A play-money prediction market about when SRM will first be deployed

[mako yass]

I've created a prediction market about when Solar Radiation Management will first be "seriously" deployed (for a definition of "seriously", see the question). It's just on Manifold, so it's not using real money, but Manifold players take the game pretty seriously, so I expect it to still produce a somewhat meaningful probability distribution over possible dates.
In the very least, members of the list may find it fun to bet on.

The prediction market: https://manifold.markets/makoyass/when-will-serious-deployment-of-sol

ad astra.

system designer mako yass

[Andrew Song]

Why 30%? What if it's only 1% and deployment is measurable, recorded, and verified?

--
You received this message because you are subscribed to the Google Groups "geoengineering" group.
To unsubscribe from this group and stop receiving emails from it, send an email to geoengineerin...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/geoengineering/b3d9a53f-2a84-4980-9420-290268ee2bb9n%40googlegroups.com.

[mako yass]

30% strikes a balance between being serious enough that it indicates that they (whoever is doing it, anyone interested in betting on who's going to go first? :) ) probably going to go all the way, and occurring closer to the beginning than the end, so easier to estimate from political factors alone without having to guess about what techniques might be used.

Would you argue that I should lower this? Would 10% represent just as much of a commitment?

[Andrew Lockley]

30pc of the increase since 1990 is around 11y of global warming - roughly 1/6 of all GHGs emitted or around 0.2C of cooling. The most popular votes are late 2020s. The equipment to do this work simply doesn't exist - it's not even at the detailed design stage. To get this done by the late 2020s would require an Apollo style engineering project (probably smaller expenditure, probably larger fleet tonnage - but you get the idea), followed by the most sudden global cooling since Pinatubo - all without triggering a war among nuclear armed superpowers who are currently having the largest land war since WWII. 

I'm at the gung ho end of gung ho, and even I can confidently say that simply isn't going to happen. 

Andrew Lockley 

To view this discussion on the web visit https://groups.google.com/d/msgid/geoengineering/101f0bcc-85ce-48f5-b52c-f659ff415fa5n%40googlegroups.com.

[Adrian Hindes]

Agree with Andrew on this. I don't think any scientific or engineering research supports 30% (ie. 0.2C) cooling from 1990 baseline in any time within the next decade. The prediction market should really *start* at 2030 (maybe 2035, really), and go in interval stages every five years.

See Smith & Wagner (2018) for a comprehensive analysis of a 15 year ramp-up scenario, starting at 2033. They even argue 2033 is far too soon, and in that linear ramp up scenario it still takes till 2042 to reach -0.2C cooling.

Stratospheric aerosol injection tactics and costs in the first 15 years of deployment - IOPscience

[mako yass]

Okay, I also agree with Andrew, though. The reason the years end at 2030 is that I embarrassingly thought this would be a Free Response market, didn't realize we wouldn't be able to add additional response years to a Multiple Choice market. I'm probably going to have to make another market :(

So yeah, bets weren't meant to be restricted to the 2020s. (That said, I'm still a bit unclear on how opportunity cost should affect our willingness to bet over far future resolution dates. Some members of the forecaster/mechanism design community argue that we wont get informative figures from > 15 year bets until we have derivative trading, but I'm not sure how much the effects of the derivatives they mention would differ from manifold's existing loan system (you automatically receive interest free loans asymptotically approaching the sum of your outstanding bets).)

I should point out that we're betting absolutely tiny amounts here so far, due to opportunity costs, and because there probably aren't many climate intervention researchers already hanging out on manifold. I'm mostly just betting to get the ball rolling, so, heh, the more wrong I am, the better I am performing my role as market creator (the more mana you get to take from me).

I should probably investigate the geopolitical risks of climate intervention more deeply. Though my field is not geoengineering, the geopolitics of emerging technology increasingly kind of probably is going to be my field, at this point, so I should know more about this. If I had to bet now, though (and I don't, and shouldn't) I'd bet that escalation threats concerning what are essentially attempts to stabilize our own climate (a climate which we need? for agriculture?), are in the category of threats we have to vocally dismiss. It's like threatening to escalate if we dare to commit the offense of engaging in commercial fishing, or monitoring the ocean for foreign submarines. There are many threat-bargains we have to be able to reject as unreasonable, and this may be one of them.

Regarding engineering difficulties, I think it's easy to underestimate technological change. We have more researchers and engineers and modelling tools than we ever have, and technical progress follows counterintuitive laws; what the machine gives us is not just the average of the work of research scenes that we can currently see, what the machine gives us the weirdest, unprecedented best. Reliably, every year, we receive impossible findings, and then those impossible findings proliferate into all hands.

I think unexpected engineering solutions are likely to pop up really as soon as there's a market for cooling credits. (I'd expect that around, median, 2029? Maybe that would be a better topic for a manifold question!)