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 1300 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).