[ExI] homo sapiens as endangered species

[Anders Sandberg]

Typically a population under 2000 that will eventually die out can yet 
remain for many centuries. Extinction processes are lingering affairs. 
Apparently many smaller Pacific island populations were not stable, and 
would occasionally die out. There is also evidence for Tasmanian 
Technology Traps in Oceania: you need a decently large population to 
maintain even neolithic technology.

Under the right circumstances a small population can also grow to become 
a large one. The americas appear to have been initially colonized by a 
very small group, judging from genetic bottleneck effects. However, that 
group might simply have been the first one to succeed - there are no 
traces of previous failures left.


For xrisk management with refuges, there are a few issues:

* Building refuges for 2000+ people is doable but nontrivial. Storing 
food for that population for a 10+ year agricultural crash requires some 
space, and maintaining the infrastructure for the survivors will require 
some work.

* I am interested in "natural" refuges like nuclear submarines (skewed 
gender ratios, crew ~140 a bit too small, good natural protection), 
isolated low-tech populations (often have hunter-gatherer/agricultural 
skills, vulnerable to climate-affecting disasters), random resource 
concentrations (Wallmart logistics centers) and isolated high-tech 
populations (e.g. McMurdo Station, vulnerable to being cut off, 
sometimes large enough). It might be interesting to see if they could be 
made better at acting as refuges through simple modifications.

* After a typical near-xrisk that is survivable in refuges but not 
outside the problem is bootstrapping agriculture (hunter-gathering is 
inefficient and might work badly in the ecological aftermath of a 
disaster). This requires seed resources and skill (farming equipment is 
a bonus). Many natural refuges will be lacking in this.

* Depending on disaster type it might be good to have multiple smaller 
refuges to increase survival chances of individual refuges; afterwards 
there is a benefit in grouping populations together (skills, demographic 
solidity) with the caveat of epidemics.

* Solving the Tasmanian Technology Trap might be tough. Just having all 
the necessary information in some robust form is better than nothing, 
but there has to be time and interest to learn it too (= requires 
ecological surplus). Knowing about the problem might ameliorate it a 
bit, but I am not confident it will fix it over multiple generations.

* What is the population requirement for building and maintaining an 
industrial society? As far as I know there has never been a totally 
isolated industrial society. England and Wales during the industrial 
revolution had 5-7 million people, so this likely gives an upper bound.

* Some technology can likely be packaged in ways that make it available 
for a long time. Other technology is fragile and requires a working 
infrastructure. Economic incentives tends to push towards the later.

* Low genetic diversity is inconvenient but not a showstopper (as the 
America example shows). Lots of women and a convenient sperm bank might 
be ideal from a genetics point of view, but might not be practical, 
available or acceptable to the survivors. Survivors are likely more 
interested in having good lives for themselves and their families than 
the fate of the species.

* Self-sufficient space refuges would of course be the near perfect 
solution, except that the thresholds to entry are currently too high. 
Merely having a base somewhere is not enough: it needs to be able to 
function as a proper civilization seed.

* Refuges on Earth work fine against global climate disasters due to 
impacts, supervolcanos, climate changes, GRBs and nuclear winters. They 
also work against pandemics. They are likely not effective against 
intentional threats like nanowarfare or runaway AI.

-- 
Anders Sandberg,
Future of Humanity Institute
Philosophy Faculty of Oxford University