A question about computer simulation and the brain.

Recently a thought hit me that I thin was profoundly interesting if possible.

Today, we can accurately simulate, last time I read, something like 50,000 molecules. And that's semi-exponentially growing.

So here's a thought that I had in regards to that. Let's say we one day hit a point that we can simulate every molecule of a human brain. Now lets say we do just that. We simulate a human brain. But it's on hardware that can only simulate it. It has no ability to create AI, because we simply haven't mastered that ability yet. We cannot create true, self learning AI. So now, what if we make that simulator on that computer accurately simulate the human brain. In theory, is this computer not simulating something with more powerful simulating abilities than itself?

That's the main thing I realized. The human brain is an amazing simulator. It's processing power is astonishing. We can live years-worth of dream-lives in a mere 4 hours of sleep. So if you told this computer to input a simulation into this simulated brain, would it not get results more accurate and powerful than its own hardware is capable of?

Furthermore, would this simulated brain be alive? and could we alter it to be, say, savant? And in turn simulate whole other things? I mean i theory it could simulate all of pi, if there is one. You could simulate an insane person's mind, and in turn, generate whole worlds from psychopathic viewpoints.


I just wanted to get your ideas on it. This very strange idea of simulating something with more simulating abilities than the simulator is incredibly perplexing to me. And perhaps a touch of Matrix-ness in us. After all, what happens if the simulation wants out?

Id say the resources in the human brain are overplayed, i mean i cant remember much before last week, so we have various forms of memory, short term, immediate, long term. We also have pretty limited decision making ability after childhood. People seldom change their mind about subjects as they get older. You'll find an atheist or theist will need something pretty profound to happen in their lives in order to remove a well formed thought process. Its almost as if the original process needs to be destroyed so it fights to stay alive.

Simulation of the human mind is probably already going on and I would expect has been for some time.

i was reading an article in 'focus' magazine just yeaterday on this topic. it said that quantum computers should be able to process multiple info at the same time and we could see the rise of AI in the next 20-30 years, not sure on all the deatials as it was a small article. anyway its the one thats out in the next few days (i read it early as i work in a dsitribution warehouse) and its the cover article if you want a read

reply to post by Gorman91


I do not know which molecular dynamics simulation engine you reference, but if it's NAMD or GROMACS, they pretty much simulate electrostatic forces between atoms (and hence, molecules). They ignore or simply estimate the effect of the other 4 forces that determine atom interactions - they have to, as the simulations take way too much wall-clock time as it is now. Last I heard, a 2.5 million atom ribosome takes 6 months of computer time in simulation.

So the baseline simulations are a little imprecise, and they only simulate interactions between atoms/molecules, which determines relative atom/molecule placement from time step to time step in the simulation.

So let's say we could simulate all 5 forces for billions of atoms (in a part of a human brain) in a reasonable amount of time (months, not years). That's only a simulation of the structure of neurons and synapses and chemical interplay. What causes the first spark across a synapse? What I am pointing to is how do we simulate the interplay of synapses firing caused by conscious thought? There's the question that fascinates me - say we could model the brain's neuron's, how do we model thought on those simulated neurons?

I have no idea, and I find this a marvelous inquiry.

reply to post by Gravity215


I don't see how this is true. Just because most of the population is that way does not mean that all are. A lot of people with mental instability, for example, have absolutely no consistency in choice.

reply to post by macFiachna


indeed. But I think that we would just have to simulate atoms behaving under the same conditions of the brain, and then we would not need to worry about synopsis. If you got it right on the lower scale, than the higher scale will behave as it should. I suppose, if consciousness truly is as powerful as people think, then this brain without a body would simply make up its own world to live in, given enough time. I suppose we could program what it sees or hears or smells, etc etc.

reply to post by Gorman91


I understand what you are saying, and even so I am having difficulty with thinking this scenario through. It is as if my hardware is at the moment capping whatever UI (unartificial intelligence) my brain is capable of!

Perhaps that is the answer. In trying to give a mechanized creation intelligence, one would not be able to move beyond whatever AI "nuts-and-bolts" are capable of generating. Perhaps the kind of technology that would break through the bonds of a nut-and-bolt world and propel us into an era of accelerated growth, is exactly the kind that simulates real thought - the kind that is modelled on a "real, living and breathing" sentient being. Perhaps all nuts and bolts are limited in that fashion, and it takes the introduction of a modelled or simulated human brain to break beyond "the programming".

Which is not to say that our "hardware" isn't in fact in essence just some kind of nuts-and-bolts creation? Who is to say that our brains is not a foreign and introduced "simulation" into a constructed body, and whose purpose is to propel us past our hardware's basic capabilities? I will always remember the episode of Star Trek Voyager in which the debate was whether holograms like the Doctor deserved equal rights in society. And the decision was that, as holograms are just atoms held together by force-fields (which is actually how our physical bodies are also constructed when viewed on a microscopic level), holograms deserve as much recognition as a real-life individual.

What is my point? The point is that the manner in which our atoms are structured and held together may just be another more advanced lifeforms way of building a technology like a computer or robot, and that our brains may be that simulation that allows us to break past our programming.

S+F for a good question, and for giving my mind a good yoga-workout.

reply to post by Gorman91

We can live years-worth of dream-lives in a mere 4 hours of sleep.

I dont think so. As far as I know, we dream in real time.

reply to post by Maslo


I do too, but on the rare occasion I don't. While it's rare if ever now a days, when I was a kid, I would wake up with whole civilizations in my mind. I struggled with that for a while. These people, each with their own minds, their own ideas, their own pursuits, all fabricated projections of a creative mind. I suppose that may be the origins of such a thought as this, now that I think about it. My brain was simulating multiple brains, which in turn such brains were each simulating their own views of the world they lived in, a world of my own mind's creation. It's sort of like a multiple personality disorder lol. Only I know they are all fabricated people in the depths of a creative consciousness.

Originally posted by Maslo
reply to post by Gorman91

 

We can live years-worth of dream-lives in a mere 4 hours of sleep.

I dont think so. As far as I know, we dream in real time.

I donit think thats accurate, ive had many of times woken up from full dreams , with intricate detail, story and time lines, only to find out that i had been asleep for less then 10 min.......and looking back and remembering the dream and thinking, theres no way that was a 10 minute dream............I would have figured i was out the whole night

Originally posted by Gorman91
So here's a thought that I had in regards to that. Let's say we one day hit a point that we can simulate every molecule of a human brain. Now lets say we do just that. We simulate a human brain. But it's on hardware that can only simulate it. It has no ability to create AI, because we simply haven't mastered that ability yet. We cannot create true, self learning AI. So now, what if we make that simulator on that computer accurately simulate the human brain. In theory, is this computer not simulating something with more powerful simulating abilities than itself?

For simulations of that precision, you would need a computer much more powerful than the brain. For reasons of processing constraints, we simulate the brain, but when we know more about the brain and it's firmware and software, it would be faster to emulate it on dedicated hardware.

That's the main thing I realized. The human brain is an amazing simulator. It's processing power is astonishing. We can live years-worth of dream-lives in a mere 4 hours of sleep. So if you told this computer to input a simulation into this simulated brain, would it not get results more accurate and powerful than its own hardware is capable of?

No. Binary computers are limited to a certain precision in everything they do. The brain is not. The brain is a massively parallel analogue computer. The results on your computer would be constrained by the software simulation of binary to analogue or analogue to binary.

Furthermore, would this simulated brain be alive? and could we alter it to be, say, savant? And in turn simulate whole other things? I mean i theory it could simulate all of pi, if there is one. You could simulate an insane person's mind, and in turn, generate whole worlds from psychopathic viewpoints.

I believe, from personal experimentation that you can create a sentient AI, but I don't think that we are there yet.

I just wanted to get your ideas on it. This very strange idea of simulating something with more simulating abilities than the simulator is incredibly perplexing to me. And perhaps a touch of Matrix-ness in us. After all, what happens if the simulation wants out?

Simulation is pointless if it takes too long to run a simulation.