I’ve always been interested in space exploration. But I get impatient with excessively fantastic stories. What makes it exciting is the possibility, the challenge. Simply waving an authorial wand, to summon the necessary magical technology, is cheating. If you have any respect for science, engineering, and the real work of technological research and development, then it must be believable.
The purpose of space exploration is to acquire direct observational data of other planets, including those in other star systems. That requires sending observers. But it does not require them to be people. (Space colonization is another story. But it isn’t feasible until we have done some initial scouting and survey work.)
It’s well understood that this is currently beyond our capability. At least, we haven’t devised a strategy for doing it. Doing what, exactly?
What does it take to survey a distant world? We’ve already done it within our solar system. The question is how to extend our range by four orders of magnitude. The nearest star system is around seven thousand times as far away as Pluto. In order to get to Alpha Centauri, we first have to get through our own solar system, including the Kuiper belt, and then past the heliopause, including the Oort Cloud.
We don’t really know what might lie between here and the nearest star system. There is something there, even if it is extremely sparse. Is there useful material there? That’s where comets come from. Comets are made of ice. Ice contains hydrogen, and maybe deuterium and helium. If we had the means of generating energy out there, could we harvest those materials?
We can’t send people that far away. But we can send machines. The problem is that our machines are not sufficiently sophisticated. They cannot function autonomously. They would also have a similar problem to humans: lack of energy to function. So we would have to find a way to take a source of energy out there. Fusion is the obvious choice, if we allow for fantasy technology. But we can’t build a fusion reactor. Can we send a fission reactor?
We won’t be sending any complex machines into distant space until we’ve first sent them to closer places, first. The Moon, Mars, and into orbits or onto moons of the gas giants. The big challenge with sending anything is propulsion. The more mass, the more propulsion needed. The more propulsion, the more reaction mass. As with rockets, this could be made easier with multi-stage designs, to avoid carrying mass you don’t need, and by doing all the acceleration early—except that it takes just as much energy to decelerate. Although you can use friction and gravity to help slow you down, just like you can use gravity to help speed you up. We’ve already done this with various space probes.
Even if we can devise a system for getting an autonomous probe to another planet, we don’t actual have any truly autonomous robotic probes. That’s why we need some form of artificial intelligence, or whatever you want to call it. What we have now, in 2024, is not AI, even if it can do some cute tricks. Generative algorithms, no matter how impressive their outputs, are not even remotely autonomous. Maybe they could become so, if we can redesign them to use their sensors as their input, and behaviours as their output. But we would have to figure out how to remove the randomness and confabulation that characterizes their generation today.
Autonomy means the ability to observe and respond to one’s environment, in order to perform tasks, and avoid harm. In fact, we want to go even beyond autonomy, to the ability to reproduce. Or at least, if not make exact copies, than be able to produce other autonomous systems that can do other work, including harvesting raw materials, and manufacturing useful things, including factories and communications systems. But the holy grail is self-perpetuation.
Self-perpetuating robots is a dangerous idea. What if they get out of hand? They would need to be designed to limit their own propagation, and keep their own numbers under control. They would have no interest in working around their limits, unless someone was foolish enough to program them with strange lifelike drives to reproduced indiscriminately.
Then again, accidents happen. Bugs happen. If they happen out in space, far from Earth, it’s probably not a big deal. But who knows? That’s where we have to use our imagination, within limits. Horror stories are fun, but they don’t teach us anything useful about how to achieve difficult tasks. They only exist to scare us. They cannot provide useful insights about the future.
Robots that explore space don’t need to be self-aware, or otherwise similar to human beings. They simply need to know how to make sense of and adapt to their circumstances, in order to fulfill the tasks they have been assigned. They don’t have to have emotions, or feelings, or self-consciousness. They don’t have to appreciate art, science, or nature. They just have to do their jobs well. They aren’t artificial people; we shouldn’t think of them as such. Or build them to be such.
Everything interesting that happens on Earth is thanks to the energy of the Sun. In deep space, there is no Sun. There is no easy access to energy, unless our explorers bring their own. There is, however, matter. And we might be able to invent a number of different ways to send energy resources with them. We can send fissile material and a nuclear reactor, for starters. But maybe we can come up with something else. Or maybe there is fissile material out there to be discovered, although it’s unlikely to be easy to find.
Alternately, we can beam the energy of the Sun directly into space. Radiation from the sun spreads outward in all directions. So by the time it gets really far, it is very diffuse. But maybe we can find a way to reduce or eliminate that spread, at least in a very narrow corridor. We already know how to do that: lasers. The question then is how large of a laser can we make? And how accurately can we target it? Convert the laser light into electricity, charge a battery, power a machine. Ideally, you power a local factory that can harvest and refine raw materials, and then build new machines, including a new power planet, if you can find what you need to feed it.
The key to all of this is automation on a scale that we have never achieved on Earth. The challenge is to find the shortest route to self-perpetuation, which is when the machines can find energy to power themselves, and materials to build new machines, without human intervention. Initially, we have to sacrifice a small amount of the energy and materials (and talent) on Earth, in order to build that starter package. That is the initial investment.
Of course, such a self-sufficient machine economy would have serious implications for life on Earth, even if we dismiss exaggerated horror story predictions. It implies that most human labour is no longer necessary, because machines could do it all. It implies a drastic re-organization of the economy and how we live on Earth. It might not be appealing to everyone.
Some of those already in power will try to take advantage, and monopolize control over those machines. Greedy people refuse to share, no matter how much they have. All they really want is power over others, and force them to be servants. We must learn how to restrain such dangerous people. They have already done so much damage. If we don’t restrain them, and we succeed in building highly autonomous machines, the risk of harm is incalculable. Even if such machines have no understanding of good and evil, or life and death. Like any machine, they could still be used to do evil, unless we work together to prevent it.
One day, we may have to contend with the problem of robotic armies. Whether those are workers or soldiers, is up to us. We don’t need soldiers. We no longer need to conquer anything. We only need to explore, for the joy of doing so. This is going to be an issue, regardless of how you feel about space exploration. Humans like to build machines. The only question is what we will do with them.