Stephen Hawking: Why Isn't the Milky Way Crawling with Life?

[Wyrm]

Extrapolation from a single datapoint

Yes, that is your entire argument isn't it; "Self-replicating robots must be exactly like life because... because... because I'm too stupid to actually consider any of the real technical issues, so I am going to pretend they're irrelevant!"

Science builds on what is known, you fucking moron. You can't extrapolate from a total vacuum, which is exactly what you're doing when you spew about AI (and nanotech, for that matter). You think you can produce an AI without some sort of hierarchy of needs? Go do it, and once it passes the Turing test, we'll talk.

I admit the possibility is there, but until I see some goddamned evidence, you can go fuck yourself.

plus the fact that a sapience is useless if the hardware is wrecked.

So? How is that relevant to anything?

It means preserving the seat of that sapience, that is the hardware, is a primary priority, above any long-term goals. This really isn't hard.

Yeah, you may cry 'single datapoint', but it's better than extrapolating from 'no datapoint.'

Since you have proven yourself incapable of understanding functional arguments, I will simply point out that subsumption architectures (which is what your fumbling characterisation is actually grasping at) comprise a tiny fraction of current general AI research. There is Rodney Brooks (who pioneered the design) and his followers, Mark Tilden's BEAM analogue dead end and a handful of similarly patterned NN designs. Together they make up about 3% of the publication output and projects underway. There are other heirarchical control designs and most architectures are layered in some way, but not in a way that puts goal driven behaviour functionally 'below' the main intelligence.

But none of that has produced a full AI yet, has it? Here's a big spoiler: Most avenues of research are dead-ends, especially when the goal is so far into the future. Prior probability states that most of the above research avenues into AI will be ultimately morbund. At least we know that subsumption architecture intelligences actually work.

[Simon_Jester]

You're right. I reread the story, and it was a comment in the discussion thread indeed. No one proposes a concrete diplomatic approach in the story. I should've checked it before giving my example, probably.

No one could figure out how to talk to them.

How do you talk to a species that not only eats their own babies in large numbers, but considers the need to do so as the founding basis for all that is right and moral in their own culture? Especially given that you need to explain to them that:

1) You do not eat babies, and
2) You'd honestly be a lot happier if they stopped, but that
3) You aren't the Ravening Hordes From Andromeda that Must Be Stopped At All Costs.

Because when they find out you don't eat babies, they will consider you the Ravening Hordes From Andomeda. Obviously, if you don't have that basic linchpin of sanity and culture, they will think, you could be anything. You could be a complete monster, as demonstrated by your suggestion that they stop.
________

Civilization on Earth may well have been running against the clock to develop high technology. Barring technological intervention sooner or later (probably within a few tens of thousands of years) another ice age is bound to come, and it's an open question whether a preindustrial civilization could survive such a dramatic climate change.

Most Earthlike worlds would experience very dramatic periodic climate shifts because they'd lack a large moon like ours to stabilize their axis. On a world like that civilization might appear only to be knocked back to the stone age with the next major climate shift again and again.

I kind of doubt it. Most preindustrial technology is sustainable on the community level: you can have a self-sustaining Iron Age village in one little valley as long as they don't overlog the hillsides and keep digging bog iron out of the creek. So even drastic climatic shifts are unlikely to destroy much technical knowledge, even if they destroy enormous amounts of organization. It's only in modern industrial societies that our ability to keep using our knowledge depends on our ability to keep the social infrastructure running.

[Samuel]

No one could figure out how to talk to them.

How do you talk to a species that not only eats their own babies in large numbers, but considers the need to do so as the founding basis for all that is right and moral in their own culture? Especially given that you need to explain to them that:

1) You do not eat babies, and
2) You'd honestly be a lot happier if they stopped, but that
3) You aren't the Ravening Hordes From Andromeda that Must Be Stopped At All Costs.

Because when they find out you don't eat babies, they will consider you the Ravening Hordes From Andomeda. Obviously, if you don't have that basic linchpin of sanity and culture, they will think, you could be anything. You could be a complete monster, as demonstrated by your suggestion that they stop.

As I said, complete and total idiots. This is the level of thinking extreme religious fanatics and political zealots display. The inability to see that basic facets of your culture might not be shared by other beings.

Seriously, this relies upon them having a single culture which is pretty stupid, ignoring their own history and ignoring the possibility of alien aliens. What if you have xenos that are popped out by batches in test tubes? Ones that eat the aged? Ones that... I could go on, but they are retarded and it is too easy to shoot them full of holes.

I kind of doubt it. Most preindustrial technology is sustainable on the community level: you can have a self-sustaining Iron Age village in one little valley as long as they don't overlog the hillsides and keep digging bog iron out of the creek. So even drastic climatic shifts are unlikely to destroy much technical knowledge, even if they destroy enormous amounts of organization. It's only in modern industrial societies that our ability to keep using our knowledge depends on our ability to keep the social infrastructure running.

Nope. Tasmania lost just about all technology. A single village is NOT sustainable- a single famine, a single plague, a single tipping of the scales and things fall apart.

[Starglider]

Science builds on what is known, you fucking moron.

You are not doing anything that could be called 'science'. You cheerfully ignore mechanism, hell you can't even grasp basic differences between two entirely different types of entity; designed robots and evolved organisms.

You can't extrapolate from a total vacuum, which is exactly what you're doing when you spew about AI (and nanotech, for that matter).

On the contrary, many thousands of brilliant scientists and engineers have spent decades painstakingly developing the basics of these fields. In both cases enough knowledge has been built up to give a reasonable characterisation of the design space. Of course this is knowledge you do not have, and in your arrogance you assume that anything you do not know is worthless.

You think you can produce an AI without some sort of hierarchy of needs?

Not just me; the majority of people who have studied the problem seriously in an academic setting. You think you're smarter than all of them despite not devoting a lick of effort to the problems they have spent decades tackling. I would attempt to explain why a 'hierarchy of needs' doesn't imply a subsumption architecture but you clearly aren't interested in actual technical details.

It means preserving the seat of that sapience, that is the hardware, is a primary priority, above any long-term goals. This really isn't hard.

No, it is not. You are confusing systematic dependence with the causal flow of utility (or its equivalents in non-utilitarian goal systems). This is actually one of your least serious crimes, in that mistakes of this nature are very common in AGI beginners and dilettantes, but it is still a falsehood arrived at by the same mindless assumption that you use for all your other reasoning - that AIs must be like organic life. In a deductive goal system, 'preserving the hardware' is a subgoal which is implied by top-level goals. Even in designs where self-preservation is present as an independent goal, it is never placed 'above' other goals . Even Isaac Assimov, for all his crimes against accurate depiction of general AI, got this bit right - he made self-preservation the lowest priority of his three laws - but you didn't.

Together they make up about 3% of the publication output and projects underway.

But none of that has produced a full AI yet, has it?

So? Once again I will bring myself down to your level, where only the most superficial reasoning is allowed and functional detail must be ignored. Yet you still lose, because no organic life has produced interstellar starships yet, and it's an exponentially bigger jump in capability from bacteria to starships than it is from current AI prototypes to general AI systems.

Prior probability states that most of the above research avenues into AI will be ultimately morbund.

Willful ignorance of technical detail has left you with no way of judging which of those they are, so you are forced to assume the same 3% ratio for workable designs as for the entire field of designs currently being studied. Even your metric is broken; many NN designs are more biomorphic overall than Brooks type designs without being subsumption architectures.

At least we know that subsumption architecture intelligences actually work.

It's not as if AI researchers don't know that, or don't pay very close attention to human brains and organic nervous systems in general. Of course we do, but when they actually study neurology most people quickly realise that many of the structural details are byproducts of a tortuous evolutionary history with no value in a designed system.

[Junghalli]

The latter ain't no picnic either, due to the extreme difference in abundancies between hydrogen and deuterium, and the very marginal difference in their physical and chemical properties. Deuterium is a pain in the ass to extract on Earth, where natural processes have enriched deuterium to 115 ppm for normal water. In contrast, Jupiter has only 22 ppm deuterium.

If by "readily obtainable", you mean on the order of 22 ppm with long and energy intensive processing to enrich due to the very subtle differences between their physical and chemical properties. Yeah.

A fusion drive using deuterium or hydrogen, as I have said already, could easily be wilderness refueled.

22 ppm.

First off the amount of deuterium in water in the Comet Hale-Bopp was found to be twice that of Earth's oceans, so the assumption that you'll be working with the sort of low deuterium concentrations found in Jupiter's upper atmosphere is totally unfounded.

The abundance of deuterium in Comet Hale-Bopp in the form of heavy water was found to be about twice that of Earth's oceans. If Hale-Bopp's deuterium abundance is typical of all comets, this implies that although cometary impacts are thought to be the source of a significant amount of the water on Earth, they cannot be the only source.[34]

Deuterium was also detected in many other hydrogen compounds in the comet. The ratio of deuterium to normal hydrogen was found to vary from compound to compound, which astronomers believe suggests that cometary ices were formed in interstellar clouds, rather than in the solar nebula. Theoretical modelling of ice formation in interstellar clouds suggests that Comet Hale-Bopp formed at temperatures of around 25–45 kelvin.

Yes, it's Wikipedia, but I see no reason to not consider it credible unless you can give me a different statement from a better source.

Second, unless you're going to be running a deuterium-burning direct fusion drive you're not going to need much of the stuff. D-D fusion gives you 6.2 X 10^14 joules/kilogram. Assuming your self-replicating probe's industrial facility eats up 100 megawatts you could run it for 17 days on a single kilogram of deuterium, assuming 1/4 energy efficiency. You'd need more to provide energy for thermal or electric rockets, but even then unless your spacecraft is huge you're probably not going to need much. Assuming your spacecraft masses 100 tons, carries its own mass in propellant, and has an exhaust velocity like high-gear VASIMR (294 km/s) it can run through its entire propellant tank while only burning 21 kg of deuterium at 1/3 energy efficiency (plug 100 tons and 294 km/s into the kinetic energy equation and then divide the result by the energy density of deuterium fusion if you don't believe me). That gives you over 200 km/s delta V, which is plenty for going all over a solar system if you use modest energy transfer orbits. With these figures in mind the fact you'd have to process 26 tons of water to get 1 kilogram of deuterium assuming deuterium abundance similar to Hale-Bopp is not so horrible. Being able to jet all over a solar system even requires processing no more than hundreds of tons of water, you can easily find that even in a very dinky little comet, and as long as you kept a reserve of a few kilograms of deuterium you have plenty of energy for the filtration.

Of course, the bot is wearing out as you speak...

Of course, that strategy was a pretty crazy one anyway. I can think of at least three vastly better options than setting up shop on a KBO when you're dependent on solar power, starting with simply snagging one of those KBOs that have orbits that conveniently take them close to the sun at periastron (i.e. a comet).

You need startup rockets for that, dearheart.

You top off its tanks while you're braking. The magscoop can easily be turned into a brake; simply choke it off and slow down the hydrogen plowing into it. While you're at it, funnel some of that slowed-down hydrogen into your ship and store it away. Then when you want to move on to the next system use that as fuel until you're going fast enough to set out your scoop again. Or if you don't want to do that just mine some regular old hydrogen from the nearest convenient icy body; your engine needs to be able to burn it to run off the ISM plasma anyway.

Now all you need is the ridonkulously large (and flimsy) sail.

Easily manufacturable from space resources. You just need a very big, very thin piece of aluminum foil, and aluminum is readily obtainable from any reasonably metal-rich asteroid. You can also use carbon fiber, which would use even more readily obtainable raw materials. There are theoretical advanced designs for superior very light sails using carbon nanotubes, again the raw materials would be easily obtainable from any C-type asteroid or dirty snowball. The sail must be very large but it's also very thin, so for its size it doesn't use much material.

Remember that most of that is going to be infissionable U-238. You need to enrich that stuff to use it in a reactor.

What about using a thorium breeder reactor? Thorium is 40 ppb on the same chart.

So we just need a ready source of power and we're peachy.

Why not the local star? Silicon solar cells can be made from asteroidal silicon relatively readily (ref), and since you're in a zero-gravity environment you can easily build very large solar panel farms, or if you prefer you can use huge metal foil mirrors easily manufactured from asteroid resources to focus sunlight onto smaller solar panels arrays or simple boilers. Even in the inner system you're likely going to be able to find asteroids that have pockets of volatiles in their interior (according to this site a C-type asteroid could be as much as 10% water ice), and then there are the so-called "main belt comets" in the asteroid belt. Sure, the sunlight that far out is a lot weaker than it is in the inner system, but as I said, you can easily build very large solar panel arrays to catch it since you don't have gravity or wind to worry about.

I see nowhere in that quote or in the referring paper the energy cost per tonne of material processed. You're operating on an energy budget, you know.

Actually I'm not at all sure that the idea that you're "operating on an energy budget" is justified, at least insofar as you seem to be implying that the probe will have to operate under a burden of serious energy scarcity. It's pretty easy to construct very large solar power plants in space out of locally available resources.

I'm having a hard time trying to think of a situation where the bot will fulfill its mission by being wrecked.

I don't see what that has to do with anything. The point is you program the AI so that "I want to finish my mission" is the primary goal and "I don't want to be destroyed or damaged" is a secondary goal that it follows because being destroyed or damaged would impede or prevent it from fulfilling the mission, so there's no reason that it would resent being expected to shut down once the mission is over.

I have no idea what being 'a zombie (in the philosophical sense)' means.

A philosophical zombie is a creature that exhibits all the external signs of intelligence (such as being able to pass the Turing Test) but is not self aware. I use it here to indicate the possibility of creating an intelligence which is every bit as capable as a sapient one but lacks self-awareness. How hard or easy this would be depends on whether consciousness is

(1) An emergent property of sufficient mental complexity (it would be impossible without deliberately engineering subroutines that quash self-reflective thoughts as they arise, and that would probably mean performance sacrifices for an independent robot).
(2) An emergent property of survival instincts and/or self-monitoring (it would be easy, but it would be difficult to do it without sacrifices in performance for a fully independent robot).
(3) An emergent property of the reactive-emotional nature of our mental processes (you'd probably have to deliberately imitate organic mental processes to not make a zombie, and a zombie could be every bit as capable as a sapient in every respect, or perhaps even more capable).

It's fairly irrelevant though, as even a non-zombie can be made a completely nonthreatening and obedient servant if you get to design its mind from the ground-up.

You think you can produce an AI without some sort of hierarchy of needs? Go do it, and once it passes the Turing test, we'll talk.

The problem is not with the idea that an AI would have a heirarchy of needs, the problem is with the idea that its self-preservation must have the same place in the heirarchy of needs that it does in ours. A competent programmer would design our self-replicating probe so that fulfilling its mission is its first need, and self-preservation is simply an attendant need. "I want to finish my mission. If I do not survive the mission fails. Therefore I want to survive until I have finished my mission. After I have finished my mission it does not matter whether I survive or not. I will shut myself down and await further orders or the failure of my systems from entropy, whichever comes first."

It means preserving the seat of that sapience, that is the hardware, is a primary priority, above any long-term goals.

Why? The only way that making self-preservation a secondary goal would impede self-preservation is if self-preservation conflicted with one of the higher goals. Since we designed the goal system we'd very much like the intelligence to die in those circumstances. I don't see what the problem is.

[Wyrm]

Science builds on what is known, you fucking moron.

You are not doing anything that could be called 'science'. You cheerfully ignore mechanism, hell you can't even grasp basic differences between two entirely different types of entity; designed robots and evolved organisms.

Strawman. I acknowledge that there is a difference between the two. It just isn't relevant in the discussion.

You can't extrapolate from a total vacuum, which is exactly what you're doing when you spew about AI (and nanotech, for that matter).

On the contrary, many thousands of brilliant scientists and engineers have spent decades painstakingly developing the basics of these fields. In both cases enough knowledge has been built up to give a reasonable characterisation of the design space. Of course this is knowledge you do not have, and in your arrogance you assume that anything you do not know is worthless.

Translation: "People better studied than me have worked on the problem. Therefore, AIs have these characteristics because... because... I said so!"

As the saying goes, don't sing it, bring it. I made every effort to explain myself and my position. Convince me that the AIs we can use really have these characteristics.

Not just me; the majority of people who have studied the problem seriously in an academic setting. You think you're smarter than all of them despite not devoting a lick of effort to the problems they have spent decades tackling. I would attempt to explain why a 'hierarchy of needs' doesn't imply a subsumption architecture but you clearly aren't interested in actual technical details.

Translation: "Waahhh! Wyrm thinks he's smarter than me! So I'll accuse him of thinking he's smarter than the entire field! He's not interested because... because... I said so!"

Go fuck yourself, Starglider. You didn't even TRY to help me understand why you are correct. All your spew amounted to 'it just doesn't work that way' without even a token attempt to explain why. If you really understand any of this, you should be able to explain the gist of why you're correct.

It means preserving the seat of that sapience, that is the hardware, is a primary priority, above any long-term goals. This really isn't hard.

No, it is not. You are confusing systematic dependence with the causal flow of utility (or its equivalents in non-utilitarian goal systems).

Why does semantic dependence or causal flow have anything to do with it? How do they dismiss the fundamental observation that a machine wrecked through misadventure puts a definite kibosh on any future goals and doesn't do anyone, the machine or the creators, any fucking good?

This is actually one of your least serious crimes, in that mistakes of this nature are very common in AGI beginners and dilettantes, but it is still a falsehood arrived at by the same mindless assumption that you use for all your other reasoning - that AIs must be like organic life.

Again, this strawman. You think my comparing life and an AI robot and realizing the parallels in facing similar challenges somehow means that I think that AI's must be like life. Hate to break it to you, but it doesn't.

In a deductive goal system, 'preserving the hardware' is a subgoal which is implied by top-level goals.

So the subgoal... results in the preservation of the hardware which is pretty high on the 'to do' list, due to the fact that a wrecked machine hinders the top-level goal. How is this any different in practice to what I've already said? I found this subgoal by the simple observation that an AGI can't do jack-shit if it's wrecked, while you had to go through this ridiculous 'subgoal implied by top-level goals' rigamarole.

Even in designs where self-preservation is present as an independent goal, it is never placed 'above' other goals . Even Isaac Assimov, for all his crimes against accurate depiction of general AI, got this bit right - he made self-preservation the lowest priority of his three laws - but you didn't.

You know that humans can turn off the 'self-preservation' instinct by virtue of our intelligence for a higher goal, do you not? That's why WWII Japanese pilots could crash into US ships in service to their God-Emperor. We're not totally subservient to it; it just happens to be the default state.

So? Once again I will bring myself down to your level, where only the most superficial reasoning is allowed and functional detail must be ignored. Yet you still lose, because no organic life has produced interstellar starships yet, and it's an exponentially bigger jump in capability from bacteria to starships than it is from current AI prototypes to general AI systems.

Another goddamned strawman. If you bothered to read my (admittedly) long-ass response, you would have found, to your utter amazement, that I was not talking about bacteria:

Finally, the organisms would build a space ship the same way your nanobots would: by constructing larger and larger structures to help with gathering and manipulating resources. As for programming, if I'm building the organism(s) from the ground up, I can eliminate all the junk and code the instructions tightly, and genomes can code for a lot. The lowly Amoeba has 290 billion base pairs to play with, and if you work with communities of organisms, you can work with a library as sizable as any VNM you can build.

The obvious choice for the larger and larger structures: the creating race, or something like it. You know, the race that had sent the spacecraft in the first place and, so doing, demonstrated their ability to produce starships.

Now fuck off, you dishonest cuntrag!

Willful ignorance of technical detail has left you with no way of judging which of those they are, so you are forced to assume the same 3% ratio for workable designs as for the entire field of designs currently being studied. Even your metric is broken; many NN designs are more biomorphic overall than Brooks type designs without being subsumption architectures.

That's your excuse for everything, isn't it? I'm 'willfully ignorant', that I have 'no way of judging' the matter. However, you didn't even ATTEMPT to explain jack-shit. All you screech is that researchers greater than myself know better than my puny self and that AGIs have such-and-such properties, handed from on high.

Well, Starglider, there are two possibilities: either you know the technical details, or you don't. If you do, then you can at least lay out a basic argument of why I'm wrong. I want to see this argument — the argument you completely neglected to provide when you were countering my argument. If you don't know the technical details, then why are you so sure that you're right?

It's not as if AI researchers don't know that, or don't pay very close attention to human brains and organic nervous systems in general. Of course we do, but when they actually study neurology most people quickly realise that many of the structural details are byproducts of a tortuous evolutionary history with no value in a designed system.

And yet it works. That in itself has value. It gives us clues on how to construct a system that we would recognize as intelligent. Your objection that "it evolved" is but a canard — the optimization software package Evolver uses genetic algorithms (artificial evolution) to find the best design. Yet it uses the same kind of process that produced the human brain. You're committing the reverse naturalistic falacy: 'natural' implies 'bad'.

[Junghalli]

As I said, complete and total idiots. This is the level of thinking extreme religious fanatics and political zealots display. The inability to see that basic facets of your culture might not be shared by other beings.

Seriously, this relies upon them having a single culture which is pretty stupid, ignoring their own history and ignoring the possibility of alien aliens. What if you have xenos that are popped out by batches in test tubes? Ones that eat the aged? Ones that... I could go on, but they are retarded and it is too easy to shoot them full of holes.

Actually it was quite possible to talk to the Baby Eaters. They did try to convince us to adopt something like their offspring-culling practices, but then I imagine we'd try to convince them to adopt something our morality is comfortable with too.

[Junghalli]

So the subgoal... results in the preservation of the hardware which is pretty high on the 'to do' list, due to the fact that a wrecked machine hinders the top-level goal. How is this any different in practice to what I've already said? I found this subgoal by the simple observation that an AGI can't do jack-shit if it's wrecked, while you had to go through this ridiculous 'subgoal implied by top-level goals' rigamarole.

The difference is you've been saying an independent robot would have to be programmed for survival and then suggesting that this makes it potentially dangerous to its creators, because it may somehow come to resent or refuse orders to shut down or endanger itself or being treated as a simple disposable tool. Or at least that's what I'm getting out of your argument.

There is no reason this has to be the case. An entity can very consistently have survival as a secondary goal which is always obeyed unless it conflicts with one of the higher goals and yet always loses when such a conflict does happen. You imply (or seem to imply) that such an entity would be handicapped in its mission because it wouldn't try to survive in certain situations, but the only situations where that would be the case would be ones where its builders would want it to choose self-destruction. That's not a handicap, that's just its mission profile.

You know that humans can turn off the 'self-preservation' instinct by virtue of our intelligence for a higher goal, do you not? That's why WWII Japanese pilots could crash into US ships in service to their God-Emperor. We're not totally subservient to it; it just happens to be the default state.

I'd say that's evidence for the feasibility of constructing the kind of thing I'm talking about, not against. Even humans can be talked into committing to missions that end in their destruction, using nothing more than the blunt and feeble behavior-modification tools of peer pressure, psychological persuasion, rhetoric, and logic (none of which do anything more than manipulate the intelligence's sensory input; they don't even touch the actual software/hardware directly).

Of course, these techniques do usually rely on exploiting other in-built goals like kin preservation.

Edit: on a side note, while considering the problems Von Neumanns might pose for their organic creators, we should also take a moment to consider civilizations where the machines have already taken over, either because everyone's gone the posthumanism uploading route or because the nightmare evil robot scenario already happened and the evil robots won. It's rather likely that such civilizations would exist somewhere in the universe, and while the pansy organics might be too scared of Von Neumanns to consider them as a way to explore the galaxy mechanical civilizations might be less timid.

[Wyrm]

Yes, it's Wikipedia, but I see no reason to not consider it credible unless you can give me a different statement from a better source.

No, that's all right. However, you're assuming that Hale-Bopp was a former Kuiper belt object. It's eccentricity is high, 0.995086, with an aphelion of 370.8 AU. That is, it came from much further out than the approximate outer limit of the Kuiper belt, 55 AU. Furthermore, its inclination is 89.4° to the plane of the solar system, whereas KBO have inclinations on the order of 10°. Also, we still think that the Kuiper belt formed from the solar nebula, not the interstellar clouds that we think Hale-Bopp formed from, so the deuterium abundance in the Kuiper belt is going to be close to that of the solar abundance (which is close to that of Jupiter), and not that of Hale-Bopp. You're not going to get these kinds of abundancies in the Kuiper belt unless you snag a passing comet.

Even if you do manage to snag a passing comet, it's still pretty pitiful. 230 ppm is still not really a high abundance.

D-D fusion gives you 6.2 X 10^14 joules/kilogram. Assuming your self-replicating probe's industrial facility eats up 100 megawatts you could run it for 17 days on a single kilogram of deuterium, assuming 1/4 energy efficiency.

Where did you get this 100 MW figure?

You'd need more to provide energy for thermal or electric rockets, but even then unless your spacecraft is huge you're probably not going to need much. Assuming your spacecraft masses 100 tons, carries its own mass in propellant, and has an exhaust velocity like high-gear VASIMR (294 km/s) it can run through its entire propellant tank while only burning 21 kg of deuterium at 1/3 energy efficiency (plug 100 tons and 294 km/s into the kinetic energy equation and then divide the result by the energy density of deuterium fusion if you don't believe me). That gives you over 200 km/s delta V, which is plenty for going all over a solar system if you use modest energy transfer orbits.

A 100 megawatt plant needs a radiator area of 35,000 m² to dissipate waste heat (500 K). This is over three football fields. If the generator is only 1/4 efficient, then the area is quadrupled. Why does your spacecraft only weigh 100 tonnes?

With these figures in mind the fact you'd have to process 26 tons of water to get 1 kilogram of deuterium assuming deuterium abundance similar to Hale-Bopp is not so horrible.

Even if I accept the abundance, how much energy will you need to extract this deuterium?

Being able to jet all over a solar system even requires processing no more than hundreds of tons of water, you can easily find that even in a very dinky little comet, and as long as you kept a reserve of a few kilograms of deuterium you have plenty of energy for the filtration.

A VASIMR is not a torch drive. Even using modest energy transfer orbits, trips are going to take you months to years. Unless your fusion plant has a good throttle on it, or you're doing something in the interrim, you're going to be using batteries to power your electronics. If fusion is your primary source, there'll be quite a bit of charging cycles.

Of course, the bot is wearing out as you speak...

Of course, that strategy was a pretty crazy one anyway. I can think of at least three vastly better options than setting up shop on a KBO when you're dependent on solar power, starting with simply snagging one of those KBOs that have orbits that conveniently take them close to the sun at periastron (i.e. a comet).

Hale-Bopp is not a KBO — it's orbit doesn't come anywhere near the Kuiper belt. Also, if you're hitching a ride on a comet, it can be just as bad. Hale-Bopp has a period of thousands of years, most of which are spent in deep space beyond the Kuiper belt. Even short-period comets like Halley spend most of its time quite far from the sun. If you're solar powered, then you'll spend only a few months out of decades active, spending the rest of the time trying to keep warm (cold is deadly to electronics, too). The thermal cycling stresses are going to be non-trivial.

Now all you need is the ridonkulously large (and flimsy) sail.

Easily manufacturable from space resources. You just need a very big, very thin piece of aluminum foil, and aluminum is readily obtainable from any reasonably metal-rich asteroid. You can also use carbon fiber, which would use even more readily obtainable raw materials. There are theoretical advanced designs for superior very light sails using carbon nanotubes, again the raw materials would be easily obtainable from any C-type asteroid or dirty snowball. The sail must be very large but it's also very thin, so for its size it doesn't use much material.

Thurst is limited to the strain load of the thin aluminum film/nanotubes. Otherwise, the sail tears from the radiation pressure. Ultimate strength?

What about using a thorium breeder reactor? Thorium is 40 ppb on the same chart.

You may assume a fissile seed.

Why not the local star? Silicon solar cells can be made from asteroidal silicon relatively readily (ref),

What dopants are you planning to use?

and since you're in a zero-gravity environment you can easily build very large solar panel farms, or if you prefer you can use huge metal foil mirrors easily manufactured from asteroid resources to focus sunlight onto smaller solar panels arrays or simple boilers.

If you want them to be stationary, yes. Otherwise, you need some load-bearing structure, and the ultimate strength of the members limits the size/acceleration.

I see nowhere in that quote or in the referring paper the energy cost per tonne of material processed. You're operating on an energy budget, you know.

Actually I'm not at all sure that the idea that you're "operating on an energy budget" is justified, at least insofar as you seem to be implying that the probe will have to operate under a burden of serious energy scarcity. It's pretty easy to construct very large solar power plants in space out of locally available resources.

Why? How long does it take to construct those power plants, and how much energy do you need? Until you build the first solar plants, you have only a limited amount of energy/power availible to do things, because you can only carry so much stuff into the system.

I don't see what that has to do with anything. The point is you program the AI so that "I want to finish my mission" is the primary goal and "I don't want to be destroyed or damaged" is a secondary goal that it follows because being destroyed or damaged would impede or prevent it from fulfilling the mission, so there's no reason that it would resent being expected to shut down once the mission is over.

Nonetheless, it will have the impulse/instinct/whatever, correct?

It's fairly irrelevant though, as even a non-zombie can be made a completely nonthreatening and obedient servant if you get to design its mind from the ground-up.

I'll believe that when non-trivial computers are running non-trivial programs bug-free.

You think you can produce an AI without some sort of hierarchy of needs? Go do it, and once it passes the Turing test, we'll talk.

The problem is not with the idea that an AI would have a heirarchy of needs, the problem is with the idea that its self-preservation must have the same place in the heirarchy of needs that it does in ours.

I didn't mean it like that. For the most part, there is practically no difference between the failure of the top-level goal and the failure of the bot's survival. Exceptions can be programmed for by overriding the bot's survival instinct. Note that we can also override our own survival instinct for a greater good, so it's not as if having exactly the same hierarchy of needs is an insuperable problem.

A competent programmer would design our self-replicating probe so that fulfilling its mission is its first need, and self-preservation is simply an attendant need. "I want to finish my mission. If I do not survive the mission fails. Therefore I want to survive until I have finished my mission.

These few simple statements represent a large amount of detailed calculation, Junghali. The statement, "If I do not survive the mission fails," represents a decision between various actions with varying degree of utility to the ultimate goal, and some of which will result in the bot being destroyed. If the bot has to evaluate how each action resulting in the bot's destruction enhances or detriments the Purpose, it's got a lot of calculations to slog through. If it takes destruction as de facto detriment to its goals, it can start avoiding immediately — and most of the time the choice will be correct.

True, you can just load up a beefy processor, but slower processors are more robust and forgiving of material flaws.

After I have finished my mission it does not matter whether I survive or not. I will shut myself down and await further orders or the failure of my systems from entropy, whichever comes first."

Why all the rigamarole? Why not, "I've run out of things to do. I will shut myself down and await further orders." That will work just as well.

It means preserving the seat of that sapience, that is the hardware, is a primary priority, above any long-term goals.

Why? The only way that making self-preservation a secondary goal would impede self-preservation is if self-preservation conflicted with one of the higher goals. Since we designed the goal system we'd very much like the intelligence to die in those circumstances. I don't see what the problem is.

I must confess that I didn't have top-level goals like 'serve the humans' in mind when I wrote this. The long-term goals were more concrete, like 'build an infrastructure' or 'spread to other stars', both of which are better served by putting them off until the bot is out of immediate danger. I made no assumption that the bots were not doodling around just doing their thing, without knowing why they do their thing beyond simply 'it's what we do', even though their activities would ultimately serve some purpose for the humans. Humans could be kept reasonably safe by having an instruction, 'don't go to Sol (even though you send reports that way)' and 'if you see a human, shut down' and the like. The machines don't really need a top-level goal that sophisticated to do their job.

The difference is you've been saying an independent robot would have to be programmed for survival and then suggesting that this makes it potentially dangerous to its creators, because it may somehow come to resent or refuse orders to shut down or endanger itself or being treated as a simple disposable tool. Or at least that's what I'm getting out of your argument.

It's not. I'm talking about a mutant/bug-ridden robot that, through some happenstance, had the top-level goal lopped off. The survival goal cannot be overridden by the top-level goal because it's simply not there, or munged beyond all recognition. Resentment has nothing to do with it.

While this kind of error would be extremely unlikely and not a show-stopper, it is impossible to assure 100% and you're flirting with exponential growth here, especially if you only have a finite time for testing.

There is no reason this has to be the case. An entity can very consistently have survival as a secondary goal which is always obeyed unless it conflicts with one of the higher goals and yet always loses when such a conflict does happen. You imply (or seem to imply) that such an entity would be handicapped in its mission because it wouldn't try to survive in certain situations, but the only situations where that would be the case would be ones where its builders would want it to choose self-destruction. That's not a handicap, that's just its mission profile.

Evaluating whether a certain action that would destroy the machine would also advance the goals, that is, want to choose self-destruction, would be an additional burden to processing — it has to juggle the figures and figure out the expected utility of the actions. That means you need a beefier processor to handle the calculations. If these calculations aren't needed by some method, then the processor may be slower. A slower processer is more robust and easier to manufacture. Since ease of manufacturing is a basic consideration of the problem, making the process top-heavy is a detriment.

Edit: on a side note, while considering the problems Von Neumanns might pose for their organic creators, we should also take a moment to consider civilizations where the machines have already taken over, either because everyone's gone the posthumanism uploading route or because the nightmare evil robot scenario already happened and the evil robots won. It's rather likely that such civilizations would exist somewhere in the universe, and while the pansy organics might be too scared of Von Neumanns to consider them as a way to explore the galaxy mechanical civilizations might be less timid.

And hence the Fermi Paradox.

[Samuel]

Actually it was quite possible to talk to the Baby Eaters. They did try to convince us to adopt something like their offspring-culling practices, but then I imagine we'd try to convince them to adopt something our morality is comfortable with too.

They would- without it we would be able to outbreed them. I don't see why we want them to change to be like us.

Also, if that was a real situation I'd have the idiots responsible for the first contact protocol shot. Seriously, in a galaxy that is designed to stimulate conflict you do not give out information that lets them get a grasp on your capabilities. You try to inflate it as much as possible to dissuade them from attacking.

[Junghalli]

No, that's all right. However, you're assuming that Hale-Bopp was a former Kuiper belt object. It's eccentricity is high, 0.995086, with an aphelion of 370.8 AU. That is, it came from much further out than the approximate outer limit of the Kuiper belt, 55 AU. Furthermore, its inclination is 89.4° to the plane of the solar system, whereas KBO have inclinations on the order of 10°. Also, we still think that the Kuiper belt formed from the solar nebula, not the interstellar clouds that we think Hale-Bopp formed from, so the deuterium abundance in the Kuiper belt is going to be close to that of the solar abundance (which is close to that of Jupiter), and not that of Hale-Bopp. You're not going to get these kinds of abundancies in the Kuiper belt unless you snag a passing comet.

According to this paper long-period comets should if anything have less deuterium than KBOs because of the way they formed. And Halley's comet also has a deuterium abundance comparable to our oceans, at least according to this site; it's also thought to have started out as a long-period comet but at least we know Hale-Bopp is nothing particularly extraordinary. I'm not sure what the deal is with Jupiter. I'm genuinely curious about this actually so you're welcome to share any knowledge you have of this.

At any rate, an abundance of 22 instead of 154-300 ppm makes things more difficult, but I don't see why it would be a showstopper unless the technology was marginal.

Even if you do manage to snag a passing comet, it's still pretty pitiful. 230 ppm is still not really a high abundance.

But a solar system has lots and lots of hydrogen. As long as the filtration process is practical for ton quantities of hydrogen it's a readily available resource to you.

Where did you get this 100 MW figure?

The air; I simply used it as an example that the probe wouldn't need large quantities of deuterium even if it was using some fairly energy-intensive industrial processes in its replication.

A 100 megawatt plant needs a radiator area of 35,000 m² to dissipate waste heat (500 K). This is over three football fields. If the generator is only 1/4 efficient, then the area is quadrupled. Why does your spacecraft only weigh 100 tonnes?

I was ignoring thrust (and hence moment by moment energy usage) for the analysis. I was not assuming 100 megawatts of power output.

Even if I accept the abundance, how much energy will you need to extract this deuterium?

I have no idea, except that it is a fairly intensive process. I've looked on Google but I can't find any exact figures. The best thing I could find was this which says "Extracting deuterium from water is straightforward using electrolyses, and the cost of the fuel would be negligeable to the other costs of making electricity", on pg. 34. Of course, if you're using deuterium as a fusion fuel obviously it implies your civilization can obtain it in a reasonably energy-positive manner.

A VASIMR is not a torch drive. Even using modest energy transfer orbits, trips are going to take you months to years. Unless your fusion plant has a good throttle on it, or you're doing something in the interrim, you're going to be using batteries to power your electronics.

If you're exploring the galaxy I don't see why you'd be in a big hurry. As for instruments and such I imagine you'll probably have some sort of battery or something you charge off the reactor periodically and use to run the minor systems.

Also, if you're hitching a ride on a comet, it can be just as bad. Hale-Bopp has a period of thousands of years, most of which are spent in deep space beyond the Kuiper belt. Even short-period comets like Halley spend most of its time quite far from the sun. If you're solar powered, then you'll spend only a few months out of decades active, spending the rest of the time trying to keep warm (cold is deadly to electronics, too). The thermal cycling stresses are going to be non-trivial.

I can think of three probably better options. One is to try to replicate and refuel the probe during the short window when you have abundant sunlight. The other is to change the comet's orbit into one more suitable for your purposes using a mass driver, solar sail, or shoveling comet material through a nuclear thermal or electric rocket. The last is to suck as much as you can out of the comet, bail, and attach yourself to another one. Hundreds of small comets pass through the inner solar system every year (ref), so it will probably be faster to just jump from one to the next instead of waiting decades to millenia for the next periastron of the one you're attached to.

Thurst is limited to the strain load of the thin aluminum film/nanotubes. Otherwise, the sail tears from the radiation pressure. Ultimate strength?

With sufficiently advanced technology I've got one figure for being able to achieve a final speed of .012 c in a close slingshot of the sun (there's a link to the paper back in another post), but I don't know how that was derived. I suspect that was probably using the advanced carbon nanotube sail. The limits of something we could build I'm less clear on.

What dopants are you planning to use?

I don't think PERMANENT talks about dopants, what are the ones that are commonly used in industry?

Anyway, if photovoltaic cells can't be manufactured there is thermal solar. It's just a mirror that concentrates light on a container full of water or some other liquid or gas, which's expansion or boiling from the heat drives a turbine. Aside from the dynamo pretty much all you need is a water, an iron tub, and a big foil mirror.

If you want them to be stationary, yes. Otherwise, you need some load-bearing structure, and the ultimate strength of the members limits the size/acceleration.

Why take them with you? Build them, use them to power your replication and refueling, and then leave them behind. You can make more in the next system, assuming you have enough energy to get started (i.e. stick a much more modest solar panel on your probe that it takes with it).

Why? How long does it take to construct those power plants, and how much energy do you need? Until you build the first solar plants, you have only a limited amount of energy/power availible to do things, because you can only carry so much stuff into the system.

Yes, you have limited energy until you have your power plant set up. That just means you'll be limited while you're building your power plant and the basic infrastructure you need to manufacture its components. As you build it up you increase capacity which lets you increase the build rate along with the size of the power plant, and once it's in place you move on to other stuff, like actual self-replication and (non-urgent) self-repair.

Nonetheless, it will have the impulse/instinct/whatever, correct?

Presumably (Starglider may correct me). But there's no reason the survival impulse has to be the first priority.

I didn't mean it like that. For the most part, there is practically no difference between the failure of the top-level goal and the failure of the bot's survival. Exceptions can be programmed for by overriding the bot's survival instinct. Note that we can also override our own survival instinct for a greater good, so it's not as if having exactly the same hierarchy of needs is an insuperable problem.

Yes, that seems logical.

These few simple statements represent a large amount of detailed calculation, Junghali. The statement, "If I do not survive the mission fails," represents a decision between various actions with varying degree of utility to the ultimate goal, and some of which will result in the bot being destroyed. If the bot has to evaluate how each action resulting in the bot's destruction enhances or detriments the Purpose, it's got a lot of calculations to slog through. If it takes destruction as de facto detriment to its goals, it can start avoiding immediately — and most of the time the choice will be correct.

Yes, survival as the default assumption makes sense, but there will be situations in which the proper response is not survival (for instance, you're about to captured by possibly hostile aliens). The intelligence must be capable of recognizing these situations. There may be short-cuts used to eliminate from consideration places where mission and survival obviously parallel ("better move out of the way of that meteorite!"), but there's a limit to how miserly you want to be with processing capacity if you want a capable intelligence. A really effective probe of this type will probably need the effective intelligence of a smart animal if not a human.

Why all the rigamarole? Why not, "I've run out of things to do. I will shut myself down and await further orders." That will work just as well.

It was meant to illustrate the general thrust of how such a goal-system would be organized, not to provide an exact model of one.

I must confess that I didn't have top-level goals like 'serve the humans' in mind when I wrote this. The long-term goals were more concrete, like 'build an infrastructure' or 'spread to other stars', both of which are better served by putting them off until the bot is out of immediate danger.

Well, yes, I imagine you would probably have more concrete long-term goals between "complete the mission" (which may take millions of years) and "survive". The important thing is that their would be a clear heirarchy of goals, with higher goals always trumping lower ones when there is conflict between them, and selfish goals like survival and reproduction would not be at the top of the list. Assuming the goal system is well designed this should make the classic robot rebellion scenarios pretty much totally impossible.

It's not. I'm talking about a mutant/bug-ridden robot that, through some happenstance, had the top-level goal lopped off. The survival goal cannot be overridden by the top-level goal because it's simply not there, or munged beyond all recognition. Resentment has nothing to do with it.

This may be a concern (Starglider would probably be the better person to ask whether this would likely be possible with a plausible software architecture). Obviously, you'd want to install safeguards against such "mutation". Redundant programs and/or computer cores with defective ones automatically being shut down is one that springs most immediately to mind. Another is massively redundant error-proofing systems in copying. Another is multiple redundant subroutines that would be dedicated to determining whether an individual is a "mutant" and fixing or self-destructing it if it is.

Evaluating whether a certain action that would destroy the machine would also advance the goals, that is, want to choose self-destruction, would be an additional burden to processing — it has to juggle the figures and figure out the expected utility of the actions. That means you need a beefier processor to handle the calculations. If these calculations aren't needed by some method, then the processor may be slower. A slower processer is more robust and easier to manufacture. Since ease of manufacturing is a basic consideration of the problem, making the process top-heavy is a detriment.

Well, it's a balancing act. You want to make the entity's thought processes as streamlined as you can to reduce hardware demands (this is one area I can't help thinking I'm pretty sure humans are absolutely terrible - we daydream and go off on tangents constantly), but if you make them too streamlined they become rigid and you lose performance.

[Junghalli]

They would- without it we would be able to outbreed them.

I wouldn't be so sure. They're physiologically classic r strategists, producing vast numbers of offspring. If they wanted to match us they wouldn't have to stop culling, just cull less.

I don't see why we want them to change to be like us.

Their social system means millions of Baby Eater die every year (in fact by courtesy of Drama Biology their deaths are especially horrible as it takes something like days for the offspring's brain to die in the adult's digestive tract). Humans don't make decisions based entirely on game theory.

[Traveller]

I think SH is being a little pre-mature by phrasing the question in the way he does. To ask why the MW is not teeming with life implies you(we) know far more about the current state of affairs than we actually do. This is somewhat hubristic imo, though that may not have been his intent. For all we know, the MW is teeming with life, but we simply lack the means to identify this properly. The inverse could just as easily be the case. Consider that we don’t even know for certain that candidate planets in our own solar system may have harbored life of some form over its history, I think its pre-mature to presume we "know" the MW is devoid(or not) of life at this point. The only thing we can say for certain is that it appears to be lacking in life. This assumption rests on what one must admit, is a very modest effort on our part at this stage. I would suggest that in order to begin to even attempt to answer such a basic question, we would need the astronomical equivalent of the "Age of Exploration". Such an effort, for example, could include permanent manned observatories on the far side of moon, vastly more powerful and sensitive,both optical and non-opitcal, than our best current spaced based orbital and earth-based facilities. When(if) we ever get that stage, and have the result of decades or even longer with such improved techniques, then maybe we can make some sort of judgment along those lines.

A lot of the people asking this very question, unfortunately tend to make some un-warranted assumptions. As an example, some people automatically assume that intelligent life will attempt to colonize the galaxy in some form or another, or failing that, will make elaborate efforts to at least communicate there existence to anyone they can. While the first assumption is fraught with problems, the second one may be less so, even still, its entirely possible alien civilizations have come to the same conclusions some of our thinkers have. It may simply not be a very good ROI so to speak, to engage in either of these activities. What a lot of people don’t do when they ask "Why haven’t we heard from aliens", is take a very good look at humanities track record. It has only been in say, the last two centuries that humanity has even grasped the concept of possible other world life in a significant way. It was not very long ago, a large percent of this planets population(if they considered the question at all), were convinced the Earth was, flat, 4000 years old, and or the center of the universe. Actually, a lot of our fellow humans still think this way. The fact is, curiosity is not the natural or even pre-dominant mindset, even among our own kind. Think about how many Americans cant find there own country on a map, it is a leap to automatically assume that an alien species won’t or can’t be just as close-minded. Even if the MW is teeming with open-minded, curious aliens, its also just as probable that they ran the numbers and concluded that interstellar communication and or colonization, is simply not feasible and decide to more or less stay at home and leave the phone off the hook.

One needs to look at ALL the issues around a question, not just a few obvious ones.
-The many filters notion that was mentioned, is highly likely.
-Although we have gained an impressive knowledge of the Universe around us in a very short period of time, the scale and data we have would need to be vastly greater than it stands now before we make any un-warranted assumptions about the state of the life in the MW.
-We are separated both in time and space from possible contacts, this problem may be the most severe of all and the hardest to overcome.
-Alien species (could) be just as provincial and small minded as the bulk of humanity is, and have no more than a passing curiosity about life on other worlds, and even less interest in actually expending resources to prove it.
-Alien life could well be abundant, but because we lack the means to detect it, we automatically assume it is not there, this seems to be a very common position.
-Alien life may be common and well aware of us, but regard us a more than a little un-balanced. But as long as we are only a danger to ourselves with our primitive technology and general lack of motivation to harness resources beyond our own planets surface, are content to let us abuse ourselves and our planet and let us go on thinking were lords and masters of the universe.

I'd like to think there are as many possibilities as there are stars in sky.

[Samuel]

They would- without it we would be able to outbreed them.

I wouldn't be so sure. They're physiologically classic r strategists, producing vast numbers of offspring. If they wanted to match us they wouldn't have to stop culling, just cull less.

I don't see why we want them to change to be like us.

Their social system means millions of Baby Eater die every year (in fact by courtesy of Drama Biology their deaths are especially horrible as it takes something like days for the offspring's brain to die in the adult's digestive tract). Humans don't make decisions based entirely on game theory.

Ah, but culling less is considered immoral in their culture because it might imply you are one of those against culling... in which case you are eaten.

As for game theory, we are talking about a universe were aliens are one jump away from Earth. All our actions are going to be based on game theory because we are dealing with a situation where another lifeform can exterminate us and vice versa.

Traveller, your points have already been adressed in some form or another. Do you want point by point or are you going to review the thread?

[Junghalli]

Ah, but culling less is considered immoral in their culture because it might imply you are one of those against culling... in which case you are eaten.

I'm not sure simply culling less would be considered immoral by them. It is attempting to spare your own offspring that they consider immoral.

As for game theory, we are talking about a universe were aliens are one jump away from Earth. All our actions are going to be based on game theory because we are dealing with a situation where another lifeform can exterminate us and vice versa.

The Baby Eaters were technologically inferior to humans so the question was entirely what we would do with them. Well, until other technologically superior aliens showed up who found our own lifestyle just as repugnant as we found the Baby Eaters'.

[Glass Pearl Player]

You know, I can think of a damn good reason as to why we haven't had contact with sentient aliens if they exist.

Your arguments are sound, but I feel like playing devil's advocate today.

We are economically not worth it.
If someone wants our planet's resources, they'd have to take the planet from us by force.

They don't want our resources. They want some resources from a planet, which might be or might be not inhabitated. Option A is to send out probes to reachable systems, leave the inhabited ones be and exploit the unclaimed ones. Yes, this saves the expenses for using force, but TANSTAAFL means you have to wait for the results to return before you decide which systems to exploit. This means - given a non-zero interest rate - that they might gain more if they start right now and take a big stick with them (or they could throw rocks at inhabitated planets, which conveniently happen to have been placed in advance in our system).

Technology? The odds are an inter-stellar civilization has superior technology.

What if it never occurred to them to invent toothpicks, hairdryers and ice cream (like Larry Niven's Kzinti)? There might be some extremely useful gadget they - for some reason - never did invent, that scratches one of their worst itches, but is sold as a children's toy by the natives of the third planet...you get the idea. There's at least another story with that theme, "In Value Deceived" by H.B. Fyfe (never read it. maybe not online?). Note, if you want to know if there's such a thing, you have to contact aliens.

Biological samples for study?

Needs you to go there, once you're there, might as well take over (if you brought an industrial starter kit and a big stick with you).

Manufacture products unique to our civilizations? Just hack into our communication networks via remote and download them from the internet and make them yourself.

A) Plans might be trade secrets and not available on the net
B) They might be, but some incredibly crucial detail is so incredibly obvious to them, but not to you, that it has been omitted from the blueprints, lest the reader's competence and intelligence might be offended.
C) Their security measures actually do work. Any sufficient network security is indistinguishable from paranoia (Or why else does SDnet offer an encrypted link?)
D) They went cable all the way, and network access implies physical access.
E) Comparative advantage. Why build the tool chain for gadget X if you can just give the natives a copy of your industry standards and get them custom made for the equivalent of glass beads?
F) Authentic handmade gadgets from planet Y are a luxury -> they are a status symbol -> having one might get you laid -> some idiot will buy it, and the seller will therefore make profit.

[Simon_Jester]

It occurs to me that the condition for intelligent life inventing the tech base for interstellar colonization may not be "one world that can support intelligent life, where intelligent life evolves, in the star system." It may be "two habitable worlds in one system, one of which produces intelligent life."

Think about it. It's honestly not obvious whether or not interplanetary colonization pays. We've done theoretical studies of the matter ("One asteroid is worth X dollars!"), but we've never actually made a profit launching stuff into deep interplanetary space, and nobody expects to do so in the immediate future. And even a lot of the serious proposals for making money in outer space will require us to lay out sums of money equivalent to months of the gross planetary product before we get back so much as a dime.

It may be that the investment barrier to entry in interplanetary colonization and resource extraction in a star system like ours is just too high. So high that such ventures are unprofitable, even at high technology levels.

But by contrast, imagine if Venus or Mars were habitable, with an oxygen atmosphere and the resources to support some kind of agriculture. In that case, the cost/benefit calculation would be rather different. Not only would it be cheaper to maintain a base on a planet where you don't have to ship all the necessities of life in, but it would also be easier for the base population to sustain itself while doing something profitable (like mining gold or studying the native plant life). That creates a much higher incentive to send a colony to the alien planet.

And since it's pretty clear that no one is going to colonize the stars without first colonizing the planets, that might help to explain why the aliens don't come to us. A lot of them couldn't figure out a way to make interplanetary travel pay, and so never built up the base of space industries needed to make interstellar colonies possible.
_______

As I said, complete and total idiots. This is the level of thinking extreme religious fanatics and political zealots display. The inability to see that basic facets of your culture might not be shared by other beings.

In our case, the idea that aliens might not share basic facets of our culture is suggested by our own history- we tend to disagree with each other on basic principles, so we confidently expect aliens to do it even more. They might not have such a long track record of irreconciliable differences among themselves, leading them to believe that all intelligent, honorable beings will behave in much the same way. So sure, they can imagine aliens who don't eat babies... but in their eyes, such aliens must necessarily be either stupid or dishonorable.

Also, before we call this an incredible and inhuman level of stupidity, compare those aliens to the humans who seriously suggest that any starfaring alien species we meet will necessarily be peaceful, having resolved their internal conflicts so as not to wipe themselves out. That's a philosophical error on par with the baby-eating aliens assuming that all morally enlightened aliens must necessarily eat babies like they do. But Carl Sagan is not considered "stupid" for conventional definitions of the word "stupid."
_________

Nope. Tasmania lost just about all technology. A single village is NOT sustainable- a single famine, a single plague, a single tipping of the scales and things fall apart.

OK, I overstated my case. In the extreme limiting case of places like Tasmania, Easter Island, and the Greenland Vikings, even Iron Age or Neolithic technology can be lost to a small and isolated enough community. But all those societies shared one important weakness- they were all islands. Once the means of communicating with the mainland or other islands was lost, all technical knowledge was on a permanent downward ratchet.

What if you don't live on an island? What if you live in a climate where survival is difficult, but where a determined man can still load up on a few days' worth of pemmican and a bundle of trade goods and walk to the next village? In that context, I submit that it's a lot harder (and less likely) to lose Iron Age or Neolithic technology outright.

Major climate shifts might well destroy civilizations, but I doubt they could trigger the same kind of total isolation found on remote islands. People are just too good at figuring out how to talk to each other.

[Samuel]

Ah, but culling less is considered immoral in their culture because it might imply you are one of those against culling... in which case you are eaten.

I'm not sure simply culling less would be considered immoral by them. It is attempting to spare your own offspring that they consider immoral.

Yeah, but your own offspring are all you have control over. Honestly they sound like a classic case of a cultural tradition getting enshrined because no one is willing to be the first to remove it.

[qupte]

As for game theory, we are talking about a universe were aliens are one jump away from Earth. All our actions are going to be based on game theory because we are dealing with a situation where another lifeform can exterminate us and vice versa.

The Baby Eaters were technologically inferior to humans so the question was entirely what we would do with them. Well, until other technologically superior aliens showed up who found our own lifestyle just as repugnant as we found the Baby Eaters'. [/quote]

Technological inferiority doesn't make them less a threat. The transit system is based on of sun, so drive speed isn't as important- all they have to do is have a suicide drive towards Earth.

The situation is very unstable as the author described it. I'm not sure if that was the intention, but he made a universe were instituitonal paranoia is a good trait. After all, it is extremely easy to get to the enemy home system and the home system is the key to your entire civilization.

What if it never occurred to them to invent toothpicks, hairdryers and ice cream (like Larry Niven's Kzinti)? There might be some extremely useful gadget they - for some reason - never did invent, that scratches one of their worst itches, but is sold as a children's toy by the natives of the third planet...you get the idea. There's at least another story with that theme, "In Value Deceived" by H.B. Fyfe (never read it. maybe not online?). Note, if you want to know if there's such a thing, you have to contact aliens.

That is extremely unlikely. A hypothetical interstellar civilization would have colonized most of its solar system and have a population possibly in the trillions.

They might not have such a long track record of irreconciliable differences among themselves, leading them to believe that all intelligent, honorable beings will behave in much the same way. So sure, they can imagine aliens who don't eat babies... but in their eyes, such aliens must necessarily be either stupid or dishonorable.

So they are still stupid because we could be dishonerable. They assume we aren't because we didn't immediately fire upon them... but we could be in another war. Or waiting to see their reaction. Or...
Basically the aliens failed to consider ANY possiblities at all.

Also, before we call this an incredible and inhuman level of stupidity, compare those aliens to the humans who seriously suggest that any starfaring alien species we meet will necessarily be peaceful, having resolved their internal conflicts so as not to wipe themselves out. That's a philosophical error on par with the baby-eating aliens assuming that all morally enlightened aliens must necessarily eat babies like they do. But Carl Sagan is not considered "stupid" for conventional definitions of the word "stupid."

That isn't a stupid position. It just relies upon the belief that all xenos go through a period where they can wipe themselves out and that they need to become relativly peaceful or else they will extreminate themselves. I would point out that Stuart holds a belief similar to that given his bioterror fears.

In a scenario with cheap FTL this goes out the window- they can get into space before they go through the winnowing process. Not to mention this ignores the possibility that they aren't peaceful towards outsiders.

they were all islands.

Well, all landmasses are islands when you think about it. Your example was a village that had just about the same characteristics. If it is not isolated in a dipshit little valley than it would probably not regress... although I'd like to point out that was the situation of the Mayans and they collapsed.

[Simon_Jester]

Technological inferiority doesn't make them less a threat. The transit system is based on of sun, so drive speed isn't as important- all they have to do is have a suicide drive towards Earth.

In point of fact, no. There is at least one, though possibly only one, star system between the first contact system and Earth. So a suicide drive isn't going to cut it; they don't even know where the wormhole to Earth is.

What if it never occurred to them to invent toothpicks, hairdryers and ice cream (like Larry Niven's Kzinti)? There might be some extremely useful gadget they - for some reason - never did invent, that scratches one of their worst itches, but is sold as a children's toy by the natives of the third planet...
That is extremely unlikely. A hypothetical interstellar civilization would have colonized most of its solar system and have a population possibly in the trillions.

Therefore, anything that can be invented has been invented, and this is a sure enough bet that the alternative isn't worth looking into?

I must have misunderstood your argument.
______

So they are still stupid because we could be dishonerable. They assume we aren't because we didn't immediately fire upon them... but we could be in another war. Or waiting to see their reaction. Or...
Basically the aliens failed to consider ANY possiblities at all.

I think you're using "stupid" to cover too broad a range of concepts.

More specifically, these aliens think intelligent beings will act predictably and rationally, and that they have enough data to define rational behavior. I wouldn't call that "stupid," because it's a failure of imagination and not of intelligence. For all we know, in their experience all rational beings do act approximately the same way, making their behavior predictable. This might happen because they tend to eat anyone who behaves "irrationally," and have done so for millenia; by process of natural selection they've wound up with a bunch of beings who all think about the same way.

And they might very well assume that any civilized species must do the same thing, or you could have all kinds of crazy people running around fighting wars with each other and smashing things. How could such a species ever get into space? How could you have a species of rational beings capable of maintaining a permanent disagreement with each other about first principles?

Which means that they can apply their version of game theory to the situation, and that they're likely to trust their results. They conclude that a more powerful ship that does not attack them must be trustworthy, and therefore ethical, and therefore (because of their idea of where ethics comes from) baby-eating. Or at least capable of understanding the necessity of baby-eating on their part, and therefore its rightness, which is to them self-evident.

Indeed, they're not entirely wrong, because we can understand why they do it, and the fact that we can't think of a better way to arrange things is one of the major reasons why the humans on the first contact ship don't launch an immediate attack.

The Babyeaters' failure is, as I say, a failure of imagination and not of intelligence. The aliens lack the ability to imagine that their fundamental assumptions about rationality and ethical behavior are wrong. So clearly they're unimaginative, at least about first principles, but does their lack of imagination mean a lack of general intelligence? I've met too many stupid people with vivid imaginations to be confident of a strong relationship between those traits.
_______

That isn't a stupid position. It just relies upon the belief that all xenos go through a period where they can wipe themselves out and that they need to become relativly peaceful or else they will extreminate themselves. I would point out that Stuart holds a belief similar to that given his bioterror fears.

In a scenario with cheap FTL this goes out the window- they can get into space before they go through the winnowing process. Not to mention this ignores the possibility that they aren't peaceful towards outsiders.

But Sagan argued that they would be peaceful towards outsiders, simply because they must be peaceful towards themselves. This is obviously wrong, but it would be incorrect to assume that Carl Sagan was a stupid man. Sagan committed a failure of imagination, assuming that he could predict how thinking beings would behave based on what he thought was rational behavior.

The only way to be safe in dealing with truly foreign minds is to make no assumptions about them except what they themselves tell you- and to accept even that only on a provisional basis. But intelligent people with too much faith in the power of what they call "reason" (including game theory) can fail to do that.
_______

Well, all landmasses are islands when you think about it. Your example was a village that had just about the same characteristics. If it is not isolated in a dipshit little valley than it would probably not regress... although I'd like to point out that was the situation of the Mayans and they collapsed.

I wasn't thinking in terms of a valley that would be completely and utterly inaccessible from outside; I should have made that clear. Sorry.

I was thinking in terms of the resource base a pre-industrial civilization would need. And that's pretty small, at least on a per-community basis; they may well be able to find everything they truly need within easy walking distance. If they're completely isolated from all other communities, they're automatically fucked the moment the blacksmith keeled over without training his apprentice, I agree. But for cultures living on a continental landmass, that's not very likely to happen. You just don't need all that much technology to go find someone who knows how to keep your technology working, or to keep it working once you've got the knowledge. There's nothing equivalent to railroads to cut, no specialized tool and die factories that can fail and leave you with no production capability. Just about anything that can be made, can be made using knowledge and skills that exist entirely in one person's hands and brain.

And that was my key point- an Iron Age society is far less sensitive to disruptions of infrastructure than a post-industrial one that has already dug up most of the world's available resources. It would be much easier to knock someone from the Atomic Age to the Iron Age than from the Iron Age to the Stone Age. Organization can be lost- you can go from an Iron Age megalopolis like classical Rome to a handful of Iron Age hicks squatting in the ruins of said megalopolis. But they're still Iron Agers, and all it takes is a few generations of good leadership and relative quiet to rebuild the place.
_______

With respect to the Mayans, it's worth noticing that they didn't lose technology, only organization- and their organization was bouncing back within a few centuries. The Mayan city-states didn't collapse all at the same time, either; some were ruined by crises that left others untouched. You can say "Mayan civilization collapsed," but it's like saying that "European civilization collapsed" during the Dark Ages. The people were still there, they still knew how to do a lot of relevant things, and they didn't have to reinvent the wheel from scratch or anything. Once the specific problem that destroyed the old civilization ended, a new one arose in relatively short order.

[Samuel]

In point of fact, no. There is at least one, though possibly only one, star system between the first contact system and Earth. So a suicide drive isn't going to cut it; they don't even know where the wormhole to Earth is.

From the colony system Huygens, the crew of the Giant Science Vessel Impossible Possible World have set out to investigate a starline that flared up with an unprecedented flux of Alderson force before subsiding. Arriving, the Impossible discovers the sparkling debris of a recent nova - and -

Most colony worlds are located only a single jump away from Earth, which remains the center of the human universe.

The odds are good that if the aliens were there first, they can tell which portal the human ship came from. If they were hostile and decided to destroy it, they could follow it back. And then they could follow any ship that reaches the colony back to Earth.

There are 1 billion nodelanes and 400 billion stars in the galaxy so transit lines are going to be really short between different alien species.

Therefore, anything that can be invented has been invented, and this is a sure enough bet that the alternative isn't worth looking into?

I must have misunderstood your argument.

If you have a civilization of a trillion than, yes, I believe all consumer product ideas have been invented. To put that into context, the human race has only had 60 billion members in its history. So unless the aliens were 20 times stupider than us they should be able to do it (this is considering a hard sci-fi universe; softer ones will not need such built up systems due to FTL allowing meeting before this happens.)

More specifically, these aliens think intelligent beings will act predictably and rationally, and that they have enough data to define rational behavior.

They have no reason to justify that assumption. Most importantly, if the opponent is rational and predictable, they will alter their behavior in order to be unpredictable.

I wouldn't call that "stupid," because it's a failure of imagination and not of intelligence.

A big enough failure of imagination IS a failure of intelligence.

This might happen because they tend to eat anyone who behaves "irrationally," and have done so for millenia; by process of natural selection they've wound up with a bunch of beings who all think about the same way.

Human society actually does this with homicidal sociopaths- people who randomly kill others tend to be put down like dogs. However, we still have them to this very day.

And they might very well assume that any civilized species must do the same thing, or you could have all kinds of crazy people running around fighting wars with each other and smashing things. How could such a species ever get into space? How could you have a species of rational beings capable of maintaining a permanent disagreement with each other about first principles?

Unless they are all clones of each other, they will have a varience of opinions about other things- what they consider nonessential. All you need to realize is that a species might extend the idea of nonessential to things like first principles. Given that evolved life forms are very bad at justifying their actions this makes perfect logical sense.

Which means that they can apply their version of game theory to the situation, and that they're likely to trust their results. They conclude that a more powerful ship that does not attack them must be trustworthy, and therefore ethical, and therefore (because of their idea of where ethics comes from) baby-eating. Or at least capable of understanding the necessity of baby-eating on their part, and therefore its rightness, which is to them self-evident.

Unless the ship got there second and wishes to see what portal you use. Or the aliens are bluffing. Or they want to exterminate your species in one swift decisive blow. Or...

Indeed, they're not entirely wrong, because we can understand why they do it, and the fact that we can't think of a better way to arrange things is one of the major reasons why the humans on the first contact ship don't launch an immediate attack.

The humans in the story are also idiots.

So clearly they're unimaginative, at least about first principles, but does their lack of imagination mean a lack of general intelligence? I've met too many stupid people with vivid imaginations to be confident of a strong relationship between those traits.

You do realize this is something that is supposed to be decided back on the capital world? You get the anthropologists, sociologists, military, astronomers, etc all together and you come up with a strategy. What this implies is NONE of them thought of it.

So not only does every single member of their species share the same assumptions, but none of them are capable of questioning it. Hell, even if we take their assumption, they still screwed up- there might be species that kill some of the young and feed it to the rest in order for them to survive and baby eating would be seen as an even worse act than we see it- after all, by eating the young you are denying food to the rest of the children!

But Sagan argued that they would be peaceful towards outsiders, simply because they must be peaceful towards themselves. This is obviously wrong, but it would be incorrect to assume that Carl Sagan was a stupid man. Sagan committed a failure of imagination, assuming that he could predict how thinking beings would behave based on what he thought was rational behavior.

It is rational though- the cost of interstellar flight means war is too expensive and assuming the aliens work of game theory (being the way they survived) they won't attack.

But intelligent people with too much faith in the power of what they call "reason" (including game theory) can fail to do that.

It works perfectly. The problem is if you misunderstand the enemies goal. Assuming an evolved life form we can reasonably predict what their goals would be. This only goes out the window in space opera because you have aliens that could not possibly evolve.

I was thinking in terms of the resource base a pre-industrial civilization would need. And that's pretty small, at least on a per-community basis; they may well be able to find everything they truly need within easy walking distance.

If that was true, why were there massive trade networks?

And that was my key point- an Iron Age society is far less sensitive to disruptions of infrastructure than a post-industrial one that has already dug up most of the world's available resources.

False. Unlike them our civilization has no danger of collapse. We could lose whole continents without it affecting our technological base. Additionally we can combat the problems and identify them in advance while they couldn't.

Having everyone live is an entirely different problem but if we were sufficiently pressed, we could survive just about any natural change.

It would be much easier to knock someone from the Atomic Age to the Iron Age than from the Iron Age to the Stone Age.

The first has not happened. The latter may have happened before.

Organization can be lost- you can go from an Iron Age megalopolis like classical Rome to a handful of Iron Age hicks squatting in the ruins of said megalopolis. But they're still Iron Agers, and all it takes is a few generations of good leadership and relative quiet to rebuild the place.

Going off wiki (if you want we can get Thanas), but Rome took 1800 years to crawl back to its population at its height.

With respect to the Mayans, it's worth noticing that they didn't lose technology, only organization- and their organization was bouncing back within a few centuries. The Mayan city-states didn't collapse all at the same time, either; some were ruined by crises that left others untouched. You can say "Mayan civilization collapsed," but it's like saying that "European civilization collapsed" during the Dark Ages. The people were still there, they still knew how to do a lot of relevant things, and they didn't have to reinvent the wheel from scratch or anything. Once the specific problem that destroyed the old civilization ended, a new one arose in relatively short order.

European civlization DID collapse. No one had the ability to create comparable structures, militaries, etc for a thousand year in Western Europe after the fall.

[Simon_Jester]

The odds are good that if the aliens were there first, they can tell which portal the human ship came from. If they were hostile and decided to destroy it, they could follow it back. And then they could follow any ship that reaches the colony back to Earth.

Aside from the point that they lack the technical capability to destroy any armed human ship, the key point here is that a suicide run aimed at passing through two wormholes and a large volume of intervening space leaves you open to interception.

There are 1 billion nodelanes and 400 billion stars in the galaxy so transit lines are going to be really short between different alien species.

Who says all the stars linked by the node network are in the same galaxy? I don't think the author did, and his statements in the attached comment threads imply that they aren't. I remember it being noted that no star linked by a wormhole has ever been found in 'telescope range' of another. I'm not sure what 'telescope range' is defined as, but that seems very unlikely if all the stars are within the Milky Way.
________

If you have a civilization of a trillion than, yes, I believe all consumer product ideas have been invented. To put that into context, the human race has only had 60 billion members in its history. So unless the aliens were 20 times stupider than us they should be able to do it (this is considering a hard sci-fi universe; softer ones will not need such built up systems due to FTL allowing meeting before this happens.)

I have a sneaking suspicion that some things are far more likely to be invented by some kinds of minds than others. The fact that my population equals that of the alien civilization does not guarantee that I am equally likely to invent all the same useful tools, even when I really need them.

I can't prove that, but I don't think it can be ruled out to the point where the alternative isn't worth considering.
________

Human society actually does this with homicidal sociopaths- people who randomly kill others tend to be put down like dogs. However, we still have them to this very day.

The key concept here is that for our imagined species of rationalist killers of irrationalists, a thinker might reasonably assume that the average of any thinking species will trend in the same direction. They might have individual freaks who for some reason haven't been offed yet, but they won't be collectively insane. Or if they were, they'd be too crazy to build the artifacts of civilized society, or at least they'd behave in ways that indicate their collective irrationality.

So the very fact that you're meeting them and that they didn't attack you on sight "proves" that they must be in the habit of killing off (or restraining) their irrationalists and that their "normal" behavior is civilized. In which case you can trust them with information saying "Look, we're civilized too!"

Remember that there are two sides to this decision- telling them about yourself might offend them when they would otherwise be friendly, but not telling them about yourself might make them more suspicious and hostile. A show of good faith will influence the odds in your favor in extended versions of the Prisoner's Dilemma.
_______

Again, this analysis is actually correct in the context of the story; humanity is a much more rational species in this setting than they are today, thanks in large part to the effects of intelligence-enhancing drugs. In-story, we are in the habit of restraining irrational individuals, and our normal behavior is civilized (as we or the aliens define "civilized"). While it's still possible that we will make what we'd call a rational decision to attack the aliens and destroy their civilization, it would be a rational decision, one that we actually sat down and thought about, not an automatic "eat hot plasma, baby-eating xenos scum!" response.

Which is more or less what they're counting on. As best as I can determine, the aliens see an unknown alien with a track record of being rational/civilized enough to follow the "cooperate" path in this variant of the Prisoners' Dilemma. They figure that the alien is less likely to attack them when given more information about them, because they expect the alien to be reassured by confirmation that they are civilized and rational.

Again, this is not unprecedented reasoning among humans- many of our own feeble attempts to send messages to aliens include unnecessary information beyond the minimum information "there is a sentient lifeform over here." We could get that by beaming pi out in binary, after all.

But instead we say things like "we come in peace for all mankind." We expect the aliens to think "Ah, they say that they come in peace, this makes it at least slightly more likely that they are not hostile." For all we know, they might instead think "Aaack, these individuals say they are coming for their entire species? They must be a sinister hive mind! Kill, kill, KILL!" We can't rule that out, but we say "we come in peace for all mankind" anyway.
______

The humans in the story are also idiots.

Wait... you encounter an alien ship and the non-idiotic response is to open fire immediately? Or is that not why they're idiots? I'm getting very confused here, because there is an enormous range of things I can think of that (I expect) you'd call idiotic.
______

It is rational though- the cost of interstellar flight means war is too expensive and assuming the aliens work of game theory (being the way they survived) they won't attack.

That's still a failure of imagination. What if they haven't thought of this lovely notion and have convinced themselves that all aliens are likely to defect in the Prisoners' Dilemma? Then it would behoove them to defect as well and start lobbing relativistic kill vehicles your way as soon as possible, to keep you from becoming dangerous. Or what if they decide that they can't assume that you will assume that they are peaceable? In that case, you might decide to kill them first, in which case they'd better nail you before you get the chance!

This can be repeated indefinitely: at any break in the chain of "I can assume that you can assume that...I can assume that you are peaceable," the logical response is immediate preemptive attack. This is a Prisoners' Dilemma variant where the odds are heavily stacked in favor of defection.

Just like our hypothetical aliens, Sagan assumed that all truly intelligent, civilized species would make the same decision that he would make. While the image of a universe populated by billions and billions of Carl Sagan clones is amusing, it is clearly unrealistic. While this is not a failure of rational thought, and not a sign that Sagan was stupid, it was a major breakdown of imagination.

==========

If that was true, why were there massive trade networks?

Luxury goods, as often as not. You don't truly need steel swords that can't be manufactured from crappy local iron ore, or spices, or even wool clothing... but those things are valuable enough to be worth trading for what is available locally. Collapse the infrastructure and those luxuries are lost to most, if not all, of what's left of the world. But the raw essentials are still there; very little has to be reinvented, and very little has to be rebuilt from the ground up the way that industrial infrastructure would.

And that was my key point- an Iron Age society is far less sensitive to disruptions of infrastructure than a post-industrial one that has already dug up most of the world's available resources.

False. Unlike them our civilization has no danger of collapse. We could lose whole continents without it affecting our technological base. Additionally we can combat the problems and identify them in advance while they couldn't.

If Iron Agers lost a whole continent, people on the other continents wouldn't even notice. Localized disruptions, even big ones, don't destroy a global intelligent species. Only global disruption could possibly do it, so this is a red herring.

Once you establish that the problem is global, we start losing technical capability a lot faster than the primitives will, because we have more to lose and more of what we do have is dependent on the tools in question. As long as the problem is limited we can contain the damage- if all microchip factories vanished and nothing else happened, we'd be OK. But once the set of breakdowns passes a crucial point we get a chain reaction.

Destroy all of Tool A that is used to make Tool B, and we may not be able to rebuild them, because we threw away the tools of type A-prime that we used to make the A's in the first place. A's work better than A-primes, after all...

We'd wind up having to settle for a lower level of technology, and not just reduced levels of luxuries within the same technological framework: no computers, not just fewer computers, at least until someone can figure out how to rebuild the semiconductor fabricators starting from a gaslight era tech base.

That can't happen in a society where the tools of technology are less interdependent. You still get massive suffering and death, but unless the entire planet becomes completely uninhabitable at the drop of a hat, life goes on somewhere. And if we were talking about the kind of world where common, random changes in the environment could render the whole world uninhabitable to a species that was presumably well adapted before... the entire idea of intelligence evolving there at all is a chimera.
_________

Organization can be lost- you can go from an Iron Age megalopolis like classical Rome to a handful of Iron Age hicks squatting in the ruins of said megalopolis. But they're still Iron Agers, and all it takes is a few generations of good leadership and relative quiet to rebuild the place.

Going off wiki (if you want we can get Thanas), but Rome took 1800 years to crawl back to its population at its height.

But Rome was very definitely a city, not just a handful of hicks, within 500 years after its fall. Its population was smaller largely because it was no longer the center of a great empire. It had a smaller hinterland to draw on, because other centers of power had arisen while it was weak.

The difference between Rome c. 400 AD and Rome c. 1000 AD was one of organization, not technology. Since Rome never became the heart of a new organization with the relative power of the Empire, it's no surprise that it took fundamentally higher technology to get it up to the same population it had before.
_________

European civlization DID collapse. No one had the ability to create comparable structures, militaries, etc for a thousand year in Western Europe after the fall.

Which, again, is a matter of organization. It's not that they didn't know how to pile rocks, it's that they couldn't get enough people working on it a given pile. The specific knowledge pertaining to details of architecture was lost, but that is a highly specialized field.

Things changed, but the raw materials was still there.

You could have a series of such collapses, one after another, for many millenia without anything fundamental changing. A series of pre-industrial civilizations rising and falling is a stable equilibrium. Losing the previous civilization does not make it more difficult to build the next one. As long as the planet stays inhabitable, someone will be in some stage of building and maintaining an advanced civilization (by local standards) somewhere.

[Samuel]

Aside from the point that they lack the technical capability to destroy any armed human ship, the key point here is that a suicide run aimed at passing through two wormholes and a large volume of intervening space leaves you open to interception.

Except the portals are all directly around stars. Which means it isn't a "large volume of open space".

I have a sneaking suspicion that some things are far more likely to be invented by some kinds of minds than others. The fact that my population equals that of the alien civilization does not guarantee that I am equally likely to invent all the same useful tools, even when I really need them.

I can't prove that, but I don't think it can be ruled out to the point where the alternative isn't worth considering.

I doubt aliens are significantly less creative than humanity.

The key concept here is that for our imagined species of rationalist killers of irrationalists, a thinker might reasonably assume that the average of any thinking species will trend in the same direction. They might have individual freaks who for some reason haven't been offed yet, but they won't be collectively insane. Or if they were, they'd be too crazy to build the artifacts of civilized society, or at least they'd behave in ways that indicate their collective irrationality.

So these aliens have never had any religions or fanatical political ideologies?

So the very fact that you're meeting them and that they didn't attack you on sight "proves" that they must be in the habit of killing off (or restraining) their irrationalists and that their "normal" behavior is civilized. In which case you can trust them with information saying "Look, we're civilized too!"

Sociopaths do not attack other people on sight. They would kill you without hesitation if it was in their advantage, but there is no reason to believe that an enemy that does not immediately try to kill you isn't a potential hostile. They are probably seizing you up.

Then you die.

but not telling them about yourself might make them more suspicious and hostile. A show of good faith will influence the odds in your favor in extended versions of the Prisoner's Dilemma.

Of course- you tell as much about yoursel without giving any information about you military potential. If you don't tell them anything you aren't communicating, and if you tell them everything they get information they can use against you. Remember- you have no way to check the validity of the info immediately. There is no reason to believe it is true if they could us the good faith to come back with you.

The only true piece of information that should be allows shared is succesul peace with other alien species. That, and that ALONE, is priceless.

Again, this analysis is actually correct in the context of the story; humanity is a much more rational species in this setting than they are today, thanks in large part to the effects of intelligence-enhancing drugs.

IQ=/ rationality.

and our normal behavior is civilized (as we or the aliens define "civilized").

It is only civilized because that is how we define normal behavior.

it would be a rational decision, one that we actually sat down and thought about, not an automatic "eat hot plasma, baby-eating xenos scum!" response.

That is a rational decision. If you consider individuals that eat babies worthy of extermination than it is perfectly rational to exterminate them.

As best as I can determine, the aliens see an unknown alien with a track record of being rational/civilized enough to follow the "cooperate" path in this variant of the Prisoners' Dilemma.

The best solution is to betray your opponent.

They figure that the alien is less likely to attack them when given more information about them, because they expect the alien to be reassured by confirmation that they are civilized and rational.

So they are too stupid to consider their own history? Do they realize that people in the past considered themselves civilized and rational while at the same time doing things they consider would violate that?

Again, this is not unprecedented reasoning among humans- many of our own feeble attempts to send messages to aliens include unnecessary information beyond the minimum information "there is a sentient lifeform over here." We could get that by beaming pi out in binary, after all.

Our universe lacks FTL.

But instead we say things like "we come in peace for all mankind." We expect the aliens to think "Ah, they say that they come in peace, this makes it at least slightly more likely that they are not hostile." For all we know, they might instead think "Aaack, these individuals say they are coming for their entire species? They must be a sinister hive mind! Kill, kill, KILL!" We can't rule that out, but we say "we come in peace for all mankind" anyway.

That is retarded. That would require a species that has never heard of delegation.

Wait... you encounter an alien ship and the non-idiotic response is to open fire immediately? Or is that not why they're idiots? I'm getting very confused here, because there is an enormous range of things I can think of that (I expect) you'd call idiotic.

They hand over enough information the aliens can tell we aren't baby eaters. Quick hint- don't show your young- you can deduce how quick they reproduce and hence how quick they are expanding. You don't want them to know this information.

Then it would behoove them to defect as well and start lobbing relativistic kill vehicles your way as soon as possible, to keep you from becoming dangerous. Or what if they decide that they can't assume that you will assume that they are peaceable? In that case, you might decide to kill them first, in which case they'd better nail you before you get the chance!

Once you have a dyson swarm, all this attack does is make your enemies angry... and make your civilization VERY visable. At which point your neighbors cook your system.

Luxury goods, as often as not.

The "as not" is important. Egypt for example had to import wood. There are going to be breaks that require trade goods to exist in or else you won't have a continuous swatch of civilization, which means it falls apart more easily.

Once you establish that the problem is global, we start losing technical capability a lot faster than the primitives will, because we have more to lose and more of what we do have is dependent on the tools in question.

Nothing naturally occuring and global could obliterate human civilization. Maybe a big enough asteroid, but in a couple decades even that won't be a problem. We are more fragile to these problems, but the shell is tough enough they will never get through.

And if we were talking about the kind of world where common, random changes in the environment could render the whole world uninhabitable to a species that was presumably well adapted before... the entire idea of intelligence evolving there at all is a chimera.

Depending on the definition of "common" Earth could fit

But Rome was very definitely a city, not just a handful of hicks, within 500 years after its fall. Its population was smaller largely because it was no longer the center of a great empire. It had a smaller hinterland to draw on, because other centers of power had arisen while it was weak.

The difference between Rome c. 400 AD and Rome c. 1000 AD was one of organization, not technology. Since Rome never became the heart of a new organization with the relative power of the Empire, it's no surprise that it took fundamentally higher technology to get it up to the same population it had before.

I would not call 20,000 a city. It took 1000 years for them to return to the same level of population they had after the total collapse of the empire.

Year Population
350 BC 30,000
250 BC 150,000
44 BC 1,000,000
120 1,000,000
330 800,000
410 700-800,000
530 90-150,000
650 70,000
1000 20,000
1400 20,000
1526 50,000–60,000
1528 20,000
Year Population
1600 100,000
1750 156,000
1800 163,000
1820 139,900
1850 175,000
1853 175,800
1858 182,600
1861 194,500
1871 212,432
1881 273,952
1901 422,411
1911 518,917
Year Population
1921 660,235
1931 930,926
1936 1,150,589
1951 1,651,754
1961 2,188,160
1971 2,781,993
1981 2,840,259
1991 2,775,250
2001 2,663,182
2007 2,718,768

You could have a series of such collapses, one after another, for many millenia without anything fundamental changing. A series of pre-industrial civilizations rising and falling is a stable equilibrium. Losing the previous civilization does not make it more difficult to build the next one. As long as the planet stays inhabitable, someone will be in some stage of building and maintaining an advanced civilization (by local standards) somewhere.

You can cut all the trees down and make a desert, leech all the nutrients out of the soil with poor crop growing methods, etc.

Edit- if anyone has any objection to the wikipedia usage please tell me and give a source that has a better source or Rome's population over time- or just say why it is horribly inaccurate.

[Wyrm]

According to this paper long-period comets should if anything have less deuterium than KBOs because of the way they formed.

My reading of this paper suggests that the the D/H ratios are to be used as a test for the orgin of both short- and long-period comets. If long-period comets like Hale-Bopp have lower D/H ratios than short-period comets, then Oort's notion —that long-period comets, paradoxically, are formed closer to the sun than the short-period comets— is correct, which results in lower D/H ratios. The overabundance of deuterium in Hale-Bopp tends to contradict this notion.

And Halley's comet also has a deuterium abundance comparable to our oceans, at least according to this site; it's also thought to have started out as a long-period comet but at least we know Hale-Bopp is nothing particularly extraordinary. I'm not sure what the deal is with Jupiter. I'm genuinely curious about this actually so you're welcome to share any knowledge you have of this.

Hale-Bopp is nothing particulary extraordinary for a long-period comet, but that's a long way from saying that that it's a former KBO. However, the Kuiper belt does have a higher deuterium abundancy than Jupiter, but only about 5x that of Jupiter, giving an abundance very close to that of the world's oceans.

At any rate, an abundance of 22 instead of 154-300 ppm makes things more difficult, but I don't see why it would be a showstopper unless the technology was marginal.

The technology is going to be as close to marginal as the problem allows. Ease of manufacture is going to be a high priority in a practical VNM. Also, the smaller the probe, the better. The probe cannot take advantage of the economy of scale availible to the home civilization, at least initially.

But a solar system has lots and lots of hydrogen. As long as the filtration process is practical for ton quantities of hydrogen it's a readily available resource to you.

I never said that the quantity of fuel isn't a problem. It's always been about the fractionation process. For fusion fuel, what you want is doubly-heavy water —ie, deuterium oxide— since singly heavy water has 50% protium impurity. This requires further refinement.

H-H fusion would solve the issue, but we're nowhere close to that stage.

Where did you get this 100 MW figure?

The air; I simply used it as an example that the probe wouldn't need large quantities of deuterium even if it was using some fairly energy-intensive industrial processes in its replication.

So you're taking 100 MW as a basic unit of power.

Even if I accept the abundance, how much energy will you need to extract this deuterium?

I have no idea, except that it is a fairly intensive process. I've looked on Google but I can't find any exact figures. <snip>

The Vamork plant produced perhaps three hundred barrels of heavy water a year with purity of around 10% - 20% using a 60 MW hydroelectric plant. That works out to about 6.3 TJ per barrel, or 42 GJ per liter/kg. This is well within the specific energy of D-D fusion. However, this is only 10% - 20% enrichment, and the major constituent is probably HDO, unless there is some chemonuclear bias I don't know about. You probably need pretty pure D2O as your feedstock to your power plant.

Of course, if you're using deuterium as a fusion fuel obviously it implies your civilization can obtain it in a reasonably energy-positive manner.

That doesn't mean the technology scales down to probe-sized with comparable efficiency.

A VASIMR is not a torch drive. Even using modest energy transfer orbits, trips are going to take you months to years. Unless your fusion plant has a good throttle on it, or you're doing something in the interrim, you're going to be using batteries to power your electronics.

If you're exploring the galaxy I don't see why you'd be in a big hurry. As for instruments and such I imagine you'll probably have some sort of battery or something you charge off the reactor periodically and use to run the minor systems.

That battery has a finite number of charges in it before it fails. Charging induces physical stresses on the battery elements because the charge carriers are being traded off between the elements, and those charge carriers take up room. Furthermore, the battery has to be kept within a critical temperature range or it stops working, and you're probe is dead. That means heating/refrigerator elements to keep within this range, which also draw power. You need to make a new battery before your current one fails, so you are in a bit of a hurry.

Also, if that battery doesn't have enough juice to start the fusion plant, then you have to keep the plant ticking over on its own. If the plant stalls below, say, 100 MW (raw production), then by your stated 6.2e14 J/kg, you need one kg every 71 days, or 5 kg a year just to keep the engine ticking over. 5 kg of deuterium is about 1.25 kilomoles, and requires 1.25 kilomoles (25 kg) of doubly-heavy water. At an abundance of 1/340,000 (less for Hale-Bopp type comets, more for Jupiter abundancies), this gives us 8517 tonnes of water processed per year just to keep the plant ticking over. That's a 125 m wide comet nucleus. This is about the size of the dinkiest of the dinky comets, so the entire comet would have to be deconstructed just for a year's fuel.

The one part in 6700 figure stated is probably for the abundance of finding at least a single deuterium in the molecule. The 340,000 figure probably for the abundance of heavy water proper, that is D2O, so the ratio of singly heavy water to doubly heavy water is 50::1, so the Vamork enrichment of doubly heavy water is around 2.96e-3. If we assume that enrichment follows an S-curve with respect to energy, with normal enrichment coincident with zero energy, we can extrapolate to 99.9% enrichment. If the energy/enrichment function is f = [tan⁻¹(a e + b) + 1]/2, with a,b to be determined, with boundary conditions of e = 0, f = 1/340,000 and e = 42 GJ/kg, f = 2.96e-3, this gives us enirchment constants of a = 4.7729032658333e-13 kg/TJ, b = -1.5573875747288. This gives us a specific enrichment energy of about 6.5 TJ/kg. This is about 1% of the energy liberated by D-D fusion — encouraging, as you can get more energy out than put in, at least for this stage, provided you can process the material fast enough.

Of course this is for a Vamork-type facility, which has the advantage of scale, but the disadvantage of being WWII techonology. This plant also produced only fission-reactor grade heavy water, not fuel-grade water. A fuel plant would require more stages and more equipment, and this is for a ~300 barrel throughput. No doubt a probe-sized plant would have reduced throughput and lower efficiency.

I can think of three probably better options. One is to try to replicate and refuel the probe during the short window when you have abundant sunlight.

A comet spends at best a year in the sunlight, and if your comet has a period of thousands of years, that's going to slow your growth rate by a factor of thousands. (And take care in your comet choice: some have periods of millions of years.)

The other is to change the comet's orbit into one more suitable for your purposes using a mass driver, solar sail, or shoveling comet material through a nuclear thermal or electric rocket.

Unless you want your orbital adjustment to take many orbits, which essentially brings you right back where you started, you'd need a torch drive; you need to make serious changes to the orbit of a massive object in a short time.

The last is to suck as much as you can out of the comet, bail, and attach yourself to another one. Hundreds of small comets pass through the inner solar system every year (ref), so it will probably be faster to just jump from one to the next instead of waiting decades to millenia for the next periastron of the one you're attached to.

If you're jumping onto and off of comets during these periods of the comets' orbits, then you're going to have expensive transfers. You'll have to spend some time on the comet to do your business, so you have to change your orbit to catch the comet as it comes it (which requires you to detect them far out, which harkens back to the discussion of detection of objects quickly enough to hitch a ride, and all of the attendant problems). That alone is going to be expensive. Then you have to bail back to the idling orbit, which will also cost you some serioud delta-v.

With sufficiently advanced technology I've got one figure for being able to achieve a final speed of .012 c in a close slingshot of the sun (there's a link to the paper back in another post), but I don't know how that was derived. I suspect that was probably using the advanced carbon nanotube sail. The limits of something we could build I'm less clear on.

Yes, I'm always somewhat suspicious by the phrase "sufficiently advanced technology".

I don't think PERMANENT talks about dopants, what are the ones that are commonly used in industry?

I've found PERMANENT to be annoyingly short on the nitty gritty, when it's the nitty gritty that kills you.

Generally, p-type dopants are IIIA elements like boron or aluminum and n-type dopants are VA elements like phosphorous and arsenic. However, the efficiency of the collecter differs a bit between combinations of dopants.

Anyway, if photovoltaic cells can't be manufactured there is thermal solar. It's just a mirror that concentrates light on a container full of water or some other liquid or gas, which's expansion or boiling from the heat drives a turbine. Aside from the dynamo pretty much all you need is a water, an iron tub, and a big foil mirror.

You also need a turbine, where even a crude example is not trivial to make. Your plants are probably going to have pitiful efficiency to be easily manufactured in situ, overcoming their inefficiency by having so damn many of them.

Why take them with you? Build them, use them to power your replication and refueling, and then leave them behind. You can make more in the next system, assuming you have enough energy to get started (i.e. stick a much more modest solar panel on your probe that it takes with it).

You need to take along your start-up power source, Tweedledum. The ones you're going to leave behind can be pretty flimsy (unless you're planning to maneuver to catch comets), but the one you take with you to the next system needs to be able to survive the trip.

Yes, you have limited energy until you have your power plant set up. That just means you'll be limited while you're building your power plant and the basic infrastructure you need to manufacture its components. As you build it up you increase capacity which lets you increase the build rate along with the size of the power plant, and once it's in place you move on to other stuff, like actual self-replication and (non-urgent) self-repair.

And your order of magnitude calculation of how much energy you need is...

But there's no reason the survival impulse has to be the first priority.

It's not a top-level goal. However, survival is going to be paramount by default, because sacrificing oneself to the top-level goal is irreversable and not to be taken lightly.

Yes, survival as the default assumption makes sense, but there will be situations in which the proper response is not survival (for instance, you're about to captured by possibly hostile aliens). The intelligence must be capable of recognizing these situations. There may be short-cuts used to eliminate from consideration places where mission and survival obviously parallel ("better move out of the way of that meteorite!"), but there's a limit to how miserly you want to be with processing capacity if you want a capable intelligence. A really effective probe of this type will probably need the effective intelligence of a smart animal if not a human.

What we have here is two opposed constraints: the probe's AI hardware needs to be simple enough to be replicatable in situ, yet beefy enough to support an effective enough intelligence, which is also safe enough to use in a von Neumann probe. This, I think, is an absolutely non-trivial set of conditions to satisfy.

Obviously, you'd want to install safeguards against such "mutation". Redundant programs and/or computer cores with defective ones automatically being shut down is one that springs most immediately to mind. Another is massively redundant error-proofing systems in copying. Another is multiple redundant subroutines that would be dedicated to determining whether an individual is a "mutant" and fixing or self-destructing it if it is.

So you have a balance between safety of the AI against the resources constraints of single and multiple processors. If you have to run multiple copies of the AI so that they can check each other, that multiplies the number of cores that must be successfully constructed by the previous VNM. It will also increase the amount of energy you need. I don't think anything less than comparing to a full copy is an adequate check of whether the AI is malfunctioning — how else do you evaluate whether the AI is operating in the best interests of the top-level than to go through the decision analysis that is really the bulk of the work?

Even multiple copies do not guard against all error, such as programming bugs. Because they are deterministic, an AI of this sort will respond (and glitch) exactly the same way.

[Junghalli]

Hale-Bopp is nothing particulary extraordinary for a long-period comet, but that's a long way from saying that that it's a former KBO. However, the Kuiper belt does have a higher deuterium abundancy than Jupiter, but only about 5x that of Jupiter, giving an abundance very close to that of the world's oceans.

Ah, thank you. Do you have a source for that? I ask not because I doubt your word but because I'm curious to read it.

That doesn't mean the technology scales down to probe-sized with comparable efficiency.

Granted.

That battery has a finite number of charges in it before it fails. Charging induces physical stresses on the battery elements because the charge carriers are being traded off between the elements, and those charge carriers take up room. Furthermore, the battery has to be kept within a critical temperature range or it stops working, and you're probe is dead. That means heating/refrigerator elements to keep within this range, which also draw power. You need to make a new battery before your current one fails, so you are in a bit of a hurry.

How much of a hurry? Regenerative braking systems on cars have charge put into them many times a day and only need replacement on a timescale of years. Flywheels often have quoted lifespans of 100K to 10 million cycles of use (ref - it's Wikipedia but it's sourced). Also, you can reduce wear on your energy storage systems by running your electronics directly off solar power when it's available (although that will create wear elsewhere; everything a machine does causes wear somewhere).

Also, if that battery doesn't have enough juice to start the fusion plant, then you have to keep the plant ticking over on its own. If the plant stalls below, say, 100 MW (raw production), then by your stated 6.2e14 J/kg, you need one kg every 71 days, or 5 kg a year just to keep the engine ticking over. 5 kg of deuterium is about 1.25 kilomoles, and requires 1.25 kilomoles (25 kg) of doubly-heavy water. At an abundance of 1/340,000 (less for Hale-Bopp type comets, more for Jupiter abundancies), this gives us 8517 tonnes of water processed per year just to keep the plant ticking over. That's a 125 m wide comet nucleus. This is about the size of the dinkiest of the dinky comets, so the entire comet would have to be deconstructed just for a year's fuel.

Kick-start from "battery" in the high MJ or low GJ range might not be infeasible. Modern FES systems can have energy densities in the 500 KJ/kg range (ref); 100 MJ could supplied by a 20 kg flywheel. Nevertheless, the problem you site is a valid concern. If one must keep the reactor constantly ticking over at high energy then probably be best strategy would be seek out relatively large icy bodies (in the range of several hundred meters to a kilometer or more). The decreasing surface to volume ratio of larger bodies means you will be able to get much more material from a somewhat larger body, and so will waste less time constantly looking for fresh deuterium sources.

The one part in 6700 figure stated is probably for the abundance of finding at least a single deuterium in the molecule. The 340,000 figure probably for the abundance of heavy water proper, that is D2O, so the ratio of singly heavy water to doubly heavy water is 50::1, so the Vamork enrichment of doubly heavy water is around 2.96e-3. If we assume that enrichment follows an S-curve with respect to energy, with normal enrichment coincident with zero energy, we can extrapolate to 99.9% enrichment. If the energy/enrichment function is f = [tan⁻¹(a e + b) + 1]/2, with a,b to be determined, with boundary conditions of e = 0, f = 1/340,000 and e = 42 GJ/kg, f = 2.96e-3, this gives us enirchment constants of a = 4.7729032658333e-13 kg/TJ, b = -1.5573875747288. This gives us a specific enrichment energy of about 6.5 TJ/kg. This is about 1% of the energy liberated by D-D fusion — encouraging, as you can get more energy out than put in, at least for this stage, provided you can process the material fast enough.

Of course this is for a Vamork-type facility, which has the advantage of scale, but the disadvantage of being WWII techonology. This plant also produced only fission-reactor grade heavy water, not fuel-grade water. A fuel plant would require more stages and more equipment, and this is for a ~300 barrel throughput. No doubt a probe-sized plant would have reduced throughput and lower efficiency.

Thank you, it helps to actually have some numbers to work with. This is relatively encouraging - even at an order of magnitude less efficiency than the Vamork facility you'll still be getting a lot more energy out of the fuel than you put into extracting it.

A comet spends at best a year in the sunlight, and if your comet has a period of thousands of years, that's going to slow your growth rate by a factor of thousands. (And take care in your comet choice: some have periods of millions of years.)

I think you misunderstand. I meant the probe replicates rapidly during the short period the comet is in sunlight once (i.e. within a time span of months), so by the time the comet gets too far away from the sun for solar power to be much use it has already finished.

Unless you want your orbital adjustment to take many orbits, which essentially brings you right back where you started, you'd need a torch drive; you need to make serious changes to the orbit of a massive object in a short time.

True. On the plus side, you have access to a large heat sink, many thousands of tons of potential propellant, and abundant solar power (as long as you can build solar a power plant out of the locally available materials - structural strength will be a concern but not very much of one, as acceleration will be quite gentle). A torch drive is much more doable under these circumstances than it is for the probe itself. Solar powered mass drivers would probably be the way to go.

If you're jumping onto and off of comets during these periods of the comets' orbits, then you're going to have expensive transfers.

If you can obtain your propellant readily from the comets, or if you're using a solar sail that doesn't need propellant, I don't see why this is a big problem.

Yes, I'm always somewhat suspicious by the phrase "sufficiently advanced technology".

The best thing I could find was this paper which talks about propelling a spacecraft by accelerating many small lightsails with lasers and having them slam into a magsail as plasma (vaporized by an onboard laser on the ship). The light sails are on the order of 10 cm in diameter, have a mass of milligrams, and the paper talks about being able to launch one every few seconds. Since the sails must travel at .1 c to be able to catch up with the ship at the end of its acceleration this suggests acceleration of hundreds of thousands of G. They confirm in the paper they are talking about "a microsail capable of accelerating from rest to relativistic velocities in less than a second." This is using "only known physics and materials, although maximum system performance depends on improvements in materials, especially sail properties", and they recommend a diamond film sail. By a quick calculation, a 1 km sail might be able to take accelerations of around 30 G based on that.

Generally, p-type dopants are IIIA elements like boron or aluminum and n-type dopants are VA elements like phosphorous and arsenic. However, the efficiency of the collecter differs a bit between combinations of dopants.

Aluminum should certainly be readily available from metallic or C-type asteroids, and according to this chart red phosphorus should also be pretty common in C-type asteroids (1.1 million ppb), and arsenic is also available but in lower quantity (1800 ppb).

You need to take along your start-up power source, Tweedledum. The ones you're going to leave behind can be pretty flimsy (unless you're planning to maneuver to catch comets), but the one you take with you to the next system needs to be able to survive the trip.

Yes, you need a start-up power source, but my point is you're only on a tight energy budget while you're still limited to it. Save for the ones that may be involved in building your power plant you can delay any high-energy processes you need to do until your power plant is complete. Also, if you want to take a large flimsy construct with you, you can always limit acceleration to some small fraction of a G. Highly efficient rockets tend to have low thrust anyway. A bigger concern is the extra fuel you will need to accelerate a large power plant.

And your order of magnitude calculation of how much energy you need is...

How much energy does it take to manufacture a solar panel?

According to this site this company invests 45 kWh in the manufacture of a solar panel. A kilowatt-hour is 3.6 megajoules, so that's 162 megajoules to make a panel. They estimate 768.1 kWh with the sourcing and processing of raw materials. This probably involves shipping it over much of the planet, so it's probably pessimistic, but let's use that. We get 2.8 gigajoules per panel. Unfortunately it doesn't say how big their panels are, but 1 m^2 sounds reasonable. To manufacture 1 m^2 of panel per day would therefore require a generating capacity of 32.4 kW. According to Wikipedia, sunlight conversion rates of solar panels vary between 5-18%, let's use 10%, which gives us 137 watts/m^2. 32.4 kW would then require a square solar panel between 6-7 meters on a side. Our panel will require around 8 months to "pay back" the energy invested in its manufacture. Assuming speed and acceleration are no great priorities I see no reason the probe couldn't carry a "seed" panel array of considerably greater size than a 7 meter square. It might also be profitable to invest some of that energy in creating a large foil mirror to reflect more sunlight on the panel and speed up the work, depending on how much light your panels can take before it starts to damage them. Or, as a foil mirror could be quite large and light, the probe might already have one with it from the previous replication.

Note that this assumes little better than present technology, save in the ability to miniaturize the solar panel factory and its support infrastructure. For a civilization that's seriously exploring the stars, for instance, 10% efficient solar panels strike me as a probably overly pessimistic assumption.

It's not a top-level goal. However, survival is going to be paramount by default, because sacrificing oneself to the top-level goal is irreversable and not to be taken lightly.

As I said, in most circumstances survival will be the default assumption, but there will be instances where it isn't and the entity must be capable of recognizing those instances.

What we have here is two opposed constraints: the probe's AI hardware needs to be simple enough to be replicatable in situ, yet beefy enough to support an effective enough intelligence, which is also safe enough to use in a von Neumann probe. This, I think, is an absolutely non-trivial set of conditions to satisfy.

You are correct. I never said a self-replicating probe was a trivial thing to build. By our standards it would require incredible technological prowess. But then, by the standards of the Middle Ages a jet fighter would require incredible technological prowess. I tend not to see "requires incredible technological prowess" to be something that is likely to stop civilizations hundreds or thousands of years more advanced than ours. I realize this sounds like a vague appeal to superior technology, but I see no reason not to assume that what looks very difficult to us won't be quite doable to a civilization of sufficient technological sophistication. Exceptions, obviously, for things that just break physics or any remote degree of practicality, like FTL.

[Junghalli]

Ack, I screwed up one of my calculations. A 100 MJ charge could be stored on a 200 kg flywheel, not a 20 kg one.