Why Future Humans Will Become Godlike Supercomputers

Table of Contents

• A small set of initial assumptions

  • 1. No faster than light space travel

  • 2. Entropy can't be reversed

  • 3. Physics is only locally flawed

  • 4. Digital life is possible

• Mining Stars, Self-Replicating Robots, and Exponen …

  • 1. Labor: Self-Replicating Robots

  • 2. Time: Exponential Growth

  • 3. Matter: Budget one planet mass? Peanuts!

• Biology vs Computronium

  • 1. On Planets:

  • 2. On Space Habitats:

    • a. Air to breathe:

    • b. Natural Landscapes: Plants, Rivers, and Hills

    • c. Material Abundance & Not Dying

    • d. Gravity

  • 3. In Digital Worlds

• Future civilizations, including ours, will follow …

• Epistemic Status

Humanity will inevitably become a digital species.

Your descendants won’t be beings of meat. They will inhabit supercomputers. They will think at light speed.

This is not an ideological transhumanist take. This is inevitable. Physics and a basic grasp of evolutionary theory are all one needs.

This is the logic of evolution. Genes that are better fitted to the environment will survive in the long run.

So, what is the environment that matters? How can one discuss genes in the context of predicting the future?

These are hard questions. 

I believe the logic of physics and evolution can tell us what we need to know. Those systems are incredibly complex. Yet, some details are general enough to be useful for making the predictions I want to make.

Physics describes the environment. Its limits are what everything must adapt to.

Genes are simply structures. Structures that replicate themselves. And that create something in the world. Those creations carry the genes. If they're well-adapted, the genes will survive. Otherwise, they disappear.

Nothing else is needed to foretell the future. The very long-term future.

Why live in a damp cave and burn your food storage for warmth?

So, how does one go about making such a prediction?

One examines the structures that exist in the present.

One considers what sort of structures are possible under known science.

One considers what the goals of the systems, people, and beings; the expressions of the Genes, are.

Then, one sees how such a path might progress.

One truth reigns supreme in this analysis when looking at humanity and its environment.

PLANETS AROUND STARS ARE SHITTY REAL ESTATE.

The issue is Entropy. There is only so much available matter and energy in the universe. Nature tends to a state of perfect balance. Stable, static, eternal, and dead. This is where things are going. Everywhere and Always.

Yet, life is the process of locally reversing entropy. It will fail in the end. But there can be so much beauty. So much experience. If one is wise about using the available resources.

So, how wise are we about using our entropy resources?

Light is the precondition for all life on Earth. The Earth gets about 1.74 * 10^17 Watts from the sun. As this energy is produced via fusion in the sun, we can calculate how much hydrogen was burnt to keep us warm.

Einstein’s E = mc^2 tells us that the sun uses more than 2 kg of mass per second to keep us cozy.

Consider that this is not like burning 2 kg of coal. The coal oven uses up the chemical binding energy. All the atoms are still there.

The 2 kg of mass are gone forever. 

Yes, most of the mass of the hydrogen atoms is turned into Helium. But some of it is “lost”. As gamma rays, kinetic energy, and neutrinos. 

True, some of it gets chemically stored on Earth. But, most of it disappears into the great, infinite blackness. The potential of that energy is gone, and disorder in the universe has increased. There is no way to put Humpty Dumpty back together.

But it gets worse. The sun doesn’t just illuminate the Earth. It illuminates equally in all directions. Think for a second. How many kilograms of hydrogen does it use per second?

Well, it produces 3.828 10^26 watts. This absolutely and utterly uses up 4.26 10^9 kg of hydrogen.

Yes.

Over 4 million tons per second.

That’s the mass of the Great Pyramid of Giza. Each second.

We are living in a damp and dangerous cave. Earth.

That is heated by a wildfire in the forest in front of it. The Sun.

We, and our children, our very future, need to live off that forest. But it is blazing away.

And we think this is good? And would anything sane accept that state of being?

A small set of initial assumptions

To discuss how sane sentients are likely to handle this situation, we need to agree on the ground rules of this discussion. I will propose several fantastical-seeming concepts.

The basic assumptions are the following:

1. No faster than light space travel

This prediction market on Manifold Markets suggests an 87% chance of this. 

Prediction markets are the most reliable tools for discussing the future, so I will trust them.

FTL would change all the rules. The faster it is, the worse it becomes. More importantly, it messes with physics on a deep level. According to modern physics, time travel is one of the main consequences.

So, FTL would automatically violate conditions 2 and 3.

2. Entropy can't be reversed

No one has figured out a way around the dictates of thermodynamics so far. 

Should someone do that, the fundamental interactions between life, physics, and economics change. 

Resources suddenly don’t matter. 

Running out of energy just ceased to be a thing. 

I won’t make any prediction about a world of physically infinite resources. That changes the logic of all the fundamental limits, which allows me to construct a picture of plausible future behavior in the first place.

3. Physics is only locally flawed

This is a pet peeve of mine. It is a common idea that Einstein disproved Newton's ideas of physics. This is correct, but it is too simplistic.

Physics, all science for that matter, lives off models. To keep a model that doesn’t approximate reality to a useful degree, one must be a social scientist. 

Engineering keeps the natural sciences honest. 

Subjective estimate about how much technology depends on what sort of physics

The table suggests that much of technology relies on wrong and outdated science. How can that be?

Pragmatically speaking, a technology only needs a sufficiently good model to explain it. A better model isn’t necessary. Or even helpful. 

Calculating the stability of a bridge using quantum physics is possible, assuming you have an infinite amount of time.

No, what happens is that local inconsistencies in the model are found. 

Newton works unless you consider the very small or the very fast and massive. Suddenly, things break down. That is essentially what physicists try to do with those huge and expansive particle accelerators.

Here, things get interesting. We developed Relativity and Quantum Physics, respectively. We gained new capabilities by finding the “place” where higher resolution was needed and useful.

So, that was a fascinating tangent. Why was it necessary?

Most of our knowledge remains valid as long as we only ever find local “exploits” in the world. 

We may develop new theories of physics to describe and manipulate the core of a neutron star. This might give us new fascinating capabilities, like supercomputers operating at the limits of physics.

But, and this is what matters for this point, the rest of physics keeps being usefully correct. We don’t figure out how to use the cores of neutron stars to edit reality at our whim.

As with the point above, that changes the rules. So much that I can't talk about it anymore in a coherent fashion.

4. Digital life is possible

Finally, this essentially means that evolution and design can keep optimizing consciousness. The exact details don’t matter that much.

Minds that think using ever faster, denser and more efficient “brains” will become the dominant life forms. Over millions of years.

There might be limits. At some point, your computer could collapse into a black hole. Your insurance won’t like that.

But I only need those minds to approach those ultimate limits over a very long time. Machine-gun-wielding monkeys putting neuro-links into their brains is a sufficient first step.

As long as this happens to be true, I can predict how evolution will shape life in the long term. Though I can't speak about the path it will take.

Mining Stars, Self-Replicating Robots, and Exponential Growth

Anything beyond those ideas are just massively scaled-up versions of things that are trivial. Star-spanning supercomputers? A trillion city-state space stations with the land area of Asia around our sun? Deconstructing a star? Seems ridiculous? Well…

Consider the idea of an American carrier battle group. Take on the perspective of an ancient Roman. A mighty formation of ships, which together weigh about 200,000 tons.

The Roman Empire produced about 80,000 tons of iron annually at its peak. 

For the needs of its entire economy.

Modern America uses over 2 years of total Roman iron production for a naval formation. Oh, and they have 11 of them. 

So when I start talking about solar system scale supercomputers, consider this:

Can we build one percent of one percent of one percent of something like this? 

What if you take a lot of time, several times the Earth's mass in materials, and robots building copies of themselves and the project itself into account? Could you conceive that we could do it with that?

Let's go with maybe for now. To discuss the future of human civilization, we need to address three core questions. I’ll address each of those in turn.

To build anything, you need three things: 

1. Labor: the ability to form matter into a desired shape

2. Time: changes at any scale won’t be instantaneous

3. Matter: resources you can change from one configuration into another

The solutions to all of those problems are as follows: 

1. Labor: Self-Replicating Robots

All life is made up of self-replicating machines. You are a copy of LUCA, the last universal common ancestor. A copy with a lot of beneficial errors added onto it, but nonetheless a copy.

LUCA came into being about 3.6 billion years ago. Probably in some deep-sea thermal vent. It was just some small biological machine, alone on a dead rock.

Yet, “only” 1.6 billion years later, copies of it had taken over almost the entire Earth. Sure, they were just microbial mats. But they started from a minuscule space. And they took over their entire world.

Today, we are building robots. Machines building more machines is an old idea.

• Gerald O’Neil worked out that some initial space infrastructure and population growth would change the solar system.

• Eric Drexler discusses how nanotechnology would be the perfect tool for this.

• NASA has run studies on doing this on the moon in the form of robotic factories.

• Canadian Professor Alex Ellery is seriously trying to revive this old idea. 3D printers are a promising avenue, though many obstacles must still be overcome. 

The fact is, automated manufacturing is coming. And as soon as you have robots capable of building robots that build more robots again, your productivity curve looks like this:

2. Time: Exponential Growth

This leads to the solution to the time problem. Exponential growth is the most powerful force in the universe. Let's take a closer look at it.

You might remember this from school. A pond has algae in it. They cover one unit of area. Each day, they double the area they cover.

If the pond has an area of 256 units, how many days until the pond is full?

Well, final_amount = initial_amount * exponential_base ^ number_of_generations.

Plugging in the numbers: 256 = 1 * 2^(number of generations).

This means we need Log2(256) or 8 generations.

As one generation lasts one day, it only takes 8 days.

The interesting thing is this: in the beginning, nothing seems to happen. Then, 25% of the algae-filled pond fills out a further 25% on the second to last day. On the last day, 50% of the algae-filled pond fills the entire pool. 

Ok, enough trivial math problems. How does this relate to building structures bigger than the sun? In our lifetimes, no less.

The sun outputs 3.828 * 10^26 watts.

Say that we start with one self-replicating factory.

It consumes 10^9 watts, one Gigawatt, of solar energy. Consult the table below for the size of the collector. This depends massively on its placement.

Let's go with conservative production numbers. Let's say it takes 100 days to make a copy of itself.

Let's modify our trusty formula: final_amount / amount_used_per_factory = initial_amount * exponential_base ^ growth_steps.

Plugging in the numbers: 3.828 10^26 watts / 10^9 watts = 1 2^(number of generations).

Solving log2(3.828 * 10^20) gives us 68.375 generations. Let’s make that 69 generations. Nice. 

Now, we have said that one machine generation takes 100 days. So, our factories only need 19.4 years to use all the sun's energy.

Does that sound amazing? Yes.

Is it overly optimistic? Yes.

Do logistical, resource, and thermodynamics lengthen the time? Yes.

But even if it takes 5 times as long, we can still transform the solar system in 100 years. That is safe to assume. Even the modern, bureaucratic nation-state rarely messes up projects that badly.

3. Matter: Budget one planet mass? Peanuts!

Okay, I have shown that we can get the “labor”. I have explained that we can do those things in a reasonable amount of time.

But, one can already hear something in the distance. Muffled, angry economist noises. Mysteriously finding their way from the byzantine, damp depths of their dark dwellings.

“How Can We Pay For This?”

The detailed answer to that question will be another post. On the economics of self-replicating machines.

The short answer? This is the best return on investment proposal that is possible.

Then, why have we not built those machines yet? 

Well, we are talking about the civilization that figured out the can opener about 50 years after canned food. Figuring out that relays, electromagnetic switches, can be used for calculations in computers took us over 100 years.

The rule of thumb for “obvious” technological applications is this: 

It takes several decades. Unless it is related to Porn. Then, it will happen instantly.

But even if money is a non-issue, how do we get all the materials? Building structures that are bigger than the sun needs a lot of, well, structural materials.

To abbreviate what I will work out in the replication-economy post, consider this:

Exponential growth can scale up massively.

Planets have a lot of useful mass. They can be disassembled.

Mercury and Venus are about as useful as New Mexico and Alabama. They can and should be put to better use.

Yet, two planets are just our start-up capital. Consider the diagram below.

Nearly all of the solar system's resources are in the sun. The same sun whose wasteful resource usage we have already discussed.

So, we get all the resources we need from the sun. Can we even do that?

Lifting a kilogram of mass from the sun's surface into Mercury's orbit takes about 1.9 * 10^10 Joules. 

One way to do this is to heat a section of the sun's equator with lasers. The lasers impart momentum to the plasma most loosely bound to the sun, and the particles speed up. 

And escape the sun. 

Towards a collection point of our choosing.

Here, we can sort our bounty with magnetic fields. 

Assume our process uses solar energy at 10% efficiency.

That means each kilogram requires about 200 of our 1 Gigawatt, 50 by 50 m satellites. Let's assume that the industrial chain behind each satellite requires about 10,000 tons of materials.

Mercury masses 3.3 10^23 kg. Assuming we scrape off 1% of its mass, we get 3.3 10^21 kg. We construct 3.3 * 10^14 power satellites with their whole 10,000-ton industrial chains.

We need 200 power satellites to get one kilogram of matter from the sun.

With that we can mine 1.65 * 10^12 kg per second. The sun is 74% hydrogen, 24% helium, and 2% all the “other stuff”. Luckily, the sun's heat keeps the mixture even throughout the star.

We are interested in all the “other stuff.” Carbon, oxygen, iron, silicon, … all the things that matter in life.

So, we get about 33 * 10^9 kg of “useful stuff” per second. To put it in more graspable terms, this is about 2 Great Pyramids. Or 3300 new solar power satellite industrial chains.

So, we have more than enough materials available. Better still, we don’t need to vandalize the solar system to get them. The process will quickly become self-accelerating.

Until we use all the energy, the sun has to offer. And turn the Sun into something more sustainable.

Biology vs Computronium

So we, or any sane alien, will alter their sun. Nothing else is sustainable in the long run.

To further predict how advanced civilizations will behave, we must consider how they will live. There are three options here:

1. On Planets:

Humanity currently lives on Earth. 

Earth has a population of 7.9 * 10^9 humans. 

It masses 5.97 * 10^24 kg. 

Its total surface area is 5.1 10^8 km^2. For a rather generous definition of hospitable, maybe a quarter of that 1.275 10^8 km^2 is livable.

This means that each modern human has on average, about 0.065 km^2 of total surface area available. Or about 0.016 km^2 of actually useful surface area.

All of that includes the atmosphere, gravity, space, and the production that keeps us alive. At the low "price" of 7.6 10^14 kg per person. Put another way, each square meter on Earth costs 1.17 10^10 kg of useful matter.

That is pretty costly. And inefficient. And the result isn't particularly safe. Volcanoes, Hurricanes, Earthquakes, Tsunamis, …

2. On Space Habitats:

To bring human habitation into line with common-sense construction standards like “doesn’t randomly start spewing liquid rock,” we need to build the environments from the ground up.

Let's see how that could be done. Our total material budget includes all the "useful stuff" in the solar system. The material is divided into the following abundances.

How would one go about optimizing the Lebensraum per unit of mass? Humans like a variety of things. Including but not limited to:

• Air to breathe

• Natural Landscapes: Plants, Rivers, and Hills

• Material Abundance & Not Dying

• Gravity

The question is this: How much of what material do we need to fulfill those needs?

a. Air to breathe:

• Oxygen comes to mind first. It is somewhat vital.

• Carbon dioxide is produced by humans and used by the plants we like.

• Nitrogen is a bit of a complicated matter. Plants need it, and humans need it for proteins. 

  However, the N2 in the atmosphere is chemically unusable. Specialized nitrogen-fixing bacteria make it available to plants. The plants then provide it to us.

• Thermal Regulation: The air also determines how well things cool down and how fast and hot fires burn.

• A chemically inert gas will be decent in this role. So, we can choose and mix using Nitrogen, Helium, and those otherwise useless 6.5% Neon.

An Earth-like atmosphere masses roughly 10 tons per square meter. All the required elements are either very abundant or can be flexibly replaced. No bottlenecks here.

b. Natural Landscapes: Plants, Rivers, and Hills

Getting all the stuff for this requires only a little bit more advanced chemistry. The high abundance of oxygen guarantees that this won't be a bottleneck.

Water is H2O. All the stuff we call rock is mostly oxygen. Yes. You walk on Air. Metalized air.

Oxygen will react in various combinations with iron, silicon, magnesium, aluminum, carbon, and sulfur to get us the rocks we know and love. For example, Earth consists of about 30% of Oxygen. Most of that is bound up with Iron (30% of Earth) or Silicon and Magnesium (15% of Earth each).

Furthermore, we cheat by making mountains hollow. Over those metal skeletons of the landscape, we lay 5 to 15 meters of landscape.

Landscapes vary in density, depending on the setup. Per square meter, we get the following weights for one meter of the following substances:

• Water: 1000 kg/(m*m^2)

• Soil: 1100 to 1600 kg/(m*m^2)

• Rock: 2600 to 3000 kg/(m*m^2)

Lakes and rivers will be limited in depth, but the rest will not be too different from nature on Earth. When did you last dig a hole deeper than 5 meters?

For nature, we need all the organic elements: hydrogen, carbon, oxygen, nitrogen, sulfur, and phosphorus, as well as a huge variety of trace elements.

Phosphorus is a potential bottleneck here. The solar system might only contain about 2.33% of the Earth's mass of it.

The bad news is that, on the high end, an ecosystem has 50 kg of living biomass per m2.

The good news is that phosphorus only makes up less than 1% of that biomass.

This means that we can "only" cover 546000 times the Earth's surface area in biomass. 

Assuming Tropical vegetation levels. 

And phosphorous consumption on the biological high end.

With clever design, we might cover more than 54.6 million times the Earth's surface area with life. That ought to keep even the most discerning connoisseur of nature's beauty occupied for some time.

How much mass are we going to need for this?

The biomass will never amount to much in aggregate. 50 kg per square meter is a bit higher than in tropical rainforests. Swamps might build up more, but those are outliers.

Half of that is carbon. Most of the rest is oxygen. We already discussed the non-issue of phosphorus. Material-wise, we will not run into limitations.

The landscape itself will be set at 15 tons per square meter. This might include 5 meters of basalt rock. Or 10 meters of deep water with some sort of seafloor.

We use Steel, Carbon-Supermaterials, and Aerogels for the hollow below-ground structures. I'll budget 950kg per m^2 because that goes nicely with the biomass.

In total, the landscaping will cost 16 tons per m^2.

c. Material Abundance & Not Dying

We humans care about both. Figuring out which is more important is left as an exercise for the reader.

On a space habitat, they are connected. This category includes:

• self-replicating robots building and maintaining the habitat

• all the industries needed to keep humans wealthy

• all the solar collectors

• the outer structural shell and armor of the habitat

• and so much more.

All in all, it would be a fool's errand to try to tally this in detail. I'd be trying to predict the supply chain of a future civilization. That's a whole post in itself, even by this blog's standards.

Because I like round sums, we will estimate that this takes a generous 14 tons per m^2.

Regarding materials, I'd expect a wild mix of biology, carbon-supermaterials, metals, and ceramics. Ceramics are oxidized metals. That would turn, be an economical use of the very abundant oxygen. 

Overall, I don't find any obvious and hard-to-substitute bottlenecks here.

d. Gravity

Jokes aside, gravity is a necessity. Unless the posthumans of this future bioengineer themselves live without it. In microgravity, water or air bubble habitats.

Even having Earth's standard gravity might not be necessary. Or desirable.

There are two schools of thought on humans in different gravity.

One is that anything except Earth standard, 1G, is unhealthy.

The other is that the issues we observe in freefall result from the absolute lack of gravity. Even Luna's 17% of Earth's gravity might be perfectly healthy.

We care about this because structural support consumes a lot of mass. If you can reduce the need for it, you will do so. As a species. Over several millennia. Or much more quickly. If you are smart.

Evolutionary pressures aside, what element and how much do we need?

Carbon is the undisputed heavyweight champion here. Carbon Nanotubes have a fantastic tensile strength. They are the only viable material for the large habitats we want.

This means we are going for rotating cylinders, so-called McKendree Cylinders.

They are huge. Diameters go up to 2000 km. If you insist on having Earth's standard gravity. Otherwise, you can go much larger. Or layer many into each other.

Steel or other metals are not worth considering. This is for 3 reasons. 

• They are orders of magnitude weaker than carbon materials. 

• They are rarer than the carbon materials. 

• They are needed for rocks, ceramics, and technology.

This, alongside the issue of reduced gravity and biological adaptations, makes estimating the mass hard. I'll guess that it is between 0 and 10 tons per m^2.

An efficient habitat design gives us Lebensraum at much better price. At most 50 tons per square meter. Potentially much lower if you get creative with your environment and biology.

How does this compare to planets? Well, per square meter, we use 0.00043% of the mass we need on Earth. Per person assuming we stay with the 0.016 km^2 livable area on average, we only need 8 * 10^5 tons to invest per person.

How many people might live in Habitats around our star?

The two bottlenecks that might limit the total population are phosphorus and carbon for structures.

Carbon is a somewhat lax limit. Freefall, Zero-Gravity habitats need very little carbon. Reinforced Water Ice is a valid option if you just want a wall.

Let's run with our 10 tons of carbon per m^2 estimate. That was how much carbon-supermaterials we budgeted for the habitat structure, which provided rotational “gravity.” And assume that we get to use 75% of the carbon in the solar system. The rest is needed for biology and industry.

Running those numbers, we get 3 * 10^17 km^2. That's about 600 million times the surface area of the Earth.

So, the bottleneck is phosphorous. 

Even if we go for sparse ecologies. 

Even if we get creative with genetic engineering. 

54.6 million times the Earth's surface area is all we can cover in life.

Assuming that this space sustains population densities like Earth, how many people would the Habitat-based solar system maximally contain?

We calculated that on Earth, each person has an average of 0.016 km^2 livable area available. This would mean that this solar system has about 1.9 * 10^19 people. The modern Earth would be a minor and obscure county in that civilization.

Could all the other matter in the solar system be used for life as well? How can a sentient being get by without phosphorus?

Before we turn to that, we can conclude this:

Space habitats are not only possible, safer and more efficient. No, they are also much more customizable than planets.

3. In Digital Worlds

What if evolution and economic pressures lead to further optimization? How could that work?

Cramming more and more people into less and less space will only get you so far. Ultimately, biological humans can’t take too much cramming.

But we are already on a path around that. People in the Western world spend between 8 and 12 hours daily in front of an artificial screen.

You are a bunch of wet neurons in a jar of bone.

Our senses receive signals from the world. The brain only receives those signals. It generates what we call reality from them, leading to the idea of a brain in a jar.

How many brains in jars could we support?

Let's assume that the infrastructure to run a brain in a jar weighs between 100 kg and 10 tons. I’ll assume half of that mass is carbon to simplify the calculations. And that there are no hidden bottlenecks.

So, there is roughly  5 * 1027 kg of carbon in Sol. Let's assume we use 70% of that for brain-in-a-jar machines. 

Running the numbers gives us between 4 10^23 to 4 10^25 human minds that could exist. They would live in digital worlds. They could live lives like ours but would have a wider range of abilities.

Economically, they will only depend on energy, some matter, maintenance, and data. The first three are trivial and unimportant unless they break. Data defines those people's worlds.

For us, it isn’t that different; we just don’t think much about it. And the first three aspects matter much more. Our world is a huge analog computer. All the data is processed by the matter around us.

While brains in a jar offer technological challenges beyond our current abilities, they haven't been shown to be fundamentally impossible. Embodied cognition means one probably needs to isolate more than the brain. Still, a brain and nervous system in a jar being required is fully within my assumptions.

Ultimately, this is an engineering issue. Using whole human bodies is well within the 10-ton end of the mass estimate. 

Once the baseline is established, optimization will make the devices more mass-efficient. Over the long run, that is inevitable.

Where would this trend of digitalized cognition and greater mass efficiency take a species?

Towards the physical limits of computation. 

Things don’t get ever more efficient. They get ever closer to the physical limits. 

Take heat engines exploiting the heat difference between two points in space, for example. How efficient can it get?

The efficiency is given by 1 - lower temperature / higher temperature. This is known as Carnot’s Theorem. It tells you how much energy the best physically possible engine can generate from a temperature difference. It doesn’t get any better than that.

The limits of computation are a more complicated question. This article provides a good, though somewhat technical, overview.

The gist is this:

Quantum physics limits how fast one can switch a bit. 

Lightspeed limits how fast information can travel in a computer. 

Miniaturization helps with the communication speed issue. Though, one can only miniaturize so far before that computer becomes a black hole.

So, the technical data for an ultimate computer looks something like this:

• Mass: 1kg

• Size: 1.485 * 10^-27 m (much smaller than any known subatomic particle)

• Operations per Second: 5.4258 * 10^50

• Bits: 3.827 * 10^16 (how much data it can store)

This is the hard limit. Any smaller than that, and you have a black hole, not a processor. This device might never be buildable. Yet, it provides us with a theoretical maximum.

Comparing that to the brain

So, how does a human brain perform in comparison? The direct analogy with computer terms can be tricky.

Still, common estimates state between 10^15 to 10^16 operations per second. Let's assume that running the perceptions, communications, and the mind's digital reality takes 10^17 operations per second.

Expressing the storage capacity of the human mind in bits is even more tricky. Human minds don’t use binary data. Research on this is ongoing, but recent studies suggest that estimating a petabyte of data storage will be safe for our purposes. 

Those are 8 10^15 bits, about the amount of data on the web in the mid-1990s. Again, let's assume we need an order of magnitude more storage per mind. That would mean we need about 8 10^16 bits per person.

All of those figures could easily be off by several orders of magnitude. But we are after a rough estimate about the deep future. So, take all of this with a barrel of salt.

The next problem is what sort of material would matter for this. 

We will use mass. 

At the densities of the ultimate computer, atoms no longer matter. How many minds could live on them if we could turn 1, 10, and 100% of the solar system's mass into those ultimate computers?

What does all of this tell us?

First, the ultimate number of sentient beings possible in a solar system is enormous. The number of minds that can be stored in the available bits is equivalent to the number of atoms in about 1 liter of water.

Second, the number of operations per second outweighs the available bits by about a factor of 10^30. What does this mean? 

Speed. I won't run the numbers here. This is a whole separate post about the nature of digital civilization. Still, those digital minds will experience each second of “real” time as centuries or millennia.

Future civilizations, including ours, will follow a path dictated by evolution and physics

The central dichotomy of civilization is this: Life is the local reversal of entropy.

In a universe where there is only so much energy to do that, efficiency is vital.

So, the fate of any sentient being is whatever allows the greatest, long-term, and stable complexity in combination with the highest efficiency.

The path might be slow. Walking it can take millennia or even millions of years.

Not all branches of a civilization will embrace or even like their fate. Some will stubbornly cling to a less efficient past.

This is as it should be. Evolution will take care of them. Or it won't. Some animals have lived in this world unaltered for over 200 million years.

Yet others will follow the siren song of physics towards the shores of optimality.

They will become a majority. Even if they start out small, they will eventually be legion.

Their societies will be richer. This richness allows them to find better-adapted structures faster. And so they will win the future.

Epistemic Status

This is speculative, drawing on extrapolations from established science. 

I've invested about 5 hours researching for this post, though I've contemplated these ideas for several years. 

Evaluate this based on your judgment. I am confident that my conclusions logically follow from the assumptions made below. My argument offers an elegant and simple model of the future of humanity.