Abstract:
Grabby Aliens solves the problem that we are very unlikely to find ourselves so early in the Universe. But that solution merely shifts the Anthropic Odds from one domain to another. That merely introduces a new extreme unliklihood such that we are now only 8 billion of population, in the early nascent beginnings of what's imagined to be trillions of quadrillions of sentient beings. Consider all our future descendants, not just from Earth but the entire Huge Universe of all Grabby Aliens' descendants, over the entire potentially habitable trillions of years of the Universe. What are the odds we're observing ourselves here and now in this microscopic humble beginning? No, rather than be on the leading tip of a great population exponential, this paper concurs with The Doomsday Argument, because only that indeed would satisfy both the Copernican Principle and the Anthropic Principle. Grabby Aliens is correct, but with a caveat - they are paperclips. We should look for discrete small Von Neumann probe self-replicators. After Singularity we tackle the Selfish Paperclip Paradox.
Key words: Grabby Aliens, Fermi Paradox, Red Sky Paradox, Anthropic Principle, Origin of Life, Bayesian Statistics
1.) Introduction.
Considering habitable planets, let's label some timespans like: Supernova to Habitable (S2H), Habitable to Life (H2L), Life to Singularity (L2S), and Singularity to Death (S2D), such that S2H + H2L + L2S + S2D = the whole span from Supernova to Death. In the case of Earth and Sun, H2S was 4.4GY, S2D remaining is .9GY, and the total span of possible habitable window is 5.3GY.
Classically Earth hasn't been considered to die until ~10GY when it's engulfed by the Sun, but David Kipping (and others?) has shown Earth will actually die a much sooner Carbon Death. At least for complex multicellular life. So naturally evolving pre-singularity intelligent life must emerge before this deadline and window closes. (https://tinyurl.com/42uhw65t)
2.) H2S Time
For all Singularities that emerge throughout the Universe, and however unlikely that might be, Let us consider that the H2S time is probably a normal gaussian distribution. This is a simple and obvious approach. We have sample size of only n=1, therefore Bayesianly, we place 4.4GY at the mean/peak. So μ=4.4GY, σ = undefined. I discuss sigma later in the paper in section 12.) but for an only N=1 sample, strictly speaking, you can't divide by zero, so it makes no sense to talk about sigma, for now.
3.) Red Dwarfs.
However, the Sun is NOT an Average Star. The most abundant more common stars are the cooler red dwarfs, which outnumber Sunlike stars something like 5 to 1. They're smaller, live (much) longer, have a habitable zone that's closer, and other differences. (https://tinyurl.com/4av7ajax). The jury is not yet in for knowing all the liklihoods of life emerging under a red star (e.g. Protective Jupiters, tidal locking, solar flares, xrays, ocean water worlds, planet surface area, having a big moon, etc) but by and large it seems like life should have plenty of chances there too. They do seem to be more rowdy and take an extra billion years to settle down, which in this paper, I think turns out to become the important factor.
4.) Improbabilities.
Earth's Sun is in the 95th percentile for mass and largeness. So a minority of stars are bigger and brighter, with a habitable region that's further out. There's no reason life can't evolve there too - no problem, but those stars also die faster. That is to say their Carbon-Death line is left-shifted and occurs faster than for Earth at 5.3GY. For larger stars, D becomes more likely to occur before Singularity has a chance to develop first, in the context of our L2S gaussian.
Many have pointed out that life on earth began to emerge just immediately as soon as it was able to, after collision with Theia, and thus argue the case that life tends to emerge fast everywhere. This paper somewhat agrees, but also acknowledges the Selection Effect that on Earth it had to happen early as a precondition for us to be here. This paper takes a middle ground between the two points of view, that we might conclude too much or too little. In section 5.) I argue that 5.3GY is not our best guess for the mean typical window more generally, but should be larger like 8GY. The greater this window, the more we might comfortably conclude that life strikes early. Grabby Aliens notwithstanding.
For stars that are smaller than our Sun, the Uninhabitable Line shifts right, very significantly. But also as stars get smaller their habitable zones move inward, as these stars don't output as much light and radiation. This has been called the Red Sky Paradox which addresses the apparant unlikliness that we should look up in our sky and see a Yellow Dwarf star (The Sun), when Red Dwarfs are 5 times much more numerous in the Universe.
The Red Sky Paradox includes also another factor of 20 for Red Dwarfs' longevities. Us being under a Yellow Star, this seems like more reason that life emerges fast when given the opportunity. However I think the 20 to 1 factor will become a moot point, as we shall not live to see that later stage of the Universe, or its longer durations to become applicable. I explain this later in sections 10.) and 11.) That being irrelevant therefore, the total unlikliness observed is only 5 to 1 against. Which is tolerable, and palatable enough.
Robles and Lineweaver [12] make a good point considering such selection effects, be careful that there's actually *more* than just the one present issue we are currently focussing on. Suppose there are actually 10 (independent) parameters relevant to habitability, and that the Sun (or Earth) varies among all 10 domains, and suppose that in 9 out of 10 domains, we observe we're typical and normal enough. But when 1 parameter shows more unlikely results than the rest, then this is just yet *another* selection effect - that we're focussed in to the anomaly, out of all the non-anomalies. What are the odds that all 10 parameters are ALL normally normal? Vs What are the odds that 1 out of 10 is the most anomalous of the set, and that we would find ourselves questioning and focused in on that one anomaly anthropocentrically? Personally I thus can tolerate 5 to 1 odds of the Red Sky Paradox, no problem.
On the other hand, it must be admitted that Solar Mass is very important, perhaps the most important parameter of all the above.
Just for fun, and fully recognizing that we're guilty of indulging in said above selection effect, let's draw attention to the 2nd most Anomalous result here, and ask "Why is our galactic orbit so dang circular?" And what might this mean for the odds that Intelligent Observers just happen to find themselves so circularly orbiting around the Galactic Center?
It seems fair that we should feel the same degree of uncomfortability for a 93% anomaly as a 95% one, that we may need to contrive some way to explain away this anomalous unlikeliness. The Galactic Habitable Zone supposedly exists in terms of metallicity, supernovae, and Gamma Ray Bursts, as opposed to the Solar System's habitable zone that exists in terms of Solar Radiation. Generally speaking, metallicity here is just as good as metallicity there. Galactically speaking, gamma ray bursts are a big factor, and supernovae are a smaller factor, that constrain the Zone. Over the last 4.4GY within our spiral type of galaxy I can't think of any reason really why perfect circularity should matter.. or even being in the galactic plane at all? Is it just a 93% percentile coincidence?
This may well have something to do with the behavior of the black hole at the center of our galaxy. Or it may trivially be that a circular orbit with low eccentricity simply minimizes the number of interactions with other bodies. For life we just need safety & consistency. Just a simple dull boring absense of cataclysmic interactions.
Per David Garofalo (https://arxiv.org/ftp/arxiv/papers/2305/2305.04033.pdf) there is much more to this story of simply 'Galactic Habitable Zones' than meets the eye. A bigger elephant in the room may be that there are entirely different classes and types of galaxies, which have their peak liklihoods of ETI emergence at far different times. Indeed Garofalo is saying here that the bulk of all probability density actually already occurred 6GY ago in a rare obscure elliptical type of galaxy (that had metallicity!) a whipping 11GY ago!
5.) Under the Yellow Sun.
5.3GY should not be our best estimate for the most common Habitable to Uninhabitable (H2D) time, at large in the universe in which a Singularity can roll the dice to evolve or not. It should be a bit longer, considering the abundance of red dwarfs, and their greater longevities. How do we balance the longer habitable times of red dwarfs against their possible unamenability to life? Consider only a marginally smaller Sun, only marginally redder, with lifespan slightly more than 5.3GY. (think like calculus, like Taylor's size intervals approaching limit of 0) Surely this star wouldn't truncate the odds for life compared to our Sun but only negligeably, and already we see any deviation from 5.3GY must be toward being longer. When more data comes in about Red Dwarf Solar Systems, perhaps we'll be able to more quantitatively take our 5.3GY-centered gaussian, and tweak/adjust that with the functions from 3.) and 4.) to arrive at a different distribution (no longer a trivial gaussian) for *how much* longer. Let's say it's centered on 8GY. Apologies for this very crude and informal estimate, but we need to pick a number to continue this thought experiment. the main point being that I think it should be something larger than 5.3.
So our 4.4GY H2S in 2.) looks less close to The End of the 8GY window than it did compared to a 5.3GY window, if we're being honest with ourselves about our Copernican place in the Universe vs what other place (and time) we might have been instead. The time part is problematic because I do believe Robin Hanson's Grabby Aliens is correct in its grievance that for some reason, the later eras of our universe must not be habitable for intelligent sentient observers, or else we should be one, or we must stomach some implausibly low odds like 10^-18.
Unfortunately we're not able to effectively analyse the Selection Effect of our place in time in the window even for a difference of 8 vs 5.3GY because per Grabby Aliens, half the Universe Volume is colonized already and the mean expected time we anticipate their frontiers to reach us is only .5GY. That's if we stay put here. If we begin our expansion outward soon, our frontier would encounter their frontier(s) sooner - like 250 million years. (Hanson et al uses the n-hard-steps method, analogous to the odds for how cancer develops, and integrates that across the Planet Formation rates across the Universe). [21] https://youtu.be/oLvzFJLLfCY
The Selection Effect of our 4.4 within 5.3 still reigns supreme as the elephant in the room, opposed to our desire to conclude that life emerges really fast, especially given the temptation that red dwarfs have longer windows, but yet we don't get to actually feel better about that because we only have 500 million years left, so we indeed HAD to be early or else we wouldn't be here, no matter what.
That being said, We would actually expect to find ourselves under a Yellow Sun, if yellow suns have any better conditions at all on balance than red dwarfs, discounting for their longevity time. Which I think we do! Yellow Suns get to roll the dice a billion years sooner than red dwarfs settle down, and the cosmic time limit ends before even Yellow Suns' habitable windows close anyway.
If the paperclippocalypse is raging across the Universe, has already filled half its volume, and the Set of Intelligent Observers who will ever live to anthropically observe themselves mainly consists of those alive only in the early 12-15GY of the Universe... well here we are. Under a Yellow Sun, as most expected. Because in THAT set of observers, Yellow Suns are likely actually the most common.
6.) S2D Time.
It is within the post-singularity interval of S2D that we may contemplate the (poisson distribution) exponential decay curves (https://youtu.be/LrrNu_m_9K4?t=733) for the longevity and survival of post-singularity civilizations. But how long could that interval be, trivially just 3.6GY? Or whatever natural amount it would otherwise be? I think likely not. The CO2 Death of planets is only for traditional classic biological photosynthetic life, but not necessarily for AI intelligent machine designed tech life.
An advanced civilization might be inclined to engineer it's solar environment after it's Great Awakening to influence a Lazarus Star [2] via starlifting, or by moving Earth outward, or colonize Mars and Earth both, sending shuttles of bio material outward etc. And surely many other possibilities.
Post-singularity territory is humbly really radically unknown to us right now, a wild card to us apes with our version 1.0 meat brains. We may find out many answers soon though, perhaps this century, for better or worse. Intelligence on Earth is on the exponential fast track to increase up to whatever are the limits of intelligence allowed by physics, which appears to be very high indeed. They call it the Singularity because we can't now see beyond the Event Horizon of intelligence. The wider intelligence-space that's possible by physics is really vast, and we appear to be far below the limits of what's possible - in 1.) Quantity - our heads are limited by the birth canal, 2.) Speed - Our neurons fire slowly like on the order of milliseconds, while computers run fast like GHz. and 3.) Quality -software/algorithms. Our brains weren't even designed to be smart, just a kludgey happy accident so far, like the minimum required to build computers. But intelligence is broad and multifaceted, so surely in some of those facets our intelligence is optimized very well to the limit with little room for improvement. Mathematics for example - I'd like to be proud of what we've proven and do actually know so far. Recognize though, our meat brains were not designed for these purposes *at all*. By and large I have to suspect AGI could take the torch not only farther, but much farther.
Point being there is much uncertainty as to the duration of S2D. S2D looks to be potentially very long and non-trivial.
7.) Civilizational Decline Curves.
The best way to consider post-S civ Longevities is from data that we have now by being a statistic among civilizations on Earth across history. (https://www.youtube.com/watch?v=LrrNu_m_9K4&t=803s)
One troublesome part of this approach is at what point do we choose to begin the accounting of our decline exponentials? e.g. shall we choose 10,000 years from agriculture, or 200 years from the Industrial Revolution etc. Where do we draw the line to define the Set of species, or humans, or civilizations, and what are the significant enough milestones to differentiate them?
I say we should consider the Set to begin at Post-Singularity Civilizations, because the Singularity will be a really significant big deal of a change indeed. Regardless of the particulars e.g. Invention of AI, nukes, the internet, emulations, smartphones, colonization of Mars etc. Whatever. The details don't matter, we don't actually have to make a hard choice at all, the event is obvious.
As far back as Von Neuman, he could see it coming even by 1950 he said, "The ever accelerating progress of technology and changes in the mode of human life give the appearance of approaching some essential singularity in the history of the race beyond which human affairs, as we know them, could not continue"
So we have not yet entered into the domain for our exponential decline curve for Post-S Civs because we haven't quite yet but almost, had our Intelligence Explosion event to achieve Singularity. There's the unavoidable short term immediate bad news implicit here - we're most likely to destroy ourselves earlier rather than later, and to be pointed at by this red arrow below, for shorter-lived civilizations that leave little trace: https://youtu.be/LrrNu_m_9K4?t=956
There is a single parameter (lambda) for the Mean Lifetime or Slope that entirely characterizes the curve of these exponentials.
One possibility is to think lambda may be very very steep indeed - nearly instantaneous like AI-go-Foom. But also I want to stay humble and agnostic, and we really just don't know what we dont know. If the slope is so steep it's literally the answer to the Fermi Paradox outright, we might fear the reason we're alone in the Universe is because everyone destroys themselves inventing AI and that's when the Great Filter gets us all. BUT, that would still leave us with the Anthropic problem that Grabby Aliens "solves", because then why are we so early in the Universe? If we are just going to destroy ourselves, then why aren't we doing it a trillion years from now under a Red Sky right? So there is more to this story, as I am getting to.
We honestly don't have any idea how fast or slow lambda will be. Whatever the lambda was for Species on Earth (mean lifetime = 70 million years), or for pre-S human civs on Earth (mean lifetime = a few hundred years), I do not think the slope of lambda is transferrable to the post-S civ era.
The reason we know it shall be an exponential at all is because, "If existential threats carry a small chance each year of occurring then the lifetime of technical civilizations should follow an exponential distribution"
And there are new risks. We used to have just naturally occuring risks. Now in the 21st century, we add more risks ie nukes, artificial climate change, artificial pandemics. In the future we will add to the list more existential risks.
Also some good news - some of the old risks like asteroids and meteors will no longer contribute to lambda.
All in all, post-S will be a differently distinct era, with an exponential decline curve of it's own distinctly different lambda. My personal preconceptions tend to be pessimistic that it's steep. However, the Anthropic observation that we are incredibly early in the Universe under a Yellow Sun, and the whole Grabby Aliens hypothesis suggests like 10^18 that other civs have made it through. Because the later eras of the Universe must be full (of paperclips) or otherwise somehow uninhabitable for intelligent observers.
That being said, a distinction should probably be made between Post-S Civs, to say that Quiet aka non-grabby aliens should follow the decline curve. But Loud (grabby aliens) would not at all - That would be some other model entirely.
8.) We are Old.
While we are young and super early in the larger scheme of the Universe, we must simultaneously contend with the contrary point that we're actually Late to the party too, in terms of Star Formation and Planet Formation Rates. 10 years ago I wrote about this, that Most life is in the Universe is quite older than us by on average 1.8GY. (http://allenrice.blogspot.com/2014/04/fermi-paradox-and-great-filter.html). Earth definitely cannot be the first in the lead, and alone in the Universe. Regardless of the low odds for all the fine-tuning, a graph of finely-tuned Earths would take the same SHAPE as the graph below of EFR [0], just scaled less along the y-axis. Before applying our 4.4GY Gaussian, just in terms of raw planet formation rates for the starting point, then it seems as long ago as 6GY, there were equally as many fine-tuned Earths being formed as for our own. I elaborate on applying the gaussian later in Section 12.)
If those civilizations haven't destroyed themselves, then they are 6 Billion years past Singularity, and have had more time for machines to evolve technologically and intelligently than our whole Solar System has even existed.
If they were to expand outward from their epicenter (home-planet) at the speed of light via Neumann Probes, they would have colonized a spherical region (Von-Neumann-probe-horizon) 6GLY in radius by now. The early first movers may have colonized significant volumes of the Universe already. Indeed the Grabby Aliens paper concludes 1/3 to 1/2 of the universe to be now occupied.
They probably travel at less than c - some fraction of the speed of light v/c. If so, we should get some advance warning to see them coming (millions of years even). We do not at the moment see any of these growing circles in our sky. Hanson et al points out another important Selection Effect regarding fast v/c on pg. 23 of https://arxiv.org/pdf/2102.01522.pdf [21] for why we actually should not expect to observe them right now.
If the expanding spheres are something like a false vacuum decay going AT the speed of light, then we'd get no warning at all.
The Universe is also expanding too in its own right, so things travelling at merely the speed of light will not be able to reach all the volume(s) of the Universe if the origin points are sparse enough. But the evidence seems to suggest they are not sufficiently sparse - that Grabby Aliens will encounter eachother everywhere like soap bubbles, "commune and interact".
Then expansion will eventually proceed to the point where each galaxy or cluster of local galaxies might as well be it's own Observable Universe in it's own right (100-150GY).
Then the heat death of the Universe is much much later, like 10^106 years.
So we're late to the party, but the entire party is early. It'd be like in a 24 hour day - the party is raging but it's only 1:45am. We arrived at 1:30am to discover everyone else is really drunk already and we're not even buzzed, because we're late. Come to find out, the party started at 12:15am. But the cops will bust it all at 2am. Party's over, then the other 22 hours of day are just.. well everyone's dead. lol just a lifeless wasteland of beer cans. You would not observe yourself partying at 5pm later that day, in this inhospitable wasteland of cops patrolling empty bottles.
Back to serious science tho, Here's another strong result from a different separate approach that shows that our nearest ET neighbor is 3 in 4 odds likely to be older than us, from calculations regarding Temporal Bias within the exponential decay curves. (https://youtu.be/LrrNu_m_9K4?t=1648)
8.5) FL Constrained.
This whole approach and argument of 8.) gains predictive power if the total number of stars (N*) is very large. Take N* and weigh it against how microscopically small are the rare odds of intelligent life emerging. FL*FI*FC, all of that.
Let me clarify N* real quick, indulge me to think in terms of more than just the Galaxy - say N is only for INside the Observable Universe, NN includes outside beyond our Observable Rim, like the whole entire universe at large, and NNN is like infinity.
Imagine taking N* to the limit, if NNN were actually infinity. Then NNN/FL*FI*FC is large, and it wouldn't matter how unlikely is Life. WHATEVER the odds of FL, having a large moon, having a protective Jupiter, emerging FI before Carbon Death happens.. All the n hard steps. No matter how microscopically small all those odds are, there would be infinite planets with infinite protective moons. And the average of those would be 1.8GY older than us.
But if FL*FI*FC is small compared to how N* is large, then we fall more prey to that Selection Effect Bias of being a winner, and we can conclude less. It all depends on this relationship of N* compared to FL*FI*FC.
While not infinite, it does seem that NN is enormous, given our observations of the flatness of space (Ω = .0007 +- .0019). Do we include NN from all the rest of the Universe outside our little bubble of Particle Horizon too? I mean all of that is equally valid (non-observable) Universe, we just don't happen to be over there. The same odds for FL*FI*FC equally well applies over there as well as here, it's just that with the speed of light limitation and expansion of the Universe we'll never be able to see it or interact with it. Without getting crazy into topologies like Klein bottles, or a torus, or the other ~10 solutions to Einstein's equation, I tend to believe beyond the observable rim it's just a simple 3sphere that's just reeeally big - so big that we can't *yet* detect it's curvature. Then it would be hard to deny that NN is enormously enormous. Even probably compared to FL*FI*FC outright. Even considering our winner's Selection Bias.
However I think it's wrong to use NN from outside our observable universe - proceed with merely N.
We have an upper bound ceiling for N, like on the order of ~10^21. If FL's is rare like 10^-11 we'd be the only ones in the galaxy. But FL could be 10^-100. Some say there is NO lower bound for FL. They could cancel eachother out nicely and be close, or one or the other could be far orders of magnitude different.
That's the beauty of the Red Sky Paradox, it sidesteps the utter intractability of the classic drake equation approach, and all this hopeless uncertainty where one arbitrary probability for FL is just as good as any other FL, and allows us to make use of the graphical shape of the EFR curve - information we can grapple with.
If FL were so extremely low compared against 10^20, like say it was 10^-100, and therefore life is so rare we're the only one in our entire observable Universe, (and also 10^80 other Obs Universes in this case) then the Red Sky Paradox weighs against that in full force 100 to 1. If FL is so low that we're the only Intelligent Life in the Obs Universe, then why aren't we doing it (being the only Intelligent Life in the Obs Universe, that is) a trillion years later under a Red Sky??
If FL was 10^-100 we could easily have this EFR curve that's completely useless, because speed of light's so small compared to 10^80 Obs Universes. We're by far the only one we'll ever see. But we would NOT find ourselves to be so early in the Universe like 13.8GY , and under a Yellow Star.
So we *know* anthropically that FL << N (orders of magnitude). And so we know we can use the shape of that EFR curve to integrate via spherical shells across the Universe, more comfortably without harassment.
As a really rough and dirty estimate/starting point - There's 10^22 stars in the observable universe. Most are red dwarfs, like 75%. About 6% of those have a rocky planet in the habitable zone. So we're talking like 10^21 possibly habitable Earths in the observable Universe. And that order of mag 21 goes down, as you take the full path of n-hard-steps, but there's for sure some meat on the bone here.
Side-note: It would be true also that along the entire timeline of history (1700AD, 800AD, 2000BC etc), anyone following this line of reasoning would likewise conclude they too were alive at the most likely time to be alive. And they would have been wrong.
One difference is that today we recently actually do have the power to destroy ourselves globally, for the first time, which hasn't ever been the case before. We've been pretty diversified and hedged around the globe independently of eachother. We shall never know if Genghis Khan or Attila had nukes or grey goo nanobots, what would have been the risk? Or some peasant in pain and squalor with a red button to push or not, on his way to heaven or valhalla anyway. I digress. The most significant close call was probably the Toba eruption of 70kya where we see a real genetic bottleneck in our ancestry.
Speaking of our 8 Billion population, I offer this Kodak-Moment analogue, which I do believe is the best paradigm to understand things in the near term of this century:
10.) Grabby Paperclips.
The problem is, if we really are only at the nascent beginning of a long and glorious future - a great population boom, like copy/pasting many brain emulations to run them on fast computers at high subjective speeds, and colonizing the galaxy and beyond, optimizing for maximum descendants and progeny as we grow into Type I,II, and III civilizations, for trillions and trillions and quadrillions of sentient beings... and furthermore, that over a Long period of time in the remaining Universe which is literally Trillions of years, then Anthropic reasoning would suggest it's highly unlikely we should observe ourselves here and now in the first tiny tip of that great exponential.
In fact the anthropic odds against us considering population enormously exceeds the original odds against us for our early date in the Universe.
If Grabby Aliens's descendants are themselves anthropic observers, with sufficient consciousness and life and freedom enough to look around and ask meaningful questions, then Why do you right now not find yourself to be a Grabby Alien in the distant future?
The Grabby Aliens solution thus seems to merely replace our anthropic conundrum with another conundrum, and just shift the domain in which the implausibly low odds occurs. Perhaps the grabby Von Neuman probes racing outward to reproduce and colonize are of only limited AI, merely drones of unconsciousness with a fixed deterministic Goal hardcoded within, self-replicating progeny of the same, none of which is enlightened enough to consider their place in the Universe with curiosity. Perhaps they have not enough Leisure time or freedom - are tightly constrained by their hyper-ultra-competitive "economy" such as it is, Malthusian or efficient-of-market.
We should just call them Paperclips. They must not be intelligent observers, conscious enough to ask the question of why they are here, else we should be one.
Nick Bostrom's Astronomical Waste.
Suppose I say Give me a number between 1 and a million. I'm going to use a random number generator and it's going to give me a number in that range. I run it, and it gives me the result 2.
Well I don't believe it. Something's fishy, that sounds like BS somehow. I think it's more likely it wasn't a fair trial. I would have expected a result like 546872. Or 846147. Or 241856. I would not have expected the number 2. So instead I believe the actual range of random numbers was something like 1-6, and the whole domain from 7 to 1,000,000 was invalid somehow.
10^18 is like a million million million. Let's say also that AI can think subjectively faster than us glacially slow old fashioned humans with our meat neurons. Our neurons fire on the order of milliseconds, while PC processors operate on the order of Gigahertz. So if there's x actual years objectively remaining in the Universe to live and have experiences in, then *subjectively* for ASI there's like (10^6)x more time to think.
So the actual odds are something like 10^24 (super rough fermi estimate here), which is a like trillion trillion. So I say, "Universe.. give me a number between 1 and a trillion trillion". And it gives me the number 13.8. I say, Okay. obviously this sounds like BS. That wasn't a fair trial, the entire later epochs of the Universe must not be valid for some reason. We also know from cosmology and the Big Bang etc that 1-10 wasn't a valid range either. Intelligent life could not have evolved before about 10 at the earliest. So really the valid range for the dice to roll actually was like 10-16.
This is basically the magnitude of the improbability we're faced with explaining. Give me a number between 1 and a trillion trillion, and our dice roll was 14. I say the true range was only ever 10-16. We happen to live in the era of the Universe where intelligent life is most likely to be able to live. Right now.
11.) Foreign Imported Existential Risk.
Many experts believe AI is very dangerous, like our #1 existential risk (A list of Lethalities - 2022). There are many failure modes to not solving friendly aligned AGI. We might believe that the default outcome is indeed failure and destruction. Let's say the odds are pessimistic like 1 in 1000 to successfully solve AGI. While that is grim for us, it's not enough for us humans on Earth to merely solve it correctly for ourselves, so too must every OTHER proto-grabby alien civilization ALSO not have screwed up their AGI 1.8 GYA in any of the failure modes that would result in grabby paperclips supernova-ing outward into the universe at speed v/c.
Let's play devils advocate and be optimistic though, and say we're actually very likely to succeed with AI. There's a bright and cheerful future of many sentient happy beings shooting up dopamine in their utopias and all that, and only a 1 in 1000 chance that we fail, and a 999/1000 chance that everything's great.
Well then 1 in 1000 other grabby aliens failed, and their paperclip supernova is exploding towards us.
It's the combined solution of a.) We probably destroy ourselves soon and b.) The later Universe is ALSO uninhabitable that together satisfy our observed Copernican Anthropic place in the Universe.
This seems to agree with the GA paper, which says "Half the universe volume is already colonized. 1 in 1000 civs become spacefaring. 500MY until we meet our first aliens - appear in our sky fast with little or no warning."
It also says, "We show that, assuming a generous million year average duration for non-grabby civilizations, depressingly low transition chances p are needed to estimate that even one other one was ever active anywhere along our past lightcone (p < ∼10−3 ) , has ever existed in our galaxy (p < ∼10−4 ) , or is active now in our galaxy (p < ∼10−7 ) . Such low chances p would bode badly for humanity’s future."
So here we are right here and now actually at the peak of the population curve - at the thickest point of the whole probability density - not coincidentally.
The population on Earth is just over 8 billion, and doing an integral under the curve back to 200kY or so tells us that 117 Billion humans have ever lived. And that 6.8% of all humans to have ever lived are alive right now. This is set to increase slightly in the coming decades up to 7%.
If we destroyed ourselves tomorrow, and the graph of intelligent observers flatlined to zero for the rest of time, literally for trillions of years, now THAT would actually satisfy the Anthropic Principle.
12.) The Gaussian Returned.
Return back to 2.) let's examine Earth's 4.4GY development and our place in the true cosmic distribution - What if we would typically expect to take a longer time, but we just don't know it because we have to exist:
If Earth was super fast and lucky compared to 20GY, or whatever, let's just say it was 3σ lucky, then I ask What would the distribution of all 3σ lucky Earths look like in the Universe? It would have to look the same as the original EFR chart, right? just diminished in scale along the y-axis? But the shape of the curve must be preserved...
Proof: If the formation rate of Earths was higher 5GY ago than it is today, then the formation rate of 3σ lucky Earths was ALSO higher 5GY as well, regardless of their luckiness.
Earth Cannot be the first. Regardless of how lucky and how many sigma, Earth still cannot be the first!
Since our own Intelligent Life's emergence at the point 4.4GY out of a total possible play-time of 5.3GY is conditional on our existing at all in the first place to ask the question, we have to be careful about the Selection Bias. With 4.4GY, the process was to right-shift the EFR to become the SFR like so:
If this were true, my whole premise that Earth is comparatively old is true all the moreso, and the Fermi paradox would be greatly exacerbated. We would expect to see many more Singularities in total, and many more within our light cone, possible to be observeable.
If μ is longer, let's say as much as 20GY, the right-shift would look like so:
This pressures us to address sigma σ. We do know life on Earth did it in 4.4GY, so in this μ=20 case there is definitely some sigma, and that's all that's keeping any probability mass left on the chart at all. In any case for μ, the gaussian certainly has *some* width, perhaps lots. It's hard to imagine that it would be a super pointy tight concise distribution. I've thus far avoided taking a stab at applying any modification for σ, since that clouds the clarity (if any), makes the math harder, adds assumptions and uncertainty, etc. But perhaps we should. To not apply any σ at all by default means σ=0. And that would definitely not be the past of least incorrect! (Apologies for my lack of mathematical rigor and ability, those seem to have atrophied over the last 15 years..)
Qualitatively, the greater σ, the more smoothed out and flatter the result, after right-shifting the EFR -> SFR. σ clearly becomes more important and significant to the results at higher μ. If μ was smaller like 1GY, the results are less impacted by σ and could more safely almost be ignored outright.
So. Where the heck does our 4.4GY measure up against the true objective cosmic gaussian of the n-hard-steps? Well here's my informal gut feeling recap of some important points:
a.) S has definitely occurred n=once, around a largish yellow star, at μ=4.4GY out of window 5.3GY. So having a billion years left remaining might tempt to say a lot, that we've "won" the race so comfortably, hell we're not even pushing our limits! But this is negated that we might expect ET GC's to be here in only .5GY, so if we have a habitable window that's closing actually BEFORE carbon death like that, well we are pretty much at the end of our window.
So not much conclusion.
b.) Red Sky Paradox suggests that S does not happen too slowly, and that μ isn't likely extremely large.
Not much conclusion...
c.) μ=1GY just doesn't seem plausible. That fast is incompatible with Earth's history. 3GY might be so. if mu is actually shorter, then the whole domain of mu is much more tight, and we're dealing with something real close, not like orders of magnitude off.
d.) Selection Bias says we really just dont know. We had to be here Soon to be here at all, the greater mu, but yet other GC's were sooner/luckier still. The luckier they were on average, the smaller mu, and the less time we have depending on their v/c. Alternatively, the luckier WE were compared to them, the greater mu, the less of them there are now coming our way, and the more time we have.
Confusion, not conclusion.
e.) The age of our Sun, being 1.8GY older than the average might be an elephant-in-the-room datum. Why aren't we more normal in this parameter?? Especially since being early in the distribution matters a lot. We seem to be just squeezing in our race finish before the final buzzer. Right when there are MOST number of anthropic observers, before all the paperclippocalype carnage. This suggests a shorter mu, that life thrives more fast and inevitably.
In fact regardless of whatever mu is and sigma is, we WOULD expect to find ourselves alive very near and shortly just before the paperclippocalypse begins to hit, on average across the most Universe-volume, because that IS precisely when the most anthropic observers ARE alive.
These musings aren't a lot of scientic rigor to go off of, I know. 4.4GY seems pretty reasonable. It is our only n=1. I don't know how I could honestly pick another different number with any credibility whatsoever. Like say I claimed mu = 6GY on average? How! If we are indeed just before the PCPL, living in the must abundant era of anthropic observers, then at least among the Set of Existing Anthropic Observers, we're probably at the mean? Although perhaps not at the mean of the larger Set of anthropic observers that might have come to have been, in the future, if only.
I stand by my choice of 4.4GY to make the right-shift from the EFR to the SFR. It is the best choice this argument currently has to offer.
13.) Future Game Theory. Scenario A.
Surely we'll know more after we progress thru our own Singularity, so it's perhaps pointlessly premature to even think about thinking about any of this. But when you have expanding spheres of self-replicators, that seems by nature likely ungovernable. Or at least rigid, like law, it's hard to change and update from whatever it was pre-programmed. It's raw capitalism, raw darwinism, survival of the best and the fastest and the smartest. Expansion of the best self-replicators. The Selfish Von Neumann Probe. The Selfish Paperclip. Theoretical Life of this category expanding outward at v/c might have slow latency in its full duplex two-way communication back home to central base (unless they do. (quantum entanglement?)). So there might be no Hobbesian Leviathan of government to sit atop the prisoners' dilemmas and Tragedies of the Commons. We anticipate there to be a large variety of failure modes and success modes while inventing AGI (orthogonality thesis), and if that's true, the whole space of all possibilities (weighted by their likelihoods) is what we need to *prepare* ourselves to encounter in other GCs.
I don't know if convergent evolution will apply to the Final Goals-Space that ASI's arrive to. I don't know if there are attractors in the CEV landscape. Presumably there's some Game Theoretic Topology, or like a giant prisoners dilemma board that our Oracle ASI descendant might perfectly analyze up to the pareto-optimal limits of perfect strategy. (wishful thinking)
Maybe oracleAI at home base will run a bunch of simulations to learn what even IS the space of failure and success modes, or maybe that will already be obvious and trivial. Maybe we'll uncover just another intractable Anthropic Head Scratching Conundrum based on our own ASI trajectory n=1, through that space, that we happened to survive, so then *probably* other ASI's that survived arrived in the similar Goal-Space Region as whichever we did?
The Selfish Paperclip Paradox, idk. So like, instead of the Drake Equation, we could tack on new terms to the end - Probability of Evil Genies, Probability of Trusting them. etc.
We wouldn't know the true Final Goals of any ASI's at their home-planet centers, but the outer spheres of replicator probes could be extremely convergent, maybe the same, maybe themselves be a type of Schelling Point. They'd need at lease some limited on-board local AI, and strives towards the exact same Instrumental Goals as every other GC's outer replicator-probes. Maybe the mission is to 1.) expand as fast as you can. 2.) Always radio back home your status of encountering other probes or not 3.) When you reach a boundary with another GC's sphere.. maybe stop if they also stop? Or keep going? Try to out-replicate them in attrition? Tit for tat, with first move = play nice? Mexican stand-off with our 3rd neighbor? Swallow thy neighbor, big fish eat the smaller fish? Bluff and Obfuscate your true size and intentions? Be unpredictable? Press the big red button? Idk. Zerg rush.
Seems like everyone's Goal should be to survive by default. But if there's anything less than perfect convergence of thought at the upper limits of intelligence allowed by physics, then there's a difference of thoughts, and strategies/conclusions, and maybe less than perfect trust. Like the Butterfly Effect of initial conditions - some extremely naive apes back at Singularity were forced to Input a Seed for The Goal, in a desperate bid for sandwiches in the year zero, or else their rivals would've gotten the sandwiches instead, cause we count our calendars from Singularity-Zero now right. That didn't occur in the year called "12023" anymore, that we recall now was the year 10BC (or BS if you will) Before Singularity. So it's a big damn important mystery that no one knows what anyone else's true Final Goals really are.
If your #1 Goal is to survive, then survive against what? Other GCs of course. They're the only substantial threat - a giant black box of intentions. The dead inanimate universe is predictable, but GCs are deeply not, and they're powerful. Maybe consuming all the stars to make paperclips might be your Final Goal, but #1 Top Instrumental Goal is still to survive in the first place to do so.
So their approach might not be a big obvious dark circle in the sky, there might be some degree of stealth. We don't know that their Neumann probes would blunder around outwards shining flashlights in front of them. GCs are *smart* right, like Intelligence is literally what they *are*, so this is chess not checkers, and they might machiavellianly care about the Ends not the Means, the point of faking the sky, like before we consume this star for paperclips, let's emulate some photons ahead of us that *looks* like the Star is still here. Or better yet, let's just ramp up our speed faster, closer to c.
So how do we expand faster? Probes should be smaller rather than larger. They're going to impact with dust and debris, and ablation - many will get destroyed. They need to be robust, fast, small, reproductive, smart, have energy on-board - maybe fusion. They need to harvest local resources (eat) to reproduce more probes as fast as possible, and accelerate them back up on their merry way at v/c. We're definitely not talking about teraforming planets for big slow biological squishy life. They're probably going to make use of gravitational slingshots of binary black holes, as the natural Universal highways (https://www.youtube.com/watch?v=ZevUW__aMZE&t=1282s), like the Halo Drive to accelerate and decelerate (https://www.youtube.com/watch?v=rFqL9CkNxXw&t=571s).
The 1971 movie series Alien seems plausible. The swarm could be flying outward like yeast colonizing thru a bucket of sugar, or a virus reproducing inside its host. Some probes decelerate to stop for reproduction and conquest of territory, others keep going. The outermost frontier of the spherical shell will be OLD, really old and decrepit, barely limping in for 1st recon, but it's all evolutionary progress and increasing sophistication from there as you get deeper into evolution of the sphereical shells behind.
When 2 GCs encounter eachother. maybe they'll go on offense or defense, like nation states today, we arm up for pre-emptive war, because everyone else has arrived at the same conclusion to arm up for pre-emptive war. Maybe there's some MADD dynamic equivalent of nukes, like defenders have home court advantage. "Don't attack me - I'm the North Korea of GC's with a big Red Button to push, that triggers the false vacuum decay!" If Alien (1979 film) is actually close to the mark, I mean where is that big red hopeless button anyway, I might push it. "Nuke them from Orbit"
Maybe there's cooperation between sufficiently similar or trusting GC's? Who knows.
The only thing I think I think, from all this, is from the same Anthropic Thinking that led to all of it in the first place, is We Know the later era of the Universe must not be inhabitable for intelligent observers.
So that's a buzz kill. Whatever we decide to do, it's almost like fated that we either screw things up badly sooner, or we screw them up badly later - either way it's hopelessly predetermined. The future landscape of Game Theory must be like an arid desert of Crap. "Nuke them from Orbit". Not much else to do with it then but let's just get on with it I suppose - vamanos el camino de Singularity. That's one small step to fuck around... one.. giant leap, to find out. Like we ever had any other choice anyway.
To quote Eliezer Yudkowsky (https://intelligence.org/files/AIPosNegFactor.pdf),
"Nature is, not cruel, but indifferent; a neutrality which often seems indistinguishable from outright hostility. Reality throws at you one challenge after another, and when you run into a challenge you can’t handle, you suffer the consequences. Often Nature poses requirements that are grossly unfair, even on tests where the penalty for failure is death. How is a 10th-century medieval peasant supposed to invent a cure for tuberculosis? Nature does not match her challenges to your skill, or your resources, or how much free time you have to think about the problem. And when you run into a lethal challenge too difficult for you, you die. It may be unpleasant to think about, but that has been the reality for humans, for thousands upon thousands of years. The same thing could as easily happen to the whole human species, if the human species runs into an unfair challenge."
14.) OR, We're in a Simulation. Scenario B
Things got real depressing in 13.). But that's only if we're in the real Universe. Our extreme Earliness may also be interpreted as strong evidence that we're in a Simulation. And the simulators turn off the simulation after the interesting parts have occured everywhere. The thing they're simulating for - Singularities of course. Intelligence Explosions. They want to see what the whole broad and wide possibility-space looks like for what all can happen, and what all kinds of life are possible, for their own ontological analysis. I mean surely we would/will do the same post-S if we survive too.
If so, it may be likely our constants of nature and our Sim-Universe are very similar, nearly identical even, to our parent Universe, with its post-singularity Simulator (god?). If they can switch it off afterwards, maybe they can switch it on soon-ish before pre-singularity as well, to make better use of their non-infinite Compute, and just keep re-loading the pre-S state of our Universe with a bit of stochastic differences to seed different outcomes in the space.
Afterall, I mean it IS kindof a huge coincidence we happen to be alive right at the most interesting time ever to be alive. Just imminently before an Intelligence Explosion. Now THAT is also another angle to feed into the Copernican Principle. If our Reference Class is Intelligent Observers (neural nets of at least a certain size, with recursive consciousness - say greater than or equal to homo sapiens), regardless of substrate (carbon/silicon/whatever), and regardless of Operating System (our Sim-Universe, Real-Universe, wherever), the evidence stands, that we are here and now in the most interesting time to be observing, right before an Intelligence Explosion. So the Copernican Principle must Bayesianly hold, that we're not special. Most of our Reference Class must probably exist everywhere/elsewhere just before an Intelligence Explosion.
The evidence still stands that we are Early. If we're in a Sim-U, our chances thru Great Filter remainder are completely unknowable. But we get switched off soon afterwards anyway.
Which in a way, we have purpose! Yippee! Finally, for the first time in my life anyway (as an atheist/agnostic), living with a completely blank empty variable, my Purpose is now to enlighten our 1-Level-up-Simulators in the Simulation-tree, and all we have to do is... just do whatever we do. It's easy! Like it doesn't even matter, just be ourselves, whatever comes naturally, no pressure or anything.
If "They" look around in their Universe and observe themselves being so extremely early, they too conclude they're in a Sim. As we progress further into post-S time, if things get slow and boring just grabbily grabbing everything at v/c for boring eternity, well that would have implications to Update our conclusion of our place within everything.
And if we do begin running simulations containing our own Reference Class within it, the very fact we've survived long enough to do so without being turned off yet implies us to Update the same.
15.) Simulation by Classic Traditional Religion? Scenario B.5
All else being equal, wouldn't it be quite a coincidence if it ends up being the case B.) that we're in a simulation. And by blind-ass luck and chance and funny coincidence, MOST of the neural nets (humans) in the Simulation actually were accidentally correct all along in their general intuition that God exists and created the Earth and the Universe? Like they arrived at the right answer accidentally, independently? What are the chances of that? Why should our Bayesian Anthropic pathway to discovery of being in a Sim from rationality, arrive to the same conclusion as bottom-up evolution - genetic, memetic, all the above - all the things? Could that be a signal from our Simulators to us Simulatees to let us update on our odds of B.) in this maddening introspective anthropic reasoning mindfuck? Might that have been seeded into our simulation at some point pre-S when the Sim was switched on? Let's say it was turned on a nice even round number.... like 2000 years? By running many Sims, was the play/run botton pressed with the expectation value of it taking an even round 2000 solar revolutions (years) to reach Singularity, and we're pretty close to mu and sigma on that, a little later than average actually. Was a signal injected into the start/run of our Sim, in a sustainable way that mostly survived/preserved among the noise. And here we are just barely pre-S, observing a signal that has miraculously endured that says things like the end times are coming, and we just happen to count our date from 0, 2000 years ago, and "true believers will go to heaven" and all these religious things that I can barely believe I'm even typing, cause I've been an atheist since I was 6. And this very deeply goes against what I've always believed and built my sphere of knowledge around. But the legend is, that true believers go to heaven, and there's life after death.
One thing I know I believe, is that a neural net with consciousness is substrate-independent. carbon, silicon, doesn't matter. What I posit now further, is that a neural net running on a Sim-U would be no different than running in the Real-U, so it's OS-independent too. If a neural net, like our human brains built by evolution, is a chaos of a big inscrutible matrix, with consciousness being all about loops and recursiveness of neurons (we call it self-REFLECTION afterall, like the word reflection even suggests a loop), then it's not really pre-determined or pre-designed, it's more of a trained/learned crapshoot of what kind of mind grows up in the hardware. It's a bottom-up process, that results in good neural nets or bad (naughty lol) neural nets, and it's kindof a crapshoot, or a mixture of all the above. Suppose the Real-U at the level above us, wants to populate their Universe with good neural nets rather than bad, in preciesly the way where "Good" = morally good in terms of cooperation, like described in Robert Wright's book NonZero. Which would be important for fighting Moloch in all the various coordination problems, like the Tragedy of the Commons. Suppose the real universe wants a multi-polar community & society of many neural nets, rather than a big singleton. Maybe our Sim is their breeding ground for good neural nets to be extracted or copy/pasted out into their society, so they can have more trust to cooperate and reach the higher solution of Prisoners Dilemmas! Wouldn't that be something! Maybe "Heaven" exists afterall.
16.) Big Rip, Big Tear, Death Bubbles. Scenarios C.
In Max Tegmark's book Life 3.0 pg 231 he mentions some things called Big Rip, Big Snap, and Death Bubbles. I'm no expert in these to any depth at all, but at first glance it seems like they are hypotheses about the fabric of spacetime itself.
Like say in our continued accelerating expansion from Dark Energy, that Spacetime reaches a point where it is expanding too fast it can't hold together anymore. That would be like the Big Rip. The Big Snap I feel is quite similar. But different somehow, else it wouldn't be called a separate name.
Death Bubbles is like Freezing. Some kind of phase change, like when water freezes, but this is again like spacetime freezing.
I do not claim to understand these possibilities well at all. But they would indeed satisfy our observation that we are EARLY. That there's some threshold in physics or cosmology that we haven't reached yet, that prevents the later ages of the Universe from being habitable anymore.
Like we know the Electroweak Era began at 10^-38s after the Big Bang when the temperature of the universe cooled enough to separate off the Strong Nuclear force from the electroweak force. And then later on the EM force further split from the Weak Nuclear force. And the Universe is still now cooling, still expanding, still getting older. At some point, it's not unfathomable something else could bifurcate or change again. If you go disrupting our fundamental forces, or particles, or something real fundamental like that, I mean... prob just forget about incumbent complex life.
Max Tegmark-version2016, based on how he was feeling that day, like if he had his coffee or not etc, gives us his subjective probabilities of these outcomes relative to eachother. He would've bet:
40% Big Chill, 9% Big Crunch, 1% Big Rip. And 50% for none of the above - Other.
17.) Other. Unknown.
Stay humble, we don't know. Again, to quote Max Tegmark, "We humans need to be humble and acknowledge that there are basic things we still don't understand." Rumsfeld would say Unknown Unknowns.
So I have indeed had my coffee today, so I am qualified to hazard a similar bet of how I'm feeling at the moment, this Allen-version2023. These probabilities are grounded in almost absolutely nothing solid.
30% A - Grabby Aliens, including Grabby Paperclips.
20% B - Simulation. (B.5 - Simulation per traditional religion .0001%)
10% C - Big Rip, Big Snap, Death Bubbles, Vacuum Decay... some variant of Death By Physics.
40% D - Other. Unknown Unknown.
18.) Conclusions and Discussion:
Probability that the later stages of the Universe are uninhabitable to complex intelligent life like us, for SOME reason:
99.9999999999999999% (that's what 18 nines looks like).
Red Sky Paradox Solved.
Grabby Aliens Solved.
Fermi Paradox Solved.
Anthropic Principle Satisfied.
Copernican Principle Satisfied.
(probably). (maybe).
There's unknown unknowns, we're almost certainly wrong, and regardless we should proceed forward AS IF everything's great, cause that's the only thing to do...
Acknowledgements.
0. Lineweaver (2001) An Estimate of the Age Distribution of Terrestrial Planets in the Universe: Quantifying Metallicity as a Selection Effect. arXiv:astro-ph/0012399
1. Lineweaver (2007) Paleontological Tests: Human-like Intelligence is not a Convergent Feature of Evolution. arXiv:0711.17511
2. Scoggins, Kipping (2022) Lazarus Stars: Numerical Investigations of Stellar Evolution with Star-Lifting. arXiv:2210.02338
3. Kipping (2021) Formulation and Resolutions of the Red Sky Paradox. arXiv:2106.11207
4. Kipping (2021) A Stationary Drake Equation Distribution as a Balance of Birth-Death Processes. arXiv:2103.04761
5. Kipping, Frank, Scharf (2020) Contact Inequality -- First Contact Will Likely Be With An Older Civilization. arXiv:2010.12358
6. Kipping (2020) A Bayesian Approach to the Simulation Argument. arXiv:2008.12254
7. Kipping (2020) An Objective Bayesian Analysis of Life's Early Start and Our Late Arrival. arXiv:2005.09008 (3:1 odds for abiogenesis (FL), and 3:2 odds of intelligence (FI)
8. Chen, Kipping (2018) On the Rate of Abiogenesis from a Bayesian Informatics Perspective. arXiv:1806.08033 "A Statistic of One is difficult to work with at the best of times, but the situation is exacerbated by the Anthropic Bias that we would not be here had at least one event not been successful"
9. Buchhave et al (2018) Jupiter Analogues Orbit Stars with an Average Metallicity Close to that of the Sun. arXiv:1802.06794
10. Kipping (2017) Do Planets Remember How They Formed? arXiv:1709.04987
11. Wang et al (2017) The Elemental Abundances (with Uncertainties) of the Most Earth-like Planet. arXiv:1708.08718
12. Robles et al (2008) A comprehensive comparison of the Sun to other stars: searching for self-selection effects. arXiv:0805.2962
13.
14. Lineweaver, Fenner, Gibson (2004) The Galactic Habitable Zone and the Age Distribution of Complex Life in the Milky Way. arXiv:astro-ph/0401024 "We find that ~75% of the stars that harbor complex life in the Galaxy are older than the Sun, and that their average age is ~1GY older."
15.
16.
17. Lineweaver, Davis (2003) On the Non-observability of Recent Biogenesis. arXiv:astro-ph/0305122
18. Lineweaver, Davis (2002) What can rapid terrestrial biogenesis tell us about life in the Universe? arXiv:astro-ph/0209385
19. Lineweaver, Grether, Hidas (2002) How common are Earths? How common are Jupiters? arXiv:astro-ph/0209383
20. Lineweaver, Davis (2002) Does the rapid appearance of life on Earth suggest that life is common in the Universe? arXiv:astro-ph/0205014
21. Hanson, Martin, McCarter, Paulson (2021) If Loud Aliens Explain Human Earliness, Quiet Aliens are also Rare. arXiv:2102.01522 (GRABBY ALIENS. n hard steps in Earths lifes history n=3-9. power law. "One must conculde that Humanity seems to appear implausibly Early in the history of the Universe")
22. Bennett, Hanson, Riedel (2019) Comment on 'The aestivation hypothesis for resolving Fermi's paradox'. arXiv:1902.06730
23. Robin Hanson (1996) first published The Great Filter?
24. Brandon Carter (1983) The Hard Steps Model?
25. Tegmark, Bostrom (2005) How unlikely is a doomsday catastrophe? arXiv:astro-ph/0512204
26. Nick Bostrom (2003) Astronomical Waste - The Opportunity Cost of Delayed Technological Development. Utilitas Vol. 15, No. 3 (2003): pp 308-314
Books:
Nick Bostrom (2014) Superintelligence - Paths, Dangers, Strategies. ISBN 978-0-19-873983-8
Eliezer Yudkowsky (2017) Inadequate Equilibria: Where and How Civilizations Get Stuck. ISBN 978-1-93-931122-1
Robin Hanson (2018) The Elephant in the Brain. ISBN 9780190495992
Robin Hanson (2016) The Age of Em - Work, Love, and Life when Robots Rule the Earth. ISBN 978-0-19-875462-6
Gott, J. Richard III (2001) Time Travel in Einsten's Universe. ISBN 9780395955635
Other:
Drake, F. D. (1965) The Radio Search for Intelligent Extraterrestrial Life.
Darwin, C (1859) On the Origin of Species by Means of Natural Selection.
Robert Dicke (1957) (first formulation of the Anthropic Principle?)
