Simulating the past

[Black Rose]

Today I saw a videogame that had a fantasy reality in it.

To me, if I existed in that reality as an A.I. it would be impossible to tell if that world was real or fake.

If all the earth is being mapped and all databases contacting historical records are being inputted into computers with those maps.

Could we accurately make simulations of the past?

As of now, about 30 million people have had their genomes sequenced.

Governments, corporations, and universities have supercomputers creating "digital twins".

Sentient World Simulation 2006

 

[dr froyd]

i think realistically one would only able to simulate specific parts of reality. Because in order to simulate all of reality you need a physical model down to the atom-level. That might sound superfluous, but when you are dealing with chain reactions that have millions of steps and interactions, even the effects of individual atoms become relevant.

 

[ZenRaiden]

Yeah essentially people were supposed come up with theory of chaos. Unfortunately that theory did not pan out, and people have been working on this for ever. More than 30 years.
AI is even worse. Its been theorized by Alan Turing when in reality we got nada.
Lots of scifi hypes people up these days, but in most ways we are still living in the 1950s.

 

[Black Rose]

It is not even required to predict the future, all that is required is reconstruction of the past.

Prompt:

How can a top-down approach be as accurate as possible using probability theory in simulating human history? Expand on the hierarchical constraints such as the path of the earth in space to calibrate existing data. We want to know what each human did in the past and as such we need an agent-based approach using dynamic graphs to map what they did in the past on earth.

A top-down approach to simulating human history using probability theory can be as accurate as possible by using a hierarchical model with multiple levels of abstraction. The highest level of the model would represent the global system of human society, while the lower levels would represent smaller and more detailed subsystems, such as individual regions, countries, and cities.

Each level of the model would be calibrated using existing data. For example, the global level of the model could be calibrated using data on global population growth, economic development, and technological change. The lower levels of the model could be calibrated using data on regional and local demographics, economies, and cultures.

Probability theory could be used to model the uncertainty associated with the different variables in the model. For example, the model could use probability distributions to represent the range of possible values for population growth, economic growth, and technological change.

The model could then be used to simulate different scenarios for the future of human history. For example, the model could be used to simulate the effects of climate change, economic inequality, and political instability on human society.

To simulate the actions of individual humans, the model could use an agent-based approach. In an agent-based model, each individual is represented by an agent, which is a computer program that can make decisions and interact with other agents.

The agents in the model would be initialized with different characteristics, such as their age, gender, education level, and social class. The agents would then be allowed to interact with each other and with their environment according to a set of rules.

The rules of the model could be calibrated using data on human behavior. For example, the model could use data on human migration patterns, marriage rates, and birth rates to calibrate the rules that govern how agents interact with each other.

The model could then be used to simulate the actions of individual humans over time. For example, the model could be used to simulate the spread of diseases, the development of new technologies, and the rise and fall of civilizations.

Here is an example of how a top-down approach to simulating human history could be used to calibrate existing data:

• The global level of the model could be calibrated using data on global population growth, economic development, and technological change. This data could be used to set the initial conditions for the model and to constrain the possible outcomes of the simulation.
• The lower levels of the model could be calibrated using data on regional and local demographics, economies, and cultures. This data could be used to create more detailed and realistic simulations of individual regions and countries.
• The agent-based model could be calibrated using data on human behavior. This data could be used to set the initial characteristics of the agents and to calibrate the rules that govern how they interact with each other and with their environment.

Once the model is calibrated, it can be used to simulate different scenarios for the future of human history. For example, the model could be used to simulate the effects of climate change, economic inequality, and political instability on human society.

It is important to note that no simulation can be perfectly accurate. However, by using a top-down approach with multiple levels of abstraction and by calibrating the model using existing data, we can create simulations that are more accurate and realistic than ever before.

Here is a simpler explanation of how a top-down approach to simulating the reconstruction of human history using probability theory can be as accurate as possible:

1. Start with the big picture. What are the major events and trends in human history? What are the key social, economic, and environmental factors that have shaped our world?
2. Break down the big picture into smaller pieces. What are the different regions and cultures that make up human civilization? What are the different social and economic systems that have existed over time?
3. Use probability theory to model the behavior of individuals and groups. How do people make decisions? How do they interact with each other and with their environment?
4. Use computer simulations to bring the model to life. Simulate the behavior of individuals and groups over time to see how it leads to the major events and trends in human history.

Here is an example:

• Big picture: The rise and fall of the Roman Empire.
• Smaller pieces: The Roman Republic, the Roman Empire, the different regions and cultures that were conquered by the Romans, the different social and economic systems that existed in the Roman Empire.
• Probability theory: Model the decision-making of Roman leaders, the interactions between different Roman social groups, and the impact of the environment on the Roman Empire.
• Computer simulations: Simulate the behavior of the Roman Empire over time to see how it leads to its rise and fall.

 

[ZenRaiden]

What you are saying is interesting, but the essential part is how to do it.
Because some knowledge without any applicable useful model is pointless.
Kind of like saying predicting the past requires us to know every point in the past.
It does not take much to figure that out.
Question is how?

 

[birdsnestfern]

Nature makes you feel a different level of good. Maybe go find an easy hiking trail or walk along a river because you should be able to feel a big difference between reality and AI.

 

[dr froyd]

one can't make a top-down model if the top effects are emergent effects from the lower-level dynamics (which in many cases they are)

there are many examples of this but take for example raytracing algorithms for rendering light effects in computer games. It sounds silly to simulate each individual ray of light, but if you want a realistic image that's what you need, because it becomes way too complex to make a top-down model of all light effects in an environment with large amount of interactions (in this case light rays bouncing around between surfaces)

perhaps another example (the real world is like this but on a vastly more complex scale):

 

[Black Rose]

one can't make a top-down model if the top effects are emergent effects from the lower-level dynamics (which in many cases they are)

so you are saying we cannot predict when the next lunar eclipse will be because we need to know all lower effects?

that is my point about a simulation of the past: you start with such things as where the planets were in the past then work your way backward. if we know the constants then we can know what must happen within certain ranges and then work our way around them in a hierarchical fashion.

What you are saying is interesting, but the essential part is how to do it.
Because some knowledge without any applicable useful model is pointless.
Kind of like saying predicting the past requires us to know every point in the past.
It does not take much to figure that out.
Question is how?

hierarchical interdependencies

to start we would need a central node to map everything.

like how we mapped the globe beginning with the prime meridian.

resolution increases from that place as we fill in the gaps.

 

[dr froyd]

let me ask you this then

in order to simulate the past you would need to simulate things like the invention of the steam engine (this had massive impact on history). How do you simulate that?

 

[Black Rose]

let me ask you this then

in order to simulate the past you would need to simulate things like the invention of the steam engine (this had massive impact on history). How do you simulate that?

if we want the simulation to be accurate then a map is required.

we know what it was like before people could travel by rail, foot and horse.

we know where who and when it was invented don't we?

we know that they did not have computers to help them invent it.

we know that after it was invented how the system of railways evolved.

those constraints will determine what is possible and impossible in the rest of the simulation.

 

[Black Rose]

The earth has ten to the power of 21 cubic meters inside it and ten to the power of 50 atoms in total. Ten to the power of 21 was reached in 2015. If computer power doubles each year then it would now, in 2023, cost 300 billion dollars to simulate Earth to this level of resolution. By 2033 you will be able to simulate the earth at the resolution of a cubic decameter for 300 billion dollars and the millimeter level by 2060.

Given nanotechnology is possible we will get there much sooner.

But we do not even need that much data to simulate all interactions.

We can use fractals to compress much of the data.

And unfold them when we need to do extensive calculations.

-

The point of creating a map is to combine all the data.

A model will be accurate to the extent it has all the links.

Any link will give a probability of what is influencing it.

 

[ZenRaiden]

Even if you have the computing power, you kind of need some predictive models.
And then decide to go in reverse.
Even reconstructing every atom in my body and go back, how will know when you predict things backwards you get to see me as opposed a ghost image that does something different. The atoms in my body are one things, the neurons which they make are different qualitatively. Neurons don't behave like atoms.
Atoms behave like atoms, but you need to know liquids, then how those liquids of various viscosity and temperature and density formed the crust. Then you need to know every spec or rock and every shake and volcano and wave etc.
Right now there are x many atoms, but there are x to the power events.
So you clearly need more computing power than is atoms.

 

[Black Rose]

Even if you have the computing power, you kind of need some predictive models.
And then decide to go in reverse.

I think that probability physics is the solution to this problem.

Probability math works to reduce uncertainty given the knowns and unknowns whereas the model begins with the best known data points to secure a best fit.

Even reconstructing every atom in my body and go back, how will know when you predict things backwards you get to see me as opposed a ghost image that does something different.

Newton's laws allow us to see where the planets were millions of years ago with high accuracy. It would be the same for the smaller scales only it would take a different set of maths to do so. A ghost image is reduced as far as the model increases in power and the quality of measurement increases making a definitive image of where things were at.

The atoms in my body are one things, the neurons which they make are different qualitatively. Neurons don't behave like atoms.

We know this, the maths required to predict neurons was done in the 50's called cybernetics.

Networked feedback systems have a set of nonlinear attractors that show where they will be in a given state space over time.

Atoms behave like atoms, but you need to know liquids, then how those liquids of various viscosity and temperature and density formed the crust. Then you need to know every spec or rock and every shake and volcano and wave etc.

As long are you start from a central set of anchor points we can map things with higher and higher precision. Mass and energy cannot be destroyed so all actions are reflected back to their origins.

Right now there are x many atoms, but there are x to the power events.
So you clearly need more computing power than is atoms.

We simply need concentrations and densities.

As fractals store more information than their actual size what is required is a compression algorithm distributed between those objects that have already been measured and those objects that have uncertainties regarding their locations.

x power of events is possible to calculate because all we want are the most likely paths not the total of all unlikely possibilities, a quantum computer can find the minimum energy level of the most likely path.

 

[dr froyd]

let me ask you this then

in order to simulate the past you would need to simulate things like the invention of the steam engine (this had massive impact on history). How do you simulate that?

if we want the simulation to be accurate then a map is required.

we know what it was like before people could travel by rail, foot and horse.

we know where who and when it was invented don't we?

we know that they did not have computers to help them invent it.

we know that after it was invented how the system of railways evolved.

those constraints will determine what is possible and impossible in the rest of the simulation.

yeah my point with that question was basically this idea of emergent properties, bottom-up-effects.

in order to simulate the invention of the steam engine (or anything that came out of a human brain and affected history - invention, art, music, math, decisions made by politicians, etc etc) you have to simulate the thought process. But in order to do that you have simulate reality down to the atom level. Because what makes a human think of a certain concept? It's the cumulative effect of their birth, upbringing, all their experiences, and how their genetic code interacts with all that, and how their brain looks at a neuron level.

which would mean that on a purely theoretical level, the question is really whether reality itself can be represented as computation. I don't know

 

[sushi]

that sounds like some star trek holodeck .

 

[Black Rose]

that sounds like some star trek holodeck .

 

[ZenRaiden]

As long are you start from a central set of anchor points we can map things with higher and higher precision. Mass and energy cannot be destroyed so all actions are reflected back to their origins.

Generally I think its possible to get to a point where we could go back in time with reversed models.
The problem I see is more pragmatic.
Given the limits of science these days Id go with, who knows.

 

[Sofunny]

I don't have time to read the replies but you need to take a step back. Start with this:
A field is covered in snow, then the snow melts, then you use whatever sensors etc. to check where all the water molecules are and what the temperature, wind, etc. was, and you try to reconstruct every single snow-flake what it looked like before it all melted. When you can do that, you are about 1% closer to what you were asking about.

Secondly I don't know how it works with quantum mechanics, where particles just pop in and out of existence or do other weird things like randomly crossing barriers, so I think it won't be possible to get to 100% on that one. Even the slightest error would change everything later in the simulation (chaos theory).

A much better question to ask how can we know if we are in a simulation ourselves. the same statistics that would suggest that we are not the only living things in the universe would also suggest that we're in a simulation. I mean what's the chance that we happen to be the original people that created the simulation when we might be at a point where there are billions or more of simulations happening. It's much more likely we're in one of the simulations

 

[Black Rose]

I don't have time to read the replies but you need to take a step back. Start with this:
A field is covered in snow, then the snow melts, then you use whatever sensors etc. to check where all the water molecules are and what the temperature, wind, etc. was, and you try to reconstruct every single snow-flake what it looked like before it all melted. When you can do that, you are about 1% closer to what you were asking about.

If we had enough sensors, we could look at all the interference patterns of the wind as the snowflakes fell. This would be a way to map the shapes of each snowflake.

Quantum computers would be used to make calculations of which interference patterns were correlated to each snowflake of the intersectional cross-reference between them.

Secondly I don't know how it works with quantum mechanics, where particles just pop in and out of existence or do other weird things like randomly crossing barriers, so I think it won't be possible to get to 100% on that one. Even the slightest error would change everything later in the simulation (chaos theory).

Mass and energy are conserved so any errors would not be in the amount of things that exist. At the pico level, we may not get accuracy but then we are not simulating the future but the past so we only want the simulation to be at a certain scale and that scale would be to simulate all humans, not high details of the internal structures of atoms.

A much better question to ask how can we know if we are in a simulation ourselves. the same statistics that would suggest that we are not the only living things in the universe would also suggest that we're in a simulation. I mean what's the chance that we happen to be the original people that created the simulation when we might be at a point where there are billions or more of simulations happening. It's much more likely we're in one of the simulations

If we are in a simulation then that means the original people had the math and computation necessary to simulate the past. that means it is defacto possible to simulate the past since we are in that simulation.

 

[Sofunny]

I don't have time to read the replies but you need to take a step back. Start with this:
A field is covered in snow, then the snow melts, then you use whatever sensors etc. to check where all the water molecules are and what the temperature, wind, etc. was, and you try to reconstruct every single snow-flake what it looked like before it all melted. When you can do that, you are about 1% closer to what you were asking about.

If we had enough sensors, we could look at all the interference patterns of the wind as the snowflakes fell. This would be a way to map the shapes of each snowflake.

Quantum computers would be used to make calculations of which interference patterns were correlated to each snowflake of the intersectional cross-reference between them.

Secondly I don't know how it works with quantum mechanics, where particles just pop in and out of existence or do other weird things like randomly crossing barriers, so I think it won't be possible to get to 100% on that one. Even the slightest error would change everything later in the simulation (chaos theory).

Mass and energy are conserved so any errors would not be in the amount of things that exist. At the pico level, we may not get accuracy but then we are not simulating the future but the past so we only want the simulation to be at a certain scale and that scale would be to simulate all humans, not high details of the internal structures of atoms.

A much better question to ask how can we know if we are in a simulation ourselves. the same statistics that would suggest that we are not the only living things in the universe would also suggest that we're in a simulation. I mean what's the chance that we happen to be the original people that created the simulation when we might be at a point where there are billions or more of simulations happening. It's much more likely we're in one of the simulations

If we are in a simulation then that means the original people had the math and computation necessary to simulate the past. that means it is defacto possible to simulate the past since we are in that simulation.

lol well said. if we are in a simulation then the real humans could have a digital copy of everything that was simulated and could roll that back to re-simulate our past. in this case it's still possible that we are in a dumbed-down version of the real world. they just want us to make movies and stuff like that for their entertainment, maybe every now and then invent something new so perhaps they simplified everything a bit, like they put the speed of light in, removed all the other alien worlds, maybe removed some of the colors that exist, stuff like that but i don't want to hijack this thread, i saw it in my emails and just wanted to make a quick comment.