Hard No
Every object would require modeling every atom which would require at minimum , that is at the theoretical limit of data compression, an equal number of atoms. So you need another universe to make a perfect model of the universe.
But then how can my brain envision all objects in the universe? Suppose I had an infinite life. I clise my eyes for long enough and let all possible melodies, visions of atoms, visions of storms while they move, or take your pick. After a finite time it will exceed the number of objects. All with the senses shut. How can so much be contained in the small brain? The information is latent there. Not explicit. The info is constant changing. All contained in the brain. Likewise if you construct a device with say 100 trillion neurons,connected with varying strengths. How much potentially contained?
@DescheleSchilder - There are a billion trillion stars in the universe. A hundred trillion neurons wouldn't be enough to maintain knowledge of every star even if each had thousands of dendrites. And that's just stars. John's got it right. Not remotely possible.
@jdunlop Well thats my point. There are more combinations of different connections and varying strengths than there are planck surfaces on the surface enclosing the cube.
@DescheleSchilder - that's... not your point at all. There aren't more combinations of different connections than there are stars in the universe, because neurons don't work that way. It's not remotely a soft yes. It's just impossible.
@jdunlop I dont speak of the connections themselves. Obviously there cant be more than there are atoms in the brain.
@jdunlop If one neuron connects to 100 trillion other neurons and the connection strengths can be varied than 1 neuron can make a combination already with 100 trillion others. Every other can make a connection with 100 trillion minus one. Etc. And then the strengths can also be varied.
This is exactly how I used to think too. But it is plainly wrong. Sorry...
Let's scale down the numbers to make it comprehensible. Suppose there are 10 "nodes" in your cube. Each node can connect to anywhere from 0 to 9 of the other nodes. That's 2^9 = 512 possibilities. Times 10 for 10 nodes = 5,120. (I'll assume connections are directional so I don't have to divide for the duplicates.) Does that mean that with 10 nodes I can store 5,120 "facts"? It's not that simple. How are you making the connections between nodes? There must be something to make this connection, whether it's a physical wire or a stream of positrons or whatever. ...
... So you don't need just 10 "things" to record 5,120 facts. You need 5,120 things. And in any case, none of this helps you at all. If there is some way to connect "things" in your cube that is non-physical and "doesn't count" for whatever reason, then the same must be true of things outside your cube. So you still need just as many things in your cube as there are things outside your cube. For your cube to hold all information in the universe, it must be the size of the universe.
@DescheleSchilder short answer your brain can't it can't even come close to envisioning all the objects in the universe. and the brain is not simple, your brain can store about 3 petabytes of information, but your brain also contains 10^26 atoms and about a 100 billion cells. And yet the brain cannot even model itself perfectly.
@Jay I have done a calculation. The number of paths between 100 trillion interconnected neurons exceeds the number of atoms by quite an order. If you also assume different possible cojlping strengths the number is even bigger.
@DescheleSchilder which means those 100 trillion interconnected neurons can model the interaction of 100 trillion atoms, or about 14 orders of magnitude fewer atoms than are in the brain. Also the brain does not have that many neurons, the brain only contains about 100 billion neurons.
@John The number is many many orders (that phone...) bigger than all elementary particles in the universe! 10exp14x(10exp14-1)x(10exp14-2)x.. Im talking not of the brain but that silly cube.
Can't we use lossless compression techniques (akin to a zip file) to make the model smaller than the universe itself? Don't get me wrong, the model would still be mind-bogglingly enormous—certainly larger than 1000 m³. But I don't understand why it has to be a 1-to-1 ratio.
@MJ713 compression doesn't work on analog data, to digitize analog data, and thus compress it, you have to have finer sampling that the original analog data, the simples possible perfect model of the universe is another identical universe. keep in mind all the atom in the universe have multiple pieces of information which also interacts with other atoms. For X atoms you don't need X bits, you X bits times the number of variables/interactions, so you don't just need a bit for each atom you need one for every interaction an atom can have with another atom TIMES the total number of atoms.
@DescheleSchilder You are making some incorrect assumptions in your calculations about the "number of connections". Specifically, by using a factorial, you assume that each neuron (or neuron-like structure within the cube) is directly connected to every single other neuron. This is not even close to physically possible within the given space constraints. There's not enough room. Or to approach it from another angle: each "wire" between any two given neurons must itself be made of one or more atoms, so the number of connections must always be less than the number of atoms in the system.
@DescheleSchilder Correction to my previous comment, above: your calculation is not a factorial. I forgot that a factorial is adding numbers together, and you are multiplying them instead.
Discussion on answer by John: Can we …
Discussion on answer by John: Can we …