Imagine that you came across an enormous dry grassland, continuously covered with dense grass. It seems to go on for thousands of miles in all directions, and historical records suggest that it has been in this same dry state for millions of years. You conclude that if a spark had touched it anywhere anytime during that period, a fire would have begun that would eventually spread across the entire grassland.
In this situation you either have to believe that sparks are extremely unlikely, so that for example lightning is just a very rare thing in this world. Or you have to conclude that appearances are deceiving; there are many wide barriers that limit the spread of fire in space, or there are serious defects in your historical record. Either sparks almost never happen, fire starting in one place does not spread to the entire grasslands, or fires do periodically spread everywhere but quickly burn out and then their historical records are quickly erased.
Now imagine that you came across an enormous pleasantly-wet mildly-windy barren land; it seems to be a millions-of-years stable continuum of sand that goes on for thousands of miles. You can tell from lab tests that this wet sand could serve as fertile soil. That it is, it has sufficient nutrients, water, sunlight, temperature, pressure, etc. to enable some kinds of grass seed to grow into grass plants that send out more seeds. And yet this land has apparently remained empty and barren for millions of years; it holds neither grass nor other life that might evolve from grass.
In this situation, you either have to believe that almost no grass seed has ever fallen on this land for millions of years, or that the appearance of a stable continuum of sand is seriously misleading. Perhaps that are wide strong hidden barriers to the dispersal of seeds, such as wide barrier regions of no wind. Or perhaps some big disaster happens periodically to kill basically all seeds across this entire connected land, and then later all historical records of both the prior seeds and the event that killed them are erased.
Imagine further in these situations that we the observers making these observations and drawing these conclusions are in fact made out of, or closely connected to, fire in the first case, and seeds in the second. We would then have to believe that our origins are extremely crazy rare, that we will either remain permanently isolated behind travel barriers, or that soon we will suffer a quite thorough death that erases most all records of our existence.
These imaginary scenarios seem close analogues to humanity’s actual situation in the cosmos today. Except that now we are talking a period of fourteen billion years and a scope even more billions of lightyears wide. We seem to be close to becoming part of a fire or seed that would be capable of spreading across the cosmos, burning most all or turning most all to grass, or to some descendant life. And yet our historical records seem to be good enough to tell us that no such fire has yet happened, or that almost none of it has been turned to grass or descendant life.
We must then conclude that either (A) we not remotely as close to these expansion abilities as we think, (B) the appearance of life like us is extremely rare, or we are seriously mistaken about either (C) the feasibility of long-distance travel, or (D) the absence of frequent cosmos-wide disasters that kill everything. Yes we do know of substantial obstacles to our future evolution and long-distance travel, and of periodic large disasters that would kill many things. But our best understanding is that these evolution and travel obstacles can be plausibly overcome, and that these large disasters have a quite limited scope.
Our grabby aliens analysis suggests aliens who spread across the cosmos as would a fire or grass are in fact quite rare. They appear roughly once per million galaxies, and appear in time according to a power law that emphasizes later times which we are less able to see from here now; we’ll meet them in roughly a billion years if we expand. But in this post I want to remind us of other possibilities; maybe our future evolution or long-distance travel are much harder than they seem, or maybe there are hidden disasters much more severe and frequent than we suspect. Beware.
Re "hidden disasters much more severe and frequent than we suspect":
Are prediction markets able to assess existential risks?
It isn't clear to me how a successful prediction would be paid.
On a slightly less existential question (since it would not preclude AGI-based colonization)
"Metaculus has 68% on AGI existing by 2030 but only 2% on humans
going extinct by 2100. That seems way too optimistic about AI
alignment."
( from https://astralcodexten.subs... )As prediction markets currently run, is this truly expected to be the best available estimate of the odds? Can it be improved?
That sounds about right: older civilizations might be able to prevent younger civilizations from seeing them coming, but there would be a period when younger civilizations leak their location data to any older civilizations nearby who are looking in their direction. However, it might not be possible to detect young civilizations in other galaxies at all. I think at the moment, we can only see if a whole galaxy emits biosignatures - due to noise. Younger civilizations could possibly escape being detected by older ones - by being small and distant.