New Mars Forums

The observable horizon is not the same as the horizon-of-future-action. Since the expansion of the universe is accelerating, there are things we can observe today which at some point will red-shift to infinity and become unobservable and unreachable. I think it is quite likely that something so distant while within our past is outside of our future light cone.

I think you may still be falling short, Tom. The concern is not that some AI bot might be evil like Hitler, but it might be beyond evil like Cthulu or an arachnid intelligence. Amoral can sometimes be more scary than evil.

EDIT: You may find this article and its sequel relevant and enlightening.

Terraformer, Kurzweil would say that you're still thinking linearly. The pace of progress has increased considerably since Victorian times. The equivalent unit of progress of a few decades in Victorian times would probably be a few years today. Predictions a few years forward are still pretty accurate.

Regarding AI, the most powerful supercomputer in the world today, the Tianhe-2 in China, is powerful enough by some estimates to perform whole-brain emulation of a human. The limit on that pathway is our understanding of neurobiology, not computer hardware or software. As for de novo AI, the field is finally receiving significant interest and funding, after a many-decade winter. I would use progress over the last few years -- which has been rapid -- as a reference point rather than the past 20-30 years, during which time nearly nobody was throwing money at the problem.

Terraformer, well exponential growth won't happen forever, but there is plenty, plenty of room for growth when looking at physical limits. I find it far more likely that the point where progress begins to become sub-exponential is at a technology level far more advanced than mainstream authors like KSR tend to portray as our future. Equally important, none of the advances so far have significantly affected the limitations of biology -- the human brain and human body. We've seen minimally super-exponential progress with regular technology development. Once that technology starts affecting our intellectual capabilities -- e.g. true artificial intelligence or bionic brains -- we'll see that super-exponential term grow as well. Until we hit physical limits of course, but as I said that's pretty far away from our current position.

That's the argument at least, although we have gone significantly off topic now. I'd be happy to debate it elsewhere.

Wow, can we get back to Kim Stanley Robinson?

Josh, have you read Icehenge? It’s also set in a parallel universe and the book predates the Mars trilogy.

You could say this is a defining characteristic of KSR’s space fiction – he reuses a roughly similar but not identical timeline in multiple, sometimes unrelated stories. I have to say I much prefer this over the assumption that all plots happen in the same universe. This is fiction after all., and each story is told to accomplish some purpose. If there are adjustments to be made that allow him to better accomplish his story telling, so be it.

The complaint about the pace of technology development is dead on though. Like almost all science fiction, KSR uses the linear view of history in projecting forward. The simultaneously joyous and sad reality is that we will have transcended biology and reached physical limits of computation before the first cities are being built on Mars.

Mars is more likely to be used as a staging point for the construction of the first Jupiter brain than colonized by us primates. But that makes for less accessible science fiction.

Robert, the EML2 point has the significant advantage of lower delta-V requirements since it's pretty much at the edge of the Earth-Moon system, gravitationally, and also has ready 24/7 access to the Moon and/or Earth for gravitational slingshots. It does make sense to have a space station there, more sense in fact than LEO.

Of course direct to Mars or LEO rendezvous makes far more sense for early missions to Mars. In the long term an EML2 station makes sense, but right now it's just pork.

Impaler, the Outer Space Treaty stipulates that nation states are held liable for hardware left in space if someone in their jurisdiction owns the hardware. What naturally evolved out of that is a passing on of that liability to the private interests, which requires insurance by law to avoid the bankruptsy loophole. And no insurer would be willing to take on the risk from dead weight in a dangerous orbit with no station-keeping or avoidance fuel.

Well the annoying thing is that by international treaty you need to provide full lifecycle guarantees about any hardware you put up there. That's part of why we never kept shuttle external tanks up there, even though that would have been a tremendous resource. Without a plan for dumping them in the ocean if/when station-keeping fuel is used up, you'll never get regulatory approval. It's a hurdle that can be overcome, but "let's just keep it up there until we come up with a use" is not a good enough justification...

2nd stage recovery simply doesn't make sense. You kill a lot of performance by adding that weight and extra fuel so high in the stack. Better economics to make the 2nd stage simple and cheap.

Doesn't mean it can't be reusable though. You could start collecting these things in Earth orbit and reuse them on in-orbit constructed crafts.

I assume you've read Freitas et al on this subject:

http://www.islandone.org/MMSG/aasm/
http://www.molecularassembler.com/KSRM.htm

I also advise you to talk to some of the people on ##hplusroadmap on Freenode (IRC), the only place I know of where people are actually working on this.

They are vitrified, not frozen: ice crystals do not form. You cannot repair the damage done by vitrification with today's technology. But no one is suggesting that be done. The idea is to repair the damage (and fix whatever 'killed' you too) with future technology. And no, there is no physical law preventing that from being possible.

Besides if cryonics is an experiment, being vitrified is better than being in the control group.

The virtual plasma would be at rest relative to the drive, no matter the speed of its starship relative to Earth.

Nuclear reactors are not analogous. A nuclear device is an extremely complex precision device, requires purified fuel that is scarce in nature, and results in results in life-destroying radiation and temperatures. This device on the other hand seems to be a simple construction -- just a microwave resonant cavity, really -- without poisonous side effects. It's the kind of thing that an undirected optimization process like natural selection would be able and likely to stumble upon. So the fact that we don't see anything like it in nature, is meaningful evidence against a free energy interpretation.