Suppose that humanity wiped itself out but left behind complex multicellular life. I think there is a good chance that space-faring civilization would emerge again on Earth, and I’ve argued that it might be worthwhile to try sending them a message.
(In that post I guesstimate that if all went well you might be able to effectively reduce extinction risk by 1⁄300 by sending a message to the future; I could imagine that costing only ~$10M and leveraging interest from non-EAs, in which case it sounds like a good buy.)
Unfortunately, sending a message 500 million years into the future seems very hard. Most things that aren’t buried will get destroyed and the landscape itself is going to be completely transformed (tectonic plates will be rearranged, mountains will appear and disappear, the world will be covered in forms of life that don’t exist yet...).
H/T to Dan Kane for criticizing my original post and clarifying the actual nature of the problem.
More precisely my question is whether we can:
Spend $10M-$100M.
Encode 100MB of information.
Wait 500 million years.
Have it be found with probability >25% by a civilization as sophisticated as humanity in 1900.
I’d also be interested in relaxing any of these constraints a little bit, e.g. spending 10x more, sending 10x less data, only lasting 100M years, or only being discoverable in the 21st century.
I’m not very worried about preserving the information—I suspect that if we are willing to bury something, we can preserve 100MB pretty cheaply. My biggest concern is putting the information somewhere that our successors can find it.
My initial suggestion, clarified/improved/named by Jess Riedel, was to:
Bury a small number of expensive “payload” messages (maybe ~100)
Bury a large number of “map” messages (maybe 10,000 − 100,000).
Somehow get people to stumble across a map. Either put a lot of them in places where they might end up being easy to find, or transform the landscape in ways that might remain visible in 500M years.
Use the maps to encode the location of payloads. (With each payload also including much more detailed maps to all the other payloads.)
I now feel like both of these steps are hard:
It seems very difficult to actually put maps in places where they’ll be found. For example, my proposal of using a giant+out-of-place+slow-to-weather rock is probably very difficult given how much the landscape will change, since the rock will probably be either buried or moved.
It seems very difficult to reliably point to the location of the payload, given how drastically the world map will change. You’d probably need to combine (a) a clever way of communicating locations, with (b) some form of beacon that would be visible if people were looking for roughly ghe right thing in roughly the right place. This is further compounded by the difficulty of telling what we were saying.
I’m not sure if either of those difficulties are serious. For example, I’m not sure that the relative locations of nearby items would get scrambled too much, in which case you might be able to use local maps.
Even if those difficulties are serious, there is a huge space of possibilities and I suspect that there is something that works and is reasonably cheap:
We could potentially store messages in the ruins of prominent cities, if cities have a reasonable chance of being buried+preserved (cities have enough weird materials in them that I expect they’d leave a really visible mark). This could either be used to make small messages easier to find, a place to put payloads (which can potentially be pointed to with a map of the city), or both.
If making preserved messages (or messages with a reasonable shot at fossilization) is extremely cheap, then we could potentially send very large numbers of them. This could be used either to send a bunch of payloads and rely on redundancy, or to allow maps to be very large and expressive. It could also be used to flood the world with massive numbers of maps (>>1M), so that they can be easily found without beacons. (Really extensive flooding sounds more like a last ditch effort once we can see extinction coming, rather than something you’d do preemptively.)
There might be geologically inactive locations where you can just leave giant+out-of-place+slow-to-weather rocks and they have a reasonable probability of remaining intact. Mountain ranges form over much less than 500M years, but it’s not clear to me that the whole world churns since I don’t really know anything about geology. To do this, you’d need to find a rock that wouldn’t wear away entirely, and you’d need a location where it wouldn’t be disturbed too much or end up under ground.
There were some plausible suggestions in a Facebook thread on this topic (including putting stuff in space, defining coordinates with respect to tectonic plates)
If we could come up with a really convincing and reasonably cheap way to send a message, then I think it’s probably worth exploring this idea at least a little bit further. I think the next step would be more seriously analyzing how much good a message could potentially do (which is much more speculative than this step).
I’m in the market for certificates of impact for significant contributions to this problem (in the $100-$10k price range, depending on the size of the contribution).
Could we send a message to the distant future?
Suppose that humanity wiped itself out but left behind complex multicellular life. I think there is a good chance that space-faring civilization would emerge again on Earth, and I’ve argued that it might be worthwhile to try sending them a message.
(In that post I guesstimate that if all went well you might be able to effectively reduce extinction risk by 1⁄300 by sending a message to the future; I could imagine that costing only ~$10M and leveraging interest from non-EAs, in which case it sounds like a good buy.)
Unfortunately, sending a message 500 million years into the future seems very hard. Most things that aren’t buried will get destroyed and the landscape itself is going to be completely transformed (tectonic plates will be rearranged, mountains will appear and disappear, the world will be covered in forms of life that don’t exist yet...).
H/T to Dan Kane for criticizing my original post and clarifying the actual nature of the problem.
More precisely my question is whether we can:
Spend $10M-$100M.
Encode 100MB of information.
Wait 500 million years.
Have it be found with probability >25% by a civilization as sophisticated as humanity in 1900.
I’d also be interested in relaxing any of these constraints a little bit, e.g. spending 10x more, sending 10x less data, only lasting 100M years, or only being discoverable in the 21st century.
I’m not very worried about preserving the information—I suspect that if we are willing to bury something, we can preserve 100MB pretty cheaply. My biggest concern is putting the information somewhere that our successors can find it.
My initial suggestion, clarified/improved/named by Jess Riedel, was to:
Bury a small number of expensive “payload” messages (maybe ~100)
Bury a large number of “map” messages (maybe 10,000 − 100,000).
Somehow get people to stumble across a map. Either put a lot of them in places where they might end up being easy to find, or transform the landscape in ways that might remain visible in 500M years.
Use the maps to encode the location of payloads. (With each payload also including much more detailed maps to all the other payloads.)
I now feel like both of these steps are hard:
It seems very difficult to actually put maps in places where they’ll be found. For example, my proposal of using a giant+out-of-place+slow-to-weather rock is probably very difficult given how much the landscape will change, since the rock will probably be either buried or moved.
It seems very difficult to reliably point to the location of the payload, given how drastically the world map will change. You’d probably need to combine (a) a clever way of communicating locations, with (b) some form of beacon that would be visible if people were looking for roughly ghe right thing in roughly the right place. This is further compounded by the difficulty of telling what we were saying.
I’m not sure if either of those difficulties are serious. For example, I’m not sure that the relative locations of nearby items would get scrambled too much, in which case you might be able to use local maps.
Even if those difficulties are serious, there is a huge space of possibilities and I suspect that there is something that works and is reasonably cheap:
We could potentially store messages in the ruins of prominent cities, if cities have a reasonable chance of being buried+preserved (cities have enough weird materials in them that I expect they’d leave a really visible mark). This could either be used to make small messages easier to find, a place to put payloads (which can potentially be pointed to with a map of the city), or both.
If making preserved messages (or messages with a reasonable shot at fossilization) is extremely cheap, then we could potentially send very large numbers of them. This could be used either to send a bunch of payloads and rely on redundancy, or to allow maps to be very large and expressive. It could also be used to flood the world with massive numbers of maps (>>1M), so that they can be easily found without beacons. (Really extensive flooding sounds more like a last ditch effort once we can see extinction coming, rather than something you’d do preemptively.)
There might be geologically inactive locations where you can just leave giant+out-of-place+slow-to-weather rocks and they have a reasonable probability of remaining intact. Mountain ranges form over much less than 500M years, but it’s not clear to me that the whole world churns since I don’t really know anything about geology. To do this, you’d need to find a rock that wouldn’t wear away entirely, and you’d need a location where it wouldn’t be disturbed too much or end up under ground.
There were some plausible suggestions in a Facebook thread on this topic (including putting stuff in space, defining coordinates with respect to tectonic plates)
If we could come up with a really convincing and reasonably cheap way to send a message, then I think it’s probably worth exploring this idea at least a little bit further. I think the next step would be more seriously analyzing how much good a message could potentially do (which is much more speculative than this step).
I’m in the market for certificates of impact for significant contributions to this problem (in the $100-$10k price range, depending on the size of the contribution).