In the 17th century people thought the universe is a kind of clockwork because they were so impressed with the precision of those early machines.
Today people think the universe is a kind of simulation because they’re so impressed with how algorithms tinged with randomness can tickle their human sensibilities and create illusions of meaning.
That's an unfair take. Yes, people have always been using the most interesting or complex phenomena they knew as metaphors to explain the world, starting with animals and ending with computers. But it's not about being impressed, but rather finding similarities. And most importantly, bodily fluids != clockworks != systems of pipes != computers.
There's a qualitative jump we've made in the last ~150 years. The tight feedback loop between math, natural sciences and engineering, that was earnestly established some centuries earlier, finally picked up speed. We're no longer "impressed" by a deer or a clockwork and saying the world must work like it. We're not imagining the universe to be like something we know - we're mapping alike concepts using precise, formal, well-tested reasoning. We're applying models to the world, and we know exactly how much fidelity they have. We chose those models to be useful, not evocative.
In short, back then we were doing artistic impressions of a landscape. Nowadays, we're drawing proper maps[0].
There's a popular meme that ~150 years ago, physicists thought they had it all neatly figured out, and all that's left to do is to make numbers accurate in far decimal places - and then they stumbled on relativity, nuclear physics and quantum physics, turning everything upside down. The meme is inaccurate, and its implication - that we still don't know shit - isn't particularly convincing to me. Those new fields didn't replace our understanding of the world - they enriched it, solidified it, filled in holes. We have a more complete picture now, especially of the fundamentals. We may not have solved quantum gravity, we may not know if and what dark matter is, etc. - but we know enough to put bounds on the possible consequences and possible surprises[1].
The point I'm making is that, when we now say that e.g. the brain is a computer, it's not the same thing as people 500 years ago saying the brain is like a clockwork. We're not vaguely hinting at similarities - we're applying a specific, precise model. A model that makes concrete testable predictions. A model that can be studied to yield more understanding. A model that's tied to what we now recognize as fundamental - computation isn't some gears and belts trick, it's one of the most basic and impactful ideas in mathematics.
It's similar the case of "simulation argument". Do we live in a simulation? Who knows? We're not sure if we could tell (unless it's a really hacky one) or if it would matter much. Is it possible for our universe to be a computer simulation? Probably. We know enough about physics, biology and information theory to have a justified belief it can be done, especially if it's designed around players. Decades of videogame development experience tells us how to do it; mathematics and natural sciences give it green light and say it's an engineering problem.
And as a final point to this little rant: when I say that "we know" something today, it's not the same kind of knowing as we had 500 years ago, or even 200 years ago. Mathematics and natural sciences are thoroughly interwoven. The parts that we are sure of are all mutually reinforcing - if we're wrong about any one of it, it would mean we're wrong about most of the rest, across many fields and disciplines. 200 years ago, that might have been possible. Today? We've built so many technologies and processes based on our scientific models that everything we do today, every second of every single person's life on this planet, is its own experiment confirming that our models are good fit.
--
[0] - Comparing maps from 500 years ago to maps today is a good exercise. The difference isn't just in accuracy - it's qualitative, as we now have a deep theoretical understanding of what maps are and how to make them, capacity to make them arbitrarily good for desired purposes, and experience in putting them to actual use.
[1] - This is, sadly, what makes me very pessimistic about faster-than-light travel and certain other sci-fi dreams. We may not know enough to rule those out directly just yet, but what we do know surrounds the problem space tightly and lets us rule them out via indirect proof.
The problem with the simulation idea is one of compute. Accurately simulating the quantum interactions between 30 particles is beyond our best supercomputers [1]. Imagine what it would take to compute a cell in all its quantum glory. Even with quantum computers, I certainly can’t conceive of a computer with enough qbits to approximate how many qbits there are. I just don’t think there is anything in our experience that would lead us to believe the universe we live in is able to be simulated. Maybe it’s conceivable if you took a bajillion short cuts with the physics, but then why the hell is quantum mechanics even a thing?
- That everything that makes you you and that makes you tick is located in your body, with the higher-order things like personality and memories being located in the brain;
- In particular, it's not located somewhere else outside the body, especially not in some metaphysical realm;
- That there is structure to the brain - patterns in arrangement of physical components and their transient states (including electrical and chemical signals) - that contains the entirety of you; that structure can be studied, and over time understood to a degree;
- That once we know enough about this structure, we can reach into the brain from outside to poke it with various implements, chemicals, beams and fields, and achieve changes we predicted will happen;
- That we can use insights from information theory, and even other areas of computer science and computer engineering, to better understand and predict the structure of the brain;
- That we can replicate various substructures of the brain in a simulation, and that such replicas will be useful to study the real brains, and also could be incorporated as components into computer systems we build ourselves;
- That artificial brains are possible to build;
- That artificial brains are possible to simulate on computers we build;
Some of those have already panned out. Some are used with great success in medicine. Others may take a while to verify.
Now, I recognize that some of the points I listed invite questions along the line of, "is that a prediction specific to 'brain is a computer', or would it also be a prediction of 'brain is an X' for many other X-es?". I believe the answer to that is "both", because anything you could substitute for 'X' that would yield similar predictions can itself be viewed as a computer! Computing systems aren't just next iteration of clockworks - computation is a new framework for interpreting the reality around us and making novel, precise predictions about it.
6000 years ago they had dudes (whole tribes, actually!) who practiced ritual trepanation for reasons unknown to us.
This "mapping the human brain" is no different and no more scientific.
The human brain is an information processing machine, and the science of information and its structure was discovered less than 100 years ago.
We don't even have a name yet for this science!
And our baby steps in this scientific field are nowhere close to beginning to study the brain. To use an analogy, we discovered how to make paper airplanes, but that gets us no closer to building a supersonic jetliner.
> 6000 years ago they had dudes (whole tribes, actually!) who practiced ritual trepanation for reasons unknown to us.
I'd thought it was established it was done for medical reasons - people realized that it sometimes helps with conditions nothing else helps with. Arguably most of the things humanity did until the last few hundred years was based on empirical correlations and stores to help remember them (but lacking predictive power). It's only recently that we've developed proper theoretical understanding of most things we do, in form of specific, tested theorems with lots of predictive power.
> This "mapping the human brain" is no different and no more scientific.
It is different because nowadays we do have proper science - and more importantly, we know how proper science looks like. So even if we still know very little, we at least know what can and what cannot be done with that knowledge.
> (...) the science of information and its structure was discovered less than 100 years ago.
> We don't even have a name yet for this science!
Isn't that just "information theory"?
> To use an analogy, we discovered how to make paper airplanes, but that gets us no closer to building a supersonic jetliner.
IMO that would be a good analogy for the clockwork age. Today, we not only know how to make paper airplanes, but more importantly, we can imagine a supersonic jetliner being a thing, we have justified confidence that there's a path from here to being able to build one, and that studying the phenomena behind the flight of a paper airplane are important steps towards building a supersonic jetliner.
We do know the components of the brain. "Hypocampus", "Prefrontal cortex", etc. The AIBS is studying how these regions map down to the individual elements (neurons, peptides, chemical transports etc).
You're discounting a significant amount of experimental science that is really going on right now. I don't understand why.
That's like saying we know the components of an airplane - there is the "hard shiny bit", "the bottom rubbery bits" and the "rotator thingys".
Decomposing the brain into groups of biological clumps of cells tells us literally nothing about how the brain as an information system works.
When we want to study the brain, we do so because we want to understand its information structure and build information processing models. The biochemistry of the wetware is irrelevant, except insofar as it might help us formulate an "information science".
Here we are no closer to the goal than 6000 years ago.
But we do. Check out the research into how our visual system works, or how spatial orientation works in mice. There's been quite a lot of puzzle pieces identified over the last decades. For some of those, we have good theoretical models that can make testable predictions.
We're far from having all the puzzle pieces, and even further from fitting them together. But what we already have and the progress that's happening are both reassuring.
I don't agree.
We have some very interesting ad-hoc empirical observations, but no theory.
We're not even in "alchemy" territory here yet, we're still in the "randomly mix colored rocks and see what comes about" phase, to use an analogy from how chemistry developed.
But we don't know if we're closer more like how Worcester is closer than Boston is to Los Angeles or how Worcester is closer than Boston is to Springfield.
I think simulation is a hamfisted word for the modern take on something that people have been considering for a long time. The idea that the universe could be "virtual" and conscious experience is more "real" than physical phenomenon have been actively discussed in philosophical literature since the 16th century, and it was certainly a topic of conversation long before that.
I think there's a fundamental difference and 2 different subject here.
On a universe level, it's impossible to tell if the entire thing is just a simulation.
However, I don't think our conscious experience is any different from a physics event happening elsewhere in our universe. If the universe is indeed simulated, we just happen to evolve naturally in that simulation, and there is no special need to simulate our consciousness specifically - that's a little narcissistic to think about.
Totally agree, the people who think we're living in the matrix, all atomic minds plugged into a made up reality haven't thought about the problem enough.
That's exactly why we're tired of simulationism. Short of the simulation runners talking to us, it's impossible in principle to take the least step toward figuring out if it's true or false.
I mean not exactly... It depends on the type of simulation we could be in. If we're in the type that is monitored and if we could find a way to mess with the simulation that we'd just be rolled back to a previous state, then yes, it's impossible. If instead it's watched about as well as some of us monitor our own VMs, well, maybe a different story. Of course we're still a long way from many things in physics like a theory of everything, so no need to worry about being in a simulation at this time.
> People just can't let go of wanting some kind of God
The reverse is also true. While the inability to determine why the laws of physics are there is not evident that there must be some reason that they are there, one can conjecture a causal relation with some kind of entity that affected it. Either that be God, or the playful intern that happened to create UniverseSandbox#9971
Have you considered that fun is not always without side effects? The same machine that came up with physics, calculus and evolution came up with God and the MCU. God and the MCU are fun ideas too but just look at how much damage they've done to the world.
