3 ms·

> I think you are underestimating the challenges they have to overcome to get machines to work in space. Even if it’s hard, it’s still a lot easier than gettin

by hk__2 6y ago

> I think you are underestimating the challenges they have to overcome to get machines to work in space. Even if it’s hard, it’s still a lot easier than getting humans to work in space.

akiselev 6y ago
You don't understand the challenges. Sending humans up with semi-automated machines would actually be a lot easier than fully automated machines. The human race's entire experience so far has been with machines in 1g. We have almost zero experience making self repairing factories in any gravity. We have zero experience mining in low or zero gravity. It might take several generations of engineers and miners in space to figure out all of the problems, slowly automating each step at a time through trial and error like we've been doing for centuries. The critical point you're missing is that any mining and processing equipment will require their own life support that will already be able to support humans. Machines are significantly more complicated without a gas to convect heat away from all surfaces, they'll need massive fluid handling and filtering systems for lubricants and cooling fluid, the mining operations will need tons of oxygen and scrubbing for their chemistry, and clean water will be required by pretty much every process. The only way we'll pull this off is by using our existing experience on this planet and that requires the same conditions that support life. That sounds like I'm exaggerating but we've already spent more than a generation in space studying the effects of zero-g and we just barely got a working 3D printer on the ISS a few years ago. No advanced mining, processing, or other manufacturing equipment to speak of beyond lab scale R&D devices. We have one lab capable of continuous zero g research and a few 747s capable of achieving it a a few minutes at a time.
- runarberg 6y ago
  The problems you list seem to apply to both fully and semi automatic machines. E.g. wouldn’t human operated mining equipment need just as much lubrication as pre-programmed ones? We never explored the surface of Mars, or landed on an asteroid before we had our machines do it, but we managed it anyway with the human operators still on Earth.
  - cortesoft 6y ago
    I think the difference is that humans can fix problems and adapt to changing conditions a lot better than machines currently.
    - paulryanrogers 6y ago
      Humans plus Earth-like environment. Remember that micro environment has to go with them everywhere, and they die within minutes without it functioning within relatively narrow constraints.
    - thu2111 6y ago
      But there'll still be humans. Just not necessarily physically at the mine or factory. All you really need is a robotic chassis that has sufficiently good AI to use hands and follow basic instructions, perhaps helped by a guy in a VR suit back on earth or in a nearby spaceship. Perhaps aided by simulations to ensure low latency - things in space should be easier to simulate without the outside world interfering much.
      - yellowapple 6y ago
        > perhaps helped by a guy in a VR suit back on earth or in a nearby spaceship. Definitely not "back on earth". Lag is an issue with VR as-is with mere milliseconds of it, and you'd be calling for minutes of latency there. And "in a nearby spaceship" would, definitionally, involve getting humans most of the way there; might as well get them all the way there at that point.
        thu2111 6y ago
        I think it's interesting to explore whether remote-control VR lag can be combated with simulations. If you can simulate the other side well enough, you should be able to manipulate the simulation until you've got a recording that you're satisfied with and then send it. This is kind of how Mars rovers are steered today. Obviously doing this for robots on Earth is tricky and largely pointless, given it's probably easier to just reduce the latency directly. For remote mining on asteroids or something there's not a whole lot going on in space. You're not going to get a random pedestrian or cyclist suddenly popping out in front of your robot. It's far more practical to model and simulate it entirely in e.g. a game engine, and then record it. At least, that's my guess where things will go. The point about the spaceship is not distance related but that building life support structures into everything is quite expensive, whereas building them once in a spaceship and then extending the human presence to all surrounding structures would be a lot cheaper.
- yellowapple 6y ago
  Exactly. I can't remember which NASA official I'm paraphrasing here (I think Gene Cernan?), but humans are the cheapest, most abundant, and most "fully automated" machines known today.
  - TeMPOraL 6y ago
    And self-replicating.
Benjammer 6y ago
Is it really that much easier to build space robots to cover as much breadth of function as one astronaut-engineer in a space suit?
- hk__2 6y ago
  > Is it really that much easier to build space robots to cover as much breadth of function as one astronaut-engineer in a space suit? No, but it’s tremendously easier to fly these robots there than doing the same with humans. We can afford to let robots “die” on Mars; humans not really.

by hk__2 6y ago

> I think you are underestimating the challenges they have to overcome to get machines to work in space. Even if it’s hard, it’s still a lot easier than getting humans to work in space.

akiselev 6y ago

You don't understand the challenges. Sending humans up with semi-automated machines would actually be a lot easier than fully automated machines. The human race's entire experience so far has been with machines in 1g. We have almost zero experience making self repairing factories in any gravity. We have zero experience mining in low or zero gravity. It might take several generations of engineers and miners in space to figure out all of the problems, slowly automating each step at a time through trial and error like we've been doing for centuries. The critical point you're missing is that any mining and processing equipment will require their own life support that will already be able to support humans. Machines are significantly more complicated without a gas to convect heat away from all surfaces, they'll need massive fluid handling and filtering systems for lubricants and cooling fluid, the mining operations will need tons of oxygen and scrubbing for their chemistry, and clean water will be required by pretty much every process. The only way we'll pull this off is by using our existing experience on this planet and that requires the same conditions that support life. That sounds like I'm exaggerating but we've already spent more than a generation in space studying the effects of zero-g and we just barely got a working 3D printer on the ISS a few years ago. No advanced mining, processing, or other manufacturing equipment to speak of beyond lab scale R&D devices. We have one lab capable of continuous zero g research and a few 747s capable of achieving it a a few minutes at a time.

runarberg 6y ago
The problems you list seem to apply to both fully and semi automatic machines. E.g. wouldn’t human operated mining equipment need just as much lubrication as pre-programmed ones? We never explored the surface of Mars, or landed on an asteroid before we had our machines do it, but we managed it anyway with the human operators still on Earth.
- cortesoft 6y ago
  I think the difference is that humans can fix problems and adapt to changing conditions a lot better than machines currently.
  - paulryanrogers 6y ago
    Humans plus Earth-like environment. Remember that micro environment has to go with them everywhere, and they die within minutes without it functioning within relatively narrow constraints.
  - thu2111 6y ago
    But there'll still be humans. Just not necessarily physically at the mine or factory. All you really need is a robotic chassis that has sufficiently good AI to use hands and follow basic instructions, perhaps helped by a guy in a VR suit back on earth or in a nearby spaceship. Perhaps aided by simulations to ensure low latency - things in space should be easier to simulate without the outside world interfering much.
    - yellowapple 6y ago
      > perhaps helped by a guy in a VR suit back on earth or in a nearby spaceship. Definitely not "back on earth". Lag is an issue with VR as-is with mere milliseconds of it, and you'd be calling for minutes of latency there. And "in a nearby spaceship" would, definitionally, involve getting humans most of the way there; might as well get them all the way there at that point.
      - thu2111 6y ago
        I think it's interesting to explore whether remote-control VR lag can be combated with simulations. If you can simulate the other side well enough, you should be able to manipulate the simulation until you've got a recording that you're satisfied with and then send it. This is kind of how Mars rovers are steered today. Obviously doing this for robots on Earth is tricky and largely pointless, given it's probably easier to just reduce the latency directly. For remote mining on asteroids or something there's not a whole lot going on in space. You're not going to get a random pedestrian or cyclist suddenly popping out in front of your robot. It's far more practical to model and simulate it entirely in e.g. a game engine, and then record it. At least, that's my guess where things will go. The point about the spaceship is not distance related but that building life support structures into everything is quite expensive, whereas building them once in a spaceship and then extending the human presence to all surrounding structures would be a lot cheaper.
yellowapple 6y ago
Exactly. I can't remember which NASA official I'm paraphrasing here (I think Gene Cernan?), but humans are the cheapest, most abundant, and most "fully automated" machines known today.
- TeMPOraL 6y ago
  And self-replicating.

Benjammer 6y ago

Is it really that much easier to build space robots to cover as much breadth of function as one astronaut-engineer in a space suit?

hk__2 6y ago
> Is it really that much easier to build space robots to cover as much breadth of function as one astronaut-engineer in a space suit? No, but it’s tremendously easier to fly these robots there than doing the same with humans. We can afford to let robots “die” on Mars; humans not really.