DARPA Study on how to Terraform Mars

[Tom Kalbfus]

There are lots of golf courses in the Palm Springs area which is located in the desert...

Oh well, in which case the Sahara is practically done already. If Mars would take Tom 100 years to terraform, the Sahara which is closer, smaller and not toxic should take commensurately less time. Let's say 1/14 of the time due to the smaller area and Half that due to non toxicity and easier access to resources. Should be doable in about 3 years say?

Simon Hibbs

This is based on the availability of nanotechnology and artificial intelligence that is greater than human intelligence, the technology will be quite incomprehensible to us by 2130 unless we augment our own with artificial brain implants. Moores Law will guarantee that Machine Intelligence will surpass our own by 2040, and after that the machines will be in charge, hopefully, they'll have ingrained instincts to remain subservient to us, and we can use the technology they develop to conquer the stars and terraform new planets. Not only that, but also create artificial worlds, Bishop Rings, and larger. We can then send these Machine Intelligences out into the Galaxy to terraform and expanding sphere or worlds for us to eventually settle, and then get them out of our way once we have the technology we need to colonize these planets.

 

[dragoner]

Hey, I almost died here on earth from dihydrogen monoxide inhalation when I was a kid.

Disgusting stuff. You know what fish do in it?

It's all dinosaur pee anyway*.

Simon Hibbs

* Read the last half for the bit about dinosaurs. It's a classic.

They have even found Gerber putting it in baby food!

xkcd is great.

 

[Condottiere]

Viability in our specific situation would seem to be dependent on a Martian colony being independent, that is, self reliant on energy, food, water and air.

 

[simonh]

Moores Law will guarantee that Machine Intelligence will surpass our own by 2040, and after that the machines will be in charge.

Moorese law is about speed, not intelligence. A badly written computer program on a faster machine just produces useless results in less time, the faster computer doesn't in any way make the results any more useful or the algorithm any more intelligent. As of now nobody has even the slightest clue how to even design a general purpose intelligence system, let alone implement one. All we have are varyingly useful extremely special purpose, limited domain algorithms for solving specific problems. AI researchers have been predicting strong AI in the next 10 to 20 years since the 1960s and we're honestly not much closer to achieving it now than we were then. The more we learn about the problem, the clearer it becomes how far we are from solving it.

1960s Herbert Simmons predicts "Machines will be capable, within 20 years, of doing any work a man can do."

1993 - Vernor Vinge predicts super-intelligent AIs 'within 30 years'.

2011 ray Kurzweil predicts the singularity (enabled by super-intelligent AIs) will occur by 2045, 34 years after the prediction was made.

So the distance into the future before we achieve strong AI and hence the singularity is, according to it's most optimistic proponents, receding by more than 1 year per year. So I predict that when we get to 2045 strong AI will be on the slate to be achieved by about 2090.

I'm poking fun of course. I'm not at all saying that general purpose strong AI is not possible. We are living proof that it is. We'll figure it out eventually. I just think it's likely to take generations, not decades. The thing is, if you don't actualy have a plan for how to achieve something, you can't come up with a reasonable estimate of how long it will take. All you can do is make blind guesses.

We can't rely on magic wands like nanotechnology and AI to wish away our problems in the real world, no matter how much fun it is to do so in our fictional worlds.

Simon Hibbs

 

[heron61]

This is based on the availability of nanotechnology and artificial intelligence that is greater than human intelligence, the technology will be quite incomprehensible to us by 2130 unless we augment our own with artificial brain implants. Moores Law will guarantee that Machine Intelligence will surpass our own by 2040, and after that the machines will be in charge, hopefully, they'll have ingrained instincts to remain subservient to us, and we can use the technology they develop to conquer the stars and terraform new planets. Not only that, but also create artificial worlds, Bishop Rings, and larger. We can then send these Machine Intelligences out into the Galaxy to terraform and expanding sphere or worlds for us to eventually settle, and then get them out of our way once we have the technology we need to colonize these planets.

I'm uncertain but hopeful that hyper-intelligent AI will be possible. The sort of nanotechnology you're describing (that can terraform a planet in a remotely reasonable period of time) is self-replicating nanotech, and I'm vastly less optimistic about that. Far too many proposals I've seen about what self-replicating nanotech can do rely upon the sort of near infinite geometric growth we don't see in the natural world because the one form of self-replicating molecules we know of (carbon-based life) typically runs into some sort of limit fairly rapidly. Nanotech that can terraform Martian soil & rock, replicate itself using ingredients if can obtain from Martian soil and rock, and do both on a less than geological time scale sounds suspiciously like magic, and not the Clarke's third law sort as much as the wish fullfilment sort.

We know for a fact that carbon-based life takes a very long time to terraform a planet, creating a stable oxygen atmosphere took approximately 1.5 billion years. However, much of that involved very primitive life. So, say you use life that is specially engineered to make oxygen, and to oxidize rock and water, that's 1,000 times better than the life that actually did all this. Then the process "only" takes 1.5 million years. You are positing nanotech that is 15,000 times faster than this and 15 million times faster than the life that actually oxygenated our planet. I'm strongly betting that tech like that would break more than one physical law.

 

[Reynard]

And this explains the Long Night and Rebellion Era. Worlds with hostile environments to humans demand life support dependent facilities and not all have local access to the parts and materials to maintain it. When the ships stop coming the world dies. Unless any colony in our system develop intensive industrial infrastructure, they will always be vulnerable to the elements. This is far different than colonies on Earth set up in places not so different than the home land and makes future attempts at independence easier.

 

[heron61]

And this explains the Long Night and Rebellion Era. Worlds with hostile environments to humans demand life support dependent facilities and not all have local access to the parts and materials to maintain it. When the ships stop coming the world dies. Unless any colony in our system develop intensive industrial infrastructure, they will always be vulnerable to the elements. This is far different than colonies on Earth set up in places not so different than the home land and makes future attempts at independence easier.

Absolutely. I'm certain that after the Long Night, explorers visiting various very thin, exotic, and vacuum atmosphere planets that had been inhabited before found a whole lot of dead worlds (as well as presumably some that managed to survive and thrive). Inhospitable worlds where the colony's life support failed and the last human died 1,000+ years ago would be deeply creepy and could definitely be interesting locations for adventures.

Throw in one or more of: Rumored lost treasure that others are hunting, dangerous native life that has broken into some of the dead colony (not likely on vacuum worlds, but possible elsewhere), or a heavily automated TL 12 or 13 colony which died but still has power and automation, which isn't functioning as well as it should (malfunctioning robots and automation are always fun) add in the inherent creepiness of the situation and let the PCs loose.

 

[Reynard]

You just described Marc Miller's Traveller Millieu 0 campaign.

 

[Condottiere]

Water on Mars exists today almost exclusively as ice, with a small amount present in the atmosphere as vapour.[1] The only place where water ice is visible at the surface is at the north polar ice cap.[2] Abundant water ice is also present beneath the permanent carbon dioxide ice cap at the Martian south pole and in the shallow subsurface at more temperate latitudes.[3][4][5][6] More than five million cubic kilometers of ice have been identified at or near the surface of modern Mars, enough to cover the whole planet to a depth of 35 meters.[7] Even more ice is likely to be locked away in the deep subsurface.[8]

If this is true, it would form the basis of a limited colonization of Mars, probably with a controlled population to ensure that the resources never become strained.

 

[Tom Kalbfus]

Moores Law will guarantee that Machine Intelligence will surpass our own by 2040, and after that the machines will be in charge.

Moorese law is about speed, not intelligence. A badly written computer program on a faster machine just produces useless results in less time, the faster computer doesn't in any way make the results any more useful or the algorithm any more intelligent. As of now nobody has even the slightest clue how to even design a general purpose intelligence system, let alone implement one. All we have are varyingly useful extremely special purpose, limited domain algorithms for solving specific problems. AI researchers have been predicting strong AI in the next 10 to 20 years since the 1960s and we're honestly not much closer to achieving it now than we were then. The more we learn about the problem, the clearer it becomes how far we are from solving it.

1960s Herbert Simmons predicts "Machines will be capable, within 20 years, of doing any work a man can do."

1993 - Vernor Vinge predicts super-intelligent AIs 'within 30 years'.

2011 ray Kurzweil predicts the singularity (enabled by super-intelligent AIs) will occur by 2045, 34 years after the prediction was made.

So the distance into the future before we achieve strong AI and hence the singularity is, according to it's most optimistic proponents, receding by more than 1 year per year. So I predict that when we get to 2045 strong AI will be on the slate to be achieved by about 2090.

I'm poking fun of course. I'm not at all saying that general purpose strong AI is not possible. We are living proof that it is. We'll figure it out eventually. I just think it's likely to take generations, not decades. The thing is, if you don't actualy have a plan for how to achieve something, you can't come up with a reasonable estimate of how long it will take. All you can do is make blind guesses.

We can't rely on magic wands like nanotechnology and AI to wish away our problems in the real world, no matter how much fun it is to do so in our fictional worlds.

Simon Hibbs

So how long do you suppose the human brain will remain an inscrutable black box, who's inner workings we can't analyze? Either we figure it out within this century or we discover the existance of the human soul and God by deductive reasoning, by saying there is something within the human cranial cavity that we can't simulate with computers, thus we have proved the existance of the Soul and possibly the Afterlife as well. How can that be? Easy somebody already beat us in the AI race and we are AIs ourselves, and out whole world including the Universe we observe through telescopes and our various instruments is simulated too. I think if we can't produce full AI within a century we must conclude that we are AI living within a simulation created by someone else. Maybe as a consolation prize, whoever created us will allow us to have FTL to create all sorts of interesting adventures for us instead of stranding up on just one simulated planet surrounded by stars we can't reach. We may need some "Divine Intervention" to get to the stars however or at least a change in the programming. Also maybe a civilization might suddenly pop up at Alpha Centauri just after we mysteriously "discover" the FTL drive. If that doesn't happen, we can always build our own artificial universe with multiple simulated planets perfectly suitable for sim humans to populate. While were doing that, we can also terraform some real planets for those reluctant to upload their consciousness into a computer.

 

[Condottiere]

We've got seventy odd years to accumulate and interpret data.

In theory, a computer that is continually upgraded and never turned off, will not only accumulate vastly more data, will have access to data accumulated in other databases, and will be able to process it remarkably faster than we can. Infinitely, if the infrastructure to support it's operations is maintained.

The linchpin would be the capability to accurately interpret that data and act on it.

 

[Tom Kalbfus]

This is based on the availability of nanotechnology and artificial intelligence that is greater than human intelligence, the technology will be quite incomprehensible to us by 2130 unless we augment our own with artificial brain implants. Moores Law will guarantee that Machine Intelligence will surpass our own by 2040, and after that the machines will be in charge, hopefully, they'll have ingrained instincts to remain subservient to us, and we can use the technology they develop to conquer the stars and terraform new planets. Not only that, but also create artificial worlds, Bishop Rings, and larger. We can then send these Machine Intelligences out into the Galaxy to terraform and expanding sphere or worlds for us to eventually settle, and then get them out of our way once we have the technology we need to colonize these planets.

I'm uncertain but hopeful that hyper-intelligent AI will be possible. The sort of nanotechnology you're describing (that can terraform a planet in a remotely reasonable period of time) is self-replicating nanotech, and I'm vastly less optimistic about that. Far too many proposals I've seen about what self-replicating nanotech can do rely upon the sort of near infinite geometric growth we don't see in the natural world because the one form of self-replicating molecules we know of (carbon-based life) typically runs into some sort of limit fairly rapidly. Nanotech that can terraform Martian soil & rock, replicate itself using ingredients if can obtain from Martian soil and rock, and do both on a less than geological time scale sounds suspiciously like magic, and not the Clarke's third law sort as much as the wish fullfilment sort.

We know for a fact that carbon-based life takes a very long time to terraform a planet, creating a stable oxygen atmosphere took approximately 1.5 billion years. However, much of that involved very primitive life. So, say you use life that is specially engineered to make oxygen, and to oxidize rock and water, that's 1,000 times better than the life that actually did all this. Then the process "only" takes 1.5 million years. You are positing nanotech that is 15,000 times faster than this and 15 million times faster than the life that actually oxygenated our planet. I'm strongly betting that tech like that would break more than one physical law.

It might be faster to manufacture a breathable atmosphere off planet if plant life won't do it fast enough, after all, it only takes energy to break the molecular bonds of carbon-dioxide, so the methodology would be this, we remove the carbon-dioxide from Venus, break it up into carbon and oxygen in space using solar collectors to gather the energy to do this, then we dump it back on Venus. We remove some hydrogen from Callisto, combine it with the oxygen we liberate from carbon-dioxide and dump it back on Venus, and if we do it in parallel in space rather than on Venus, we don't have to worry about the waste heat created in this process, we save the nitrogen we gathered, add most of it to the Venusian rocks as they cool to form nitrates, we probably would have to remove some soil and do it in space, we dump it back on Venus alone with excess nitrogen, oxygen, water vapor and a trace of carbon-dioxide for the plants, this could all occur within a century if we have nanotechnology to do this with. All this space manufacturing will also allow us to produce the mirrors that will give Venus a 24-hour day/night cycle and by the way a 360 day seasonal cycle as well. Mars could be handled in a similar manner, beyond those two, we can look in the Kuiper belt and Oort cloud for similar sized bodies we can terraform. Mercury can be terraformed, we'd have to add stuff to it which got boiled off by the Sun during its formation the Asteroid belt could be made into spinning habitats which we can inhabit. If we give Jupiter a giant artificial ring system, that will soak up the charged particles in its radiation belts and make its inner satellites habitable and terraformable

 

[silburnl]

Aerostats floating in Saturn's atmosphere would also work

Aerostats on Venus are also feasible - conditions are the most earth-like in the solar system at an altitude of ~50km (1 bar of pressure, temp in the liquid phase of water, radiation shielding from the remaining atmosphere) and a breathable mix of N2/O2 acts as a lifting gas in the venusian atmosphere.

Plus Venus is the easiest planet to get to in the solar system.

Regards
Luke

 

[Tom Kalbfus]

Aerostats floating in Saturn's atmosphere would also work

Aerostats on Venus are also feasible - conditions are the most earth-like in the solar system at an altitude of ~50km (1 bar of pressure, temp in the liquid phase of water, radiation shielding from the remaining atmosphere) and a breathable mix of N2/O2 acts as a lifting gas in the venusian atmosphere.

Plus Venus is the easiest planet to get to in the solar system.

Regards
Luke

Getting from is a diffrent story.