physically work while also sticking to the point of running an ancestor simulation.

> What would be the point?

We do pared-down simulations of things all the time. I think the rest of the comment presumes that the only type of simulation we should conceive of is one that is an exact mapping of some external world, and that external world adheres the same laws that our "simulation" has.

Don't get me wrong - i'm not for or against the simulation argument, I just don't understand the argument that a simulation of "a" universe would consume all the energy of some "other unknown" universe. Nor do I think that the simulation would necessarily have to map 1:1 for us to be in a simulation.

You should read the papers associated with the simulation argument. There are, of course, loads of simulations you could run that don’t fit the confines of the argument. But if you don’t have a well-defined definition of simulation you can’t reasonably talk about the probably that you’re in one.

Yes, you can simplify a great deal. You don’t need to simulate the moons of Jupiter before Galileo comes along. You can’t simplify an algorithm that is at its theoretically most efficient. That’s when you start running into the 1:1 problem.

My argument is that the only way to avoid the 1:1 problem is to craft the simulation in such a way that it’s not an ancestor simulation as defined in the papers.

Since causality exists, it would have to run everything simultaneously even if it isn't being observed.

Cause the universe from which we are being simulated, adheres to the same exact laws of physics? Who's to say a 9-volt battery in that universe can't power Manhattan for a year? I mean if this were like an incremental game, it'd just be a matter of raising or lower multiplication thresholds for complexity.

The dinosaurs thought the same thing about computers.... or not...

Reply on news.ycombinator.com