Imagine that advanced life like us is terribly rare in the universe. So damn rare that if we had not shown up, then our region of the universe would almost surely have forever remained dead, for eons and eons. In this case, we should still be able to predict
Some assumptions in the model are that intelligent life is rare in that a lot of habitable worlds don't produce an intelligent space faring species even if the planet and star is perfect, as otherwise life's arrival on earth only really tells you about ~mode arrival time of life on a planet, and not relative rank. The paper the figure is from actually shows that humans aren't necessarily even 10% early, however they only manage this if the gamma ray bursts and close supernovae aren't important to habitability and ~solar like stars are almost the only type which is habitable. If you allow red dwarfs, or even K types then human beings are still really early.
I see this as a counter to the rare earth or rare intelligence hypothesis, in that even if rare, if Grabby they're still everywhere. If they're not-rare then they're still everywhere. At least that was what I gathered.
Can you detail the mechanism by which you use future events to explain current state of the universe? Or reference me to a post explaining it?
Can you contrast the following two theories?
"Life appeared on Earth at this early time because later Earth would be already taken by grabby aliens."
"Tea appeared surprisingly early in human history because in the future Starbucks is destined to monopolize all political power on Earth and destroy all samples of tea in existence, to promote coffee-drinking."
Does this model also take into account that an estimated "95% of all stars that will ever exist, already exist", and that Earth may be around the 50th percentile of all planetary systems that will exist?https://thegreatatuin.wordp...
The jury appears to be out on whether we are in fact early.
The following may also be relevant:>although smaller stars are much more common than larger stars (the Sun is actually larger than over 80% of stars in the universe) stars smaller than about 0.5 solar masses (and thus 0.08 solar luminosities) are usually ‘flare stars’ – possessing very strong convoluted magnetic fields and periodically putting out flares and X-ray bursts that would frequently strip away the ozone and possibly even the atmosphere of an earthlike planet.
I see you mention ocean worlds, but why should they be immune to flares? More generally, you do not seem to make any arguments for why we should expect life to be able to emerge beyond "a narrow range of paths similar to Earth’s path".
GRB are included in most estimates of habitable star formation and occupation rates. They are why such distributions tend to peak after 12Gyr, instead of the 4Gyr time when star formation peaks. This is very much included, indirectly, in our analysis.
Is there a chance gamma ray bursts kept resetting terrestrial and shallow ocean life, thereby also resetting the clock on big brains? Gamma ray bursts were more numerous in the past within our own galaxy, and perhaps this is true for the evolution of galaxies across the universe. The estimates for GRB frequency seem to be pretty wide.Here is an article on that, which suggests a 60% chance 1 billion years ago and a 90% 5 billion years ago of a dangerous GRB striking us: https://physics.aps.org/art...
Other things that are pointed to:High metallicity suppresses dangerous gamma ray bursts (also relevant when considering the universe as a whole, since metallicity has increased over time).Our galaxy is said in the article to have 1/10th the average rate of bursts.Early galaxies were smaller and more dangerous.
3) well, this t^n clumping says the borders are where the complexity is. other kinds of clumping put the complexity at the nucleation centers (gravitational matter clumping, for example).
no, eventually it's too big for anyone to control it. above some size, you don't have edges, you have a crumbling into mutants whose main concern is behind them, so frontier expansion slows down and becomes irregular. parts of volumes splinter, those directions might be preferentially targeted by external competitors, local preclusion there may become less effective, etc. i'm pointing out that you're depending on uniform symmetric expansion into internally homogeneous stable volumes, but coordination definitely depends on scale.
is there any plausible way that a civ could coordinate to remain homogenous across these spatial scales? won't parts mutate and begin to compete against the center? won't this effect dominate the dynamics?
what can be learned from looking at "preclusionary" species in history of earth ecology? i'm thinking of trees fighting for canopy space for example. i think the end results seem to be one species wins the whole biome, rather than stable frontiers of interaction between different nucleation centers. ie the dynamics (or equilibrium, at least) comes from founder effects and terrain constraints, not analogs to the parameters in this model. are we then not more likely to be like a mushroom, which doesn't realize that it missed most of the real action? that is, few species could have predicted their experience using grabby assumptions about others -- most experiences play out in a landscape that isn't dominated by grabby dynamics, especially the preclusionary part.
1) Even in the cracks you don't see the others if v near c.2) filaments are not much longer than 50Mly, but we are talking 10x that in ave distance between them.3) all clumping looks similar
1) if n is high, then the most common type of civ is one that arises between the cracks of several other huge civs about to collide, among a froth of other nascents about to be steamrolled. doesn't this predict that things should look a lot more heterogeneous from where we sit? ie observed isotropy says n is low?
2) the spherically symmetric expansion depends on homogeneous space, but on the scales you're talking about, matter is distributed along webby filaments of superclusters, right? doesn't the anisotropic substrate at the relevant scale destroy a lot of the symmetries necessary for a simple model to capture the dynamics?
3) speaking of the cosmological structure, i'm struck by how the voids between the filaments, compared to the filaments themselves, mimics the exact structure of this model. could the voids be the centers of early grabbies, with the filaments being the zone of interaction/conflict and rapidly appearing latecomers?
Counterargument to (1): In the Grabby Aliens model (and, I think, in a lot of other models of alien civilizations - e.g. Dark Forest Theory): If we don't expand, then we inevitably get taken over eventually; if we expand, then we may still get taken over, but we have a chance of not being taken over. Because the space-of-all-intelligent-civilizations is pretty big, I think that means that aliens will probably not share the same values* as we do, so getting taken over will probably be bad for us. And even if the loud aliens mess up their expansion technology, the thing-that-the-expansion-technology-optimizes-for is probably not going to be aligned with our values.
Counterargument to (2): I don't think whatever we do to try to copy our minds into distant galaxies will work. Each galaxy is separated from us by many light years, meaning that our message will probably be hopelessly outdated when it reaches its recipients, and sheer distance to the stars will make targeting difficult. (Even if we sent a message now to Proxima Centauri, our closest neighboring star, it would arrive in 2025, our time.) And even if somehow the recipients get our messages, they might actively ignore them (believing them to be adversarially optimized), or they might just not understand us because of our chosen data format.
In conclusion, I am mildly in favor of us being loud, but I'm highly uncertain about this. I eagerly await more developments in ethics & the search for aliens before I can render a decent decision.
*counter-counterargument: In a 2018 ramble/post, Scott Alexander essentially derived the universal utility function, which is basically the average happiness of all sentient beings weighted by their power (I'm leaving "happiness" and "power" intentionally under-defined, but you get the point). So perhaps if aliens are smart enough to expand-at-nearly-the-speed-of-light, then they're smart enough to optimize for the universal utility function, and the same for us.
Reposting my earlier comment from Shtetl-Optimized here, lightly edited + with extra notes:
1) Has anyone considered the possibility that we’re in an expansionist civilization’s bubble right now but we just haven’t noticed it? (If any expansionist civilizations protectively enclose developing ones in spherical “containers”, which seems likely, then this situation is bound to happen many times over.) [Note added later: As you - i.e. Robin Hanson - have pointed out, if we assume the existence of expansionist-but-not-grabby civilizations, then we can no longer estimate their speed well and are back in Drake equation territory.]
2) I believe that “Dissolving the Fermi Paradox” paper has been the best treatment of the Fermi Paradox that I’ve seen so far. However, I respect Robin Hanson for his idea. [Note added later: There are *many* other proposed solutions to the Fermi paradox, such as that we haven't been looking carefully enough for the aliens - see, e.g., the Wikipedia article on this. So from a pure outside-view perspective we should be highly uncertain about what the actual reason why we don't see aliens is.]
3) The Bayesian arguments on the Fermi paradox may turn out very differently depending on whether we adopt the Self-Indication Assumption or Self-Sampling Assumption. (I currently weakly prefer Self-Indication, but I could go either way.) This needs more attention. [Note added later: I think that the Grabby Aliens model produces roughly the same number of grabby-civs in each run, and that this is uncorrelated with the future of any given grabby-civ. So I think that the choice of SSA vs. SIA doesn't matter very much.]
Some assumptions in the model are that intelligent life is rare in that a lot of habitable worlds don't produce an intelligent space faring species even if the planet and star is perfect, as otherwise life's arrival on earth only really tells you about ~mode arrival time of life on a planet, and not relative rank. The paper the figure is from actually shows that humans aren't necessarily even 10% early, however they only manage this if the gamma ray bursts and close supernovae aren't important to habitability and ~solar like stars are almost the only type which is habitable. If you allow red dwarfs, or even K types then human beings are still really early.
I see this as a counter to the rare earth or rare intelligence hypothesis, in that even if rare, if Grabby they're still everywhere. If they're not-rare then they're still everywhere. At least that was what I gathered.
Can you detail the mechanism by which you use future events to explain current state of the universe? Or reference me to a post explaining it?
Can you contrast the following two theories?
"Life appeared on Earth at this early time because later Earth would be already taken by grabby aliens."
"Tea appeared surprisingly early in human history because in the future Starbucks is destined to monopolize all political power on Earth and destroy all samples of tea in existence, to promote coffee-drinking."
Did Robins reply convince you?If so, could you share your newfound intuition?
Does this model also take into account that an estimated "95% of all stars that will ever exist, already exist", and that Earth may be around the 50th percentile of all planetary systems that will exist?https://thegreatatuin.wordp...
The jury appears to be out on whether we are in fact early.
The following may also be relevant:>although smaller stars are much more common than larger stars (the Sun is actually larger than over 80% of stars in the universe) stars smaller than about 0.5 solar masses (and thus 0.08 solar luminosities) are usually ‘flare stars’ – possessing very strong convoluted magnetic fields and periodically putting out flares and X-ray bursts that would frequently strip away the ozone and possibly even the atmosphere of an earthlike planet.
https://thegreatatuin.wordp...
I see you mention ocean worlds, but why should they be immune to flares? More generally, you do not seem to make any arguments for why we should expect life to be able to emerge beyond "a narrow range of paths similar to Earth’s path".
GRB are included in most estimates of habitable star formation and occupation rates. They are why such distributions tend to peak after 12Gyr, instead of the 4Gyr time when star formation peaks. This is very much included, indirectly, in our analysis.
Is there a chance gamma ray bursts kept resetting terrestrial and shallow ocean life, thereby also resetting the clock on big brains? Gamma ray bursts were more numerous in the past within our own galaxy, and perhaps this is true for the evolution of galaxies across the universe. The estimates for GRB frequency seem to be pretty wide.Here is an article on that, which suggests a 60% chance 1 billion years ago and a 90% 5 billion years ago of a dangerous GRB striking us: https://physics.aps.org/art...
Other things that are pointed to:High metallicity suppresses dangerous gamma ray bursts (also relevant when considering the universe as a whole, since metallicity has increased over time).Our galaxy is said in the article to have 1/10th the average rate of bursts.Early galaxies were smaller and more dangerous.
Another older paper on the topic:https://arxiv.org/abs/astro...
3) well, this t^n clumping says the borders are where the complexity is. other kinds of clumping put the complexity at the nucleation centers (gravitational matter clumping, for example).
no, eventually it's too big for anyone to control it. above some size, you don't have edges, you have a crumbling into mutants whose main concern is behind them, so frontier expansion slows down and becomes irregular. parts of volumes splinter, those directions might be preferentially targeted by external competitors, local preclusion there may become less effective, etc. i'm pointing out that you're depending on uniform symmetric expansion into internally homogeneous stable volumes, but coordination definitely depends on scale.
I don't see how that makes much difference to the volume controlled by grabby aliens. After they fight, some of them will control the result.
I'm reminded of the Kaehler--Oh--Krummenacker theory.http://hertzlinger.blogspot...
is there any plausible way that a civ could coordinate to remain homogenous across these spatial scales? won't parts mutate and begin to compete against the center? won't this effect dominate the dynamics?
what can be learned from looking at "preclusionary" species in history of earth ecology? i'm thinking of trees fighting for canopy space for example. i think the end results seem to be one species wins the whole biome, rather than stable frontiers of interaction between different nucleation centers. ie the dynamics (or equilibrium, at least) comes from founder effects and terrain constraints, not analogs to the parameters in this model. are we then not more likely to be like a mushroom, which doesn't realize that it missed most of the real action? that is, few species could have predicted their experience using grabby assumptions about others -- most experiences play out in a landscape that isn't dominated by grabby dynamics, especially the preclusionary part.
1) Even in the cracks you don't see the others if v near c.2) filaments are not much longer than 50Mly, but we are talking 10x that in ave distance between them.3) all clumping looks similar
1) if n is high, then the most common type of civ is one that arises between the cracks of several other huge civs about to collide, among a froth of other nascents about to be steamrolled. doesn't this predict that things should look a lot more heterogeneous from where we sit? ie observed isotropy says n is low?
2) the spherically symmetric expansion depends on homogeneous space, but on the scales you're talking about, matter is distributed along webby filaments of superclusters, right? doesn't the anisotropic substrate at the relevant scale destroy a lot of the symmetries necessary for a simple model to capture the dynamics?
3) speaking of the cosmological structure, i'm struck by how the voids between the filaments, compared to the filaments themselves, mimics the exact structure of this model. could the voids be the centers of early grabbies, with the filaments being the zone of interaction/conflict and rapidly appearing latecomers?
Counterargument to (1): In the Grabby Aliens model (and, I think, in a lot of other models of alien civilizations - e.g. Dark Forest Theory): If we don't expand, then we inevitably get taken over eventually; if we expand, then we may still get taken over, but we have a chance of not being taken over. Because the space-of-all-intelligent-civilizations is pretty big, I think that means that aliens will probably not share the same values* as we do, so getting taken over will probably be bad for us. And even if the loud aliens mess up their expansion technology, the thing-that-the-expansion-technology-optimizes-for is probably not going to be aligned with our values.
Counterargument to (2): I don't think whatever we do to try to copy our minds into distant galaxies will work. Each galaxy is separated from us by many light years, meaning that our message will probably be hopelessly outdated when it reaches its recipients, and sheer distance to the stars will make targeting difficult. (Even if we sent a message now to Proxima Centauri, our closest neighboring star, it would arrive in 2025, our time.) And even if somehow the recipients get our messages, they might actively ignore them (believing them to be adversarially optimized), or they might just not understand us because of our chosen data format.
In conclusion, I am mildly in favor of us being loud, but I'm highly uncertain about this. I eagerly await more developments in ethics & the search for aliens before I can render a decent decision.
*counter-counterargument: In a 2018 ramble/post, Scott Alexander essentially derived the universal utility function, which is basically the average happiness of all sentient beings weighted by their power (I'm leaving "happiness" and "power" intentionally under-defined, but you get the point). So perhaps if aliens are smart enough to expand-at-nearly-the-speed-of-light, then they're smart enough to optimize for the universal utility function, and the same for us.
Reposting my earlier comment from Shtetl-Optimized here, lightly edited + with extra notes:
1) Has anyone considered the possibility that we’re in an expansionist civilization’s bubble right now but we just haven’t noticed it? (If any expansionist civilizations protectively enclose developing ones in spherical “containers”, which seems likely, then this situation is bound to happen many times over.) [Note added later: As you - i.e. Robin Hanson - have pointed out, if we assume the existence of expansionist-but-not-grabby civilizations, then we can no longer estimate their speed well and are back in Drake equation territory.]
2) I believe that “Dissolving the Fermi Paradox” paper has been the best treatment of the Fermi Paradox that I’ve seen so far. However, I respect Robin Hanson for his idea. [Note added later: There are *many* other proposed solutions to the Fermi paradox, such as that we haven't been looking carefully enough for the aliens - see, e.g., the Wikipedia article on this. So from a pure outside-view perspective we should be highly uncertain about what the actual reason why we don't see aliens is.]
3) The Bayesian arguments on the Fermi paradox may turn out very differently depending on whether we adopt the Self-Indication Assumption or Self-Sampling Assumption. (I currently weakly prefer Self-Indication, but I could go either way.) This needs more attention. [Note added later: I think that the Grabby Aliens model produces roughly the same number of grabby-civs in each run, and that this is uncorrelated with the future of any given grabby-civ. So I think that the choice of SSA vs. SIA doesn't matter very much.]