Grey Vs. Blue? What Kind of Goo Are You?
August 4, 2006 1:33 PM Subscribe
Grey Goo? The maximum rate of global ecophagy by biovorous self-replicating nanorobots is fundamentally restricted by the replicative strategy employed...
If grey goo were possible, it would have evolved already.
posted by delmoi at 2:06 PM on August 4, 2006
posted by delmoi at 2:06 PM on August 4, 2006
I understood it, and I think the assumptions they are making are garbage. The Grey Goo argument is fundamentally silly. Cells are nanomachines, and they never run to grey goo because they are limited by competition from other cells, energy constraints, availability of trace minerals, etc. These calculations assume there's no wear & tear, that all the things you need for making diamond are in biological lifeforms, that you can perform assembly "reversably" and thus minimize entropy production. There's very little reality in these assumptions.
posted by freedryk at 2:13 PM on August 4, 2006
posted by freedryk at 2:13 PM on August 4, 2006
I, for one, welcome our biovorous self-replicating grey goo making future miniature robot leaders.
posted by Astro Zombie at 2:17 PM on August 4, 2006
posted by Astro Zombie at 2:17 PM on August 4, 2006
freedryk: I think the idea is that if grey goo is impossible and/or improbable under ideal circumstances, then it should be fairly safe to assume it is improbable and/or impossible in the real world.
posted by Freen at 2:20 PM on August 4, 2006
posted by Freen at 2:20 PM on August 4, 2006
haven't had time to read the whole thing yet, but it's interesting that people at zyvex are thinking about this stuff. they sell things used in the top-down approach to nanotech (we just bought one), as opposed to the bottom-up molecular-machinery approach. i wonder if this means they're dangling their feet into the molecular water. hmm.
freedryk: all you need for making diamond is carbon. also, for self-assembly to occur it has to be a reversible process, and that occurs really quite often. not that i disagree, mind you, just that these assumptions aren't that far out there.
posted by sergeant sandwich at 2:26 PM on August 4, 2006
freedryk: all you need for making diamond is carbon. also, for self-assembly to occur it has to be a reversible process, and that occurs really quite often. not that i disagree, mind you, just that these assumptions aren't that far out there.
posted by sergeant sandwich at 2:26 PM on August 4, 2006
Freen: do you mean that is the idea of the "Limits to Global Ecophagy" article? Because the abstract seems to be talking about how you would detect that a Grey Goo event was happening, and that it would take ~20 months to happen--which to me, indicates they think it's both a) possible and b) relatively easy to do.
posted by freedryk at 2:36 PM on August 4, 2006
posted by freedryk at 2:36 PM on August 4, 2006
If grey goo were possible, it would have evolved already.
?!
Evolution can do all kinds of stuff, but I'd guess it stops pretty well short of being an exhausitve search of all possibilities.
posted by namespan at 2:40 PM on August 4, 2006
?!
Evolution can do all kinds of stuff, but I'd guess it stops pretty well short of being an exhausitve search of all possibilities.
posted by namespan at 2:40 PM on August 4, 2006
sgt. sandwich: just because diamond is pure carbon does not mean that only carbon is needed to manufacture it. For instance, I'm betting you can't make little robot arms to manipulate atoms using only carbon. And don't reversible processes imply no entropy gain? I thought they were essentially a theoretical idea to calculate maximum possible efficiencies for a process.
posted by freedryk at 2:41 PM on August 4, 2006
posted by freedryk at 2:41 PM on August 4, 2006
Evolution can do all kinds of stuff, but I'd guess it stops pretty well short of being an exhausitve search of all possibilities.
Well, that's true. But the bigger point is that it ignores the "green goo" that already covers the planet. Self-replicating nano-bots would just be another form of life, and the question is, why would we assume that these self-replicating nano-bots would be more efficient then the life forms that have taken billions of years to optimize? Of course, the human creative process is another type of 'search' and so it's possible that we could create something novel, if evolution were stuck in a local minimum. (like wheels and roads, for example, or space-ships).
And don't reversible processes imply no entropy gain?
If two states have equal entropy, then they can happen and 'unhappen' at random. There's no way to control it.
posted by delmoi at 2:54 PM on August 4, 2006
Well, that's true. But the bigger point is that it ignores the "green goo" that already covers the planet. Self-replicating nano-bots would just be another form of life, and the question is, why would we assume that these self-replicating nano-bots would be more efficient then the life forms that have taken billions of years to optimize? Of course, the human creative process is another type of 'search' and so it's possible that we could create something novel, if evolution were stuck in a local minimum. (like wheels and roads, for example, or space-ships).
And don't reversible processes imply no entropy gain?
If two states have equal entropy, then they can happen and 'unhappen' at random. There's no way to control it.
posted by delmoi at 2:54 PM on August 4, 2006
As long as robots (self-replicating or not) can't even safely cross a street without getting run over, we're still lightyears away from grey goo, so there's not much to be worried about.
posted by sour cream at 3:03 PM on August 4, 2006
posted by sour cream at 3:03 PM on August 4, 2006
freedryk, what other elements are you suggesting one needs to manufacture diamond?
the point of molecular machinery is that you have little robot arms built from custom-designed protein molecules. i don't know what atoms one might need for a protein that one can't find in other biological systems.
posted by sergeant sandwich at 3:07 PM on August 4, 2006
the point of molecular machinery is that you have little robot arms built from custom-designed protein molecules. i don't know what atoms one might need for a protein that one can't find in other biological systems.
posted by sergeant sandwich at 3:07 PM on August 4, 2006
This biovorous self-replicating grey goo, it vibrates?
posted by Smedleyman at 3:12 PM on August 4, 2006
posted by Smedleyman at 3:12 PM on August 4, 2006
Astro Zombie - that has got to be the best example of the "I, for one, welcome ... " meme EVER! ROTFL!
posted by sirvesa at 3:18 PM on August 4, 2006
posted by sirvesa at 3:18 PM on August 4, 2006
No, really, why can’t mad science go awry to create something I can insert my penis into? Whoops, we accidentally created these crazy flying blowjob machines, sorry!
Part of the problem with the grey goo I think is that it does ultimately wear out, but not before it eliminates all the biomass. Bacteria and plankton, et.al. eventually dies but the waste plugs nutrients back into the system. Grey goo wouldn’t/doesn't.
posted by Smedleyman at 3:24 PM on August 4, 2006 [1 favorite]
Part of the problem with the grey goo I think is that it does ultimately wear out, but not before it eliminates all the biomass. Bacteria and plankton, et.al. eventually dies but the waste plugs nutrients back into the system. Grey goo wouldn’t/doesn't.
posted by Smedleyman at 3:24 PM on August 4, 2006 [1 favorite]
namespan: Evolution can do all kinds of stuff, but I'd guess it stops pretty well short of being an exhausitve search of all possibilities.
Well, I don't think you can claim that it is an exhaustive search of all possibilities. But over long periods of time, it is a highly effective one.
From a biochemestry perspective, evolution is an algorithm that seeks the maximum total efficiency from any chemical process. Even subtle variations in efficency can translate into a huge reproductive advantage over generations.
A basic problem faced by organic life is that while there is a heck of a lot of energy out there, it comes in the form of thousands of different chemical compounds. Evolution seeks the most efficient path towards cracking these compounds. So rather than using a brute force approach, enzymes function as catalysts which reduce needed energy for a reaction to happen. The notion of enzymes as little organic robots that tear apart or construct compounds is a bit of a misconception.
But, each of those thousands of different compounds has a different shape, different weak points, and different activation energies. One of the easiest ways around this is to develop specialized processes for each class of compound. The shape of many molecules varies slightly depending on temperature, so for maximum efficency, enzymes are tailored to temperature ranges. (You can go down the list with other factors as well, pH, salinity, etc., etc..)
A superbug that could eat anything would have a huge competitive advantage, and should dominate any ecosystem. The fact that we've never seen such a superbug suggests that generalization is not that efficient of a strategy.
sergeant sandwitch: all you need for making diamond is carbon.
And a huge chunk of energy, and the ability to keep that carbon from oxidizing.
Well, let's just explore that for a moment. Carbon is a limiting element for life, in spite of its ubiquity. The problem is not that carbon is rare, the problem is that most of the carbon in biosphere is in a form that requires quite a bit of work to process into a usable form.
If you are a autotroph, the problem is how to convert large quantities of inorganic carbon into something that can be used to build complex organic molecules with minimal energy expense. You need energy, but that energy needs to be in a form that is neither too volatile or too stable. You need hydrogen, which is dangerous to have around in many forms, and useless in other forms. You need to have some way of getting oxygen out of your system, because oxygen is toxic for most processes.
If you are a heterotroph, great, you can live carbon that has been fixed by others. But then again, it's not that simple. That carbon comes in a wide variety of different molecules, and an enzyme for lactose isn't going to do you any good if you are living on the bark of a tree. As with autotrophs, you have to have some way of keeping oxygen out of processes it poisons, and available for processes where it is needed. You are also going to end up creating a fair quantity of inorganic nitrogen and phosphates, which also need to be isolated.
The problem with creating diamond isn't abundance, it's the energy cost of creating relatively pure carbon, and the energy cost getting diamond rather than graphite.
posted by KirkJobSluder at 4:00 PM on August 4, 2006
Well, I don't think you can claim that it is an exhaustive search of all possibilities. But over long periods of time, it is a highly effective one.
From a biochemestry perspective, evolution is an algorithm that seeks the maximum total efficiency from any chemical process. Even subtle variations in efficency can translate into a huge reproductive advantage over generations.
A basic problem faced by organic life is that while there is a heck of a lot of energy out there, it comes in the form of thousands of different chemical compounds. Evolution seeks the most efficient path towards cracking these compounds. So rather than using a brute force approach, enzymes function as catalysts which reduce needed energy for a reaction to happen. The notion of enzymes as little organic robots that tear apart or construct compounds is a bit of a misconception.
But, each of those thousands of different compounds has a different shape, different weak points, and different activation energies. One of the easiest ways around this is to develop specialized processes for each class of compound. The shape of many molecules varies slightly depending on temperature, so for maximum efficency, enzymes are tailored to temperature ranges. (You can go down the list with other factors as well, pH, salinity, etc., etc..)
A superbug that could eat anything would have a huge competitive advantage, and should dominate any ecosystem. The fact that we've never seen such a superbug suggests that generalization is not that efficient of a strategy.
sergeant sandwitch: all you need for making diamond is carbon.
And a huge chunk of energy, and the ability to keep that carbon from oxidizing.
Well, let's just explore that for a moment. Carbon is a limiting element for life, in spite of its ubiquity. The problem is not that carbon is rare, the problem is that most of the carbon in biosphere is in a form that requires quite a bit of work to process into a usable form.
If you are a autotroph, the problem is how to convert large quantities of inorganic carbon into something that can be used to build complex organic molecules with minimal energy expense. You need energy, but that energy needs to be in a form that is neither too volatile or too stable. You need hydrogen, which is dangerous to have around in many forms, and useless in other forms. You need to have some way of getting oxygen out of your system, because oxygen is toxic for most processes.
If you are a heterotroph, great, you can live carbon that has been fixed by others. But then again, it's not that simple. That carbon comes in a wide variety of different molecules, and an enzyme for lactose isn't going to do you any good if you are living on the bark of a tree. As with autotrophs, you have to have some way of keeping oxygen out of processes it poisons, and available for processes where it is needed. You are also going to end up creating a fair quantity of inorganic nitrogen and phosphates, which also need to be isolated.
The problem with creating diamond isn't abundance, it's the energy cost of creating relatively pure carbon, and the energy cost getting diamond rather than graphite.
posted by KirkJobSluder at 4:00 PM on August 4, 2006
But the bigger point is that it ignores the "green goo" that already covers the planet. Self-replicating nano-bots would just be another form of life, and the question is, why would we assume that these self-replicating nano-bots would be more efficient then the life forms that have taken billions of years to optimize?
How long do you think it will be before we engineer them to evolve themselves?
And even if we don't; we engineer them to "die" - remember the lesson of Jurassic Park! And be very, very afraid.
Or something.
posted by hurkle at 4:01 PM on August 4, 2006
How long do you think it will be before we engineer them to evolve themselves?
And even if we don't; we engineer them to "die" - remember the lesson of Jurassic Park! And be very, very afraid.
Or something.
posted by hurkle at 4:01 PM on August 4, 2006
Part of the problem with the grey goo I think is that it does ultimately wear out, but not before it eliminates all the biomass. Bacteria and plankton, et.al. eventually dies but the waste plugs nutrients back into the system. Grey goo wouldn’t/doesn't.
Bio-waste is nutrient only because organisms evolved which could digest it. The pre-requisite is some net energy gain for the organism taking in the waste. This wouldn't necessarily be any different for any other kind of waste.
posted by Chuckles at 4:38 PM on August 4, 2006
Bio-waste is nutrient only because organisms evolved which could digest it. The pre-requisite is some net energy gain for the organism taking in the waste. This wouldn't necessarily be any different for any other kind of waste.
posted by Chuckles at 4:38 PM on August 4, 2006
hurkle: How long do you think it will be before we engineer them to evolve themselves?
You can't have something "evolve" its self.
But any population of self-replicating things will evolve in the sense that variations in replication will occur, and variations in fitness will result in changes in trait frequency over time.
I would like "Grey Goo" hypotheses more if they actually adressed some basic biochemestry. The foresight.org article just does a handwave in regards to the availability of matter by treating the biosphere as just CHON. Drexler makes a revealing error in equating photosynthesis to photovoltaics. Energy is a byproduct of photosynthesis, which developed to reduce carbon. Drexler has of late backed away from the goos, considering them improbable.
posted by KirkJobSluder at 4:43 PM on August 4, 2006
You can't have something "evolve" its self.
But any population of self-replicating things will evolve in the sense that variations in replication will occur, and variations in fitness will result in changes in trait frequency over time.
I would like "Grey Goo" hypotheses more if they actually adressed some basic biochemestry. The foresight.org article just does a handwave in regards to the availability of matter by treating the biosphere as just CHON. Drexler makes a revealing error in equating photosynthesis to photovoltaics. Energy is a byproduct of photosynthesis, which developed to reduce carbon. Drexler has of late backed away from the goos, considering them improbable.
posted by KirkJobSluder at 4:43 PM on August 4, 2006
delmoi wins! And I'm very glad that comment was made quickly. namespan, Just replace "possible" with "easy", same story practically speaking. Grey goo is fear mongering bullshit along with dirty bombs, etc. Life ain't *that* simple people.
posted by jeffburdges at 4:54 PM on August 4, 2006
posted by jeffburdges at 4:54 PM on August 4, 2006
The most interesting thing I've taken away from a nanotech class I took was how easily engineers get excited about mundane sutff.
Look! We've got a giant multimillion dollar factory that consumes megawatts of power! All to make a tiny thing that can spin. (well, one out of a hundred). It can only do it in an ultracleen room, under deep vacuum. Hope it does not get out in the wild, cause it might take over the world.
Of course the puddle in the parking lot of the factory contains billions of much more sophisticated machines that can do orders of magnitude more.
How long do you think it will be before we engineer them to evolve themselves?
You can't make them evolve themselves for a simple reason that an individual never evolves. Only populations do.
posted by c13 at 5:08 PM on August 4, 2006
Look! We've got a giant multimillion dollar factory that consumes megawatts of power! All to make a tiny thing that can spin. (well, one out of a hundred). It can only do it in an ultracleen room, under deep vacuum. Hope it does not get out in the wild, cause it might take over the world.
Of course the puddle in the parking lot of the factory contains billions of much more sophisticated machines that can do orders of magnitude more.
How long do you think it will be before we engineer them to evolve themselves?
You can't make them evolve themselves for a simple reason that an individual never evolves. Only populations do.
posted by c13 at 5:08 PM on August 4, 2006
how easily engineers get excited about mundane sutff.
I'm really torn on this.. Fact is, it is always mundane stuff. In aggregate, mundane stuff can do pretty interesting things! In fact, turn your statement around a little, and you can end up with a pretty good definition of what engineers do.
posted by Chuckles at 5:14 PM on August 4, 2006
I'm really torn on this.. Fact is, it is always mundane stuff. In aggregate, mundane stuff can do pretty interesting things! In fact, turn your statement around a little, and you can end up with a pretty good definition of what engineers do.
posted by Chuckles at 5:14 PM on August 4, 2006
This worry about self-replicating entities taking over the world seems to be somewhat borne out by the previous caterpillar post....
posted by pombe at 5:22 PM on August 4, 2006
posted by pombe at 5:22 PM on August 4, 2006
You can't have something "evolve" its self.
Oh, come on, you know what he means. He's talking about autonomous self-design. Don't be such a scientist.
Nanobots that eat the world to death? Why, that seems practically benign. Why not design nanobots that assemble and detonate nuclear weapons to disperse themselves, like some kind of apocalyptic puffball fungus? Or ones that invade and reconstruct larger organisms, grotesquely and horrifyingly reshaping the bodies of humans for their own purposes, à la Greg Bear's Blood Music.
posted by XMLicious at 5:23 PM on August 4, 2006
Oh, come on, you know what he means. He's talking about autonomous self-design. Don't be such a scientist.
Nanobots that eat the world to death? Why, that seems practically benign. Why not design nanobots that assemble and detonate nuclear weapons to disperse themselves, like some kind of apocalyptic puffball fungus? Or ones that invade and reconstruct larger organisms, grotesquely and horrifyingly reshaping the bodies of humans for their own purposes, à la Greg Bear's Blood Music.
posted by XMLicious at 5:23 PM on August 4, 2006
I appreciate the technological aspect of it. Ion implanation, e-beam machining, AFM -- all really cool stuff. But the whole grey goo stuff is just so naive its cute. Besides, the great dream is to get, after all the work done, something that maybe approaches a simplest bacterium in complexity.
posted by c13 at 5:24 PM on August 4, 2006
posted by c13 at 5:24 PM on August 4, 2006
Modeling a folding protein will happen the magic '50' years after fusion. Until that great computing triumph, leave the entropic grey goo to bacteria. They do well enough at it for now. In the most part in my pants.
posted by econous at 6:30 PM on August 4, 2006
posted by econous at 6:30 PM on August 4, 2006
Whatever Michael Crichton has argued, he is grey goo just like you, and his mommy.
posted by econous at 6:40 PM on August 4, 2006
posted by econous at 6:40 PM on August 4, 2006
I thought the Christains were worried about the gay goo bringing about the Apocalypto.
posted by orthogonality at 7:40 PM on August 4, 2006
posted by orthogonality at 7:40 PM on August 4, 2006
« Older Fokke & Sukke | Jason Rhoades, 41 Newer »
This thread has been archived and is closed to new comments
posted by isopraxis at 1:41 PM on August 4, 2006