[ExI] Warren Buffett is worried too and thinks Republicans are "asinine"

Anders Sandberg anders at aleph.se
Wed Oct 16 21:20:46 UTC 2013

> "I'm worried about damage to an asset that we've carefully cultivated 
> for 237 years [...] it would be asinine for the U.S. to risk its 
> hard-won reputation for paying its bills on time.[...]

I had a discussion this week with Bjørn Lomborg about the rationality of 
geoengineering, and I mentioned the argument that stopping aerosol 
injection causes a very rapid move to the temperatures one would have 
had without any intervention. That would obviously be very bad. However, 
Bjørn disagreed, since (1) it would be irrational to stop in this case, 
and (2) this kind of geoengineering is relatively cheap and in the case 
of some major consortium failing one could easily imagine a billionaire 
or single nation sponsoring extra efforts, either altruistically or just 
because the damage to their own interests. Point taken.

Now, in the light of the brinkmanship in the US we might see another 
twist on the problem. Geoengineering is usually framed as a game between 
national powers, and would run according to the rationality of 
game-theory on that level. But the US demonstrates that games on lower 
levels can produce irrational outcomes on higher levels: the US 
brinkmanship was (perhaps) rational for individual politicians and 
politicial coalitions, but likely made the US as an agent nearly act 
irrationally. Yes, it is asinine for the US to risk its reputation, but 
it might not be asinine for some politicians to risk it.

It is not hard to imagine other versions of this situation, where 
rational low-level games produce irrational higher-level actions. In a 
way, this might just be a formalisation of a lot of history and politics 
- a contributing explanation for the "Cock-up Theory of History" ( 
<https://en.wikipedia.org/wiki/Hanlon%27s_razor> ).

Low-level brinkmanship upsetting bigger global collective goods doesn't 
sound far-fetched, and might actually be a serious Achilles heel of 
geoengineering. I have some colleauges who plan to model it in more 
detail. In the meantime I hope that if we find ourselves in such a mess 
some Warren Buffett might unilaterally solve the problem (in exchange 
for all weather reports ending with a thank you note). Of course, the 
unilateral possibility of geoengineering leads to other fun problems ( 
http://www.nickbostrom.com/papers/unilateralist.pdf )...

(Incidentally, another senior researcher mentioned that we ought to 
develop geoengineering techniques regardless, as a just in case fix for 
methane burps - here a modest aerosol injection gets us past the worst, 
and cessation is not a major issue once past the peak.)

Anders Sandberg,
Future of Humanity Institute
Oxford Martin School
Faculty of Philosophy
Oxford University

More information about the extropy-chat mailing list