Pedro Domingos

The deeper you dig into climate models, the stronger the case for acting sooner.

The greater the uncertainty, the stronger the case for policies that work regardless of the cause of warming (geoengineering) over policies that presuppose one (decarbonization).

That's a surprising perspective. An argument against geoengineering is that it has the potential for a panoply of unpredictable side-effects, while keeping GHG levels near current ones is less likely to create tail risks. But do you agree that uncertainty is NOT our friend?

I.E. when you say you see more uncertainty than people typically assume, are you calling for more aggressive action on climate? Because many people might interpret your tweet as suggesting the opposite.

Geoengineering cheaply takes care of the tail risk. Its likely side effects are minor and easily reversed. Greater uncertainty means lower likelihood that humans are a major cause of warming. Overall, it weakens the case for decarbonization.

What do you think of this analysis? https://scholar.harvard.edu/files/weitzman/files/fattaileduncertaintyeconomics.pdf … For given average level of climate change, if the effects of GHG are less certain, that means more risk of catastrophe (since costs grow disproportionately with more warming)

How does this paper conflict at all with what I said? Geoengineering caps the tail risk. More uncertainty also means lower likelihood that GHGs are a major problem a at all. Overall, less of a case for decarbonization, which would be massively disruptive for unclear gain.

More uncertainty about the effects of GHGs and the various feedback mechanisms means that there’s *more* of a chance that GHG‘s are a catastrophic risk.