About a year or so ago, Elon Musk made an argument that we are most likely living in a simulation. I felt strong disagreement with that argument back then, but didn't know of steem or have any motivation in writing it down, so I kept this to myself. However, recently, this post reminded me about Musk's claim and I decided, it's time to share, why I think he's wrong.

Two kinds of theories

Now, when people talk about simulation theory, they talk as if it's a single idea and everyone knows what they are talking about. They use phrases like: "we are part of simulation" and then they reference movie Matrix, not realizing that those are not the same ideas. In Matrix humans were hooked up to computers, which made them experience virtual reality. So simulation did not create minds, it just created inputs for those minds.

However, it seems that people often mean something different when talking about this theory. They think that not just the world, but we, ourselves are created by the simulation. Which is quite logical assumption, if you hold the prevailing view that consciousness is created in the brain. If that view is true, when of course, it makes sense that the mind is also created by the simulation, because it creates the world and brain is part of that world.

People probably don't make a distinction between these two ideas, because they don't think that's important. For me however, there is a huge difference. The second one requires making a much bigger assumption: that it is possible to simulate (or create) consciousness in a computer.

Either way, as you might have probably noticed, I see a problem differently than Elon Musk and that's why I think his argument is off target. In this post, I will try to present reasons for why it is off target, in the context of both types of theories. But before that, let's remember what his argument actually says.

Computer game argument

His main argument is the visual realism of computer games. Computer graphics have improved so much through out the years, he says, that it's just the matter of time till it becomes indistinguishable from our reality. And when it reaches that point, we should be able to simulate the kind of reality we are in. So if we are getting near to that point (and according to him we are), that means it's quite likely that other civilizations, have come to that point as well, and that substantially increases the likelihood of us being in one of their simulations.

Well, he might be right on one point, for us computer graphics might become indistinguishable from our reality. We see a bunch of pixels on a screen and our brain automatically interprets it as something it sees from the real world. But how does it help with "we are living in a simulation argument"? It doesn't, I think. At all.

In fact, how real simulation looks from the outside is irrelevant, to this problem. We should be instead asking, how real does it look from the inside. The theory is that, we are living inside a simulation, after all.

Mind in the computer theory

If we believe that we (our minds) are simulated together with the rest of the world (second type of theory defined above), that means we are part of the simulation software. Not just that, it also means that we are part of software, which believes in the world it creates.

Now that's just ridiculous, if you ask me. How do you create software which would believe in the world it simulates Furthermore, on what level would it believe in itself? Programmers know that, software is all about abstractions. On the lowest level, it's all ones and zeros. You interpret those bits a certain way, make computations on those bits, and then interpret the output of the computation. Programs are just set of instructions, on what operations to perform on those bits. Do you see any space for conciousness in a computer defined this way? I certainly don't.

How do you create a conscious computer program? That's how this problem should be defined. It makes you understand, how huge this problem is, when you define it this way. It's certainly not something we are anywhere near of technologically. So to dismiss it as something, to be solved simply by scaling technology in a way we already have, while making a case for something which would depend on it, is a mistake in my opinion.

I realize, that many people believe that creating consciousness with computers is just a matter of computational power. I disagree with them. However, that's not the topic of this post. And the point of this section is to show you, that there is a way bigger problem, to creating a simulation of the world and minds in it, than just making it visually indistinguishable from our reality.

Matrix theory

Now if we are talking about Matrix kind of theory (mind hooked up to the simulation), the problem becomes a bit less complex. But still computer game argument does not help much. It's still not about how real does it look to the spectator from the outside, but rather, how real does it look to the mind which is hooked up to the simulation? Before that, of course, we should ask, whether we can input any kind of senses directly into the mind at all? As far as I know, currently, the answer is no. If you have a different answer I'd like to hear it (you are welcome to leave a comment).

On the other hand, you could argue that VR technologies do, just that. And you could make them add inputs for other kinds of senses, besides vision. But have you heard anyone get so immersed in it, they forgot they are playing it and that it's not real? Because it definitely seems, that if we are in that kind of simulation, we forgot we are in it.

Either way, the biggest obstacle to creating the kind of simulation we are in, doesn't seem to be visual realism. It's more about how to make the mind totally forget it's base reality and replace it with the simulated one and I think it would require some big breakthrough inventions, if we are ever to see this done.

Conclusion

As you can see, there are some big issues to be solved before we can create simulations on the same level as the reality we are in, and making games more realistic doesn't do much (if anything) to help solve them. It's very much debatable if more computational power in general would even solve them. So it doesn't make sense to base claim, that we are living in a simulation, on the idea, that it's just the matter of time before we are able to do that ourselves, because it's certainly not just the matter of time.

Now, does it mean that simulation theory should be outright rejected? No, of course not. It just requires much stronger arguments, than the one Elon Musk gave in that talk (maybe he has some better ones, who knows?). Personally, I don't think it's likely, at least not with the kind of computers we can currently conceive of. Quantum physicists seem to agree with me. So I've got that going for me, which is nice, considering I'm trying to argue with Elon Musk.