Global Warming: The Effect of Uncertainty on Policy
A new thought on global warming: even if there is a 10% chance of bad things happening because of global warming, we should have a plan to deal with it. And even if there is a 90% chance, we should have a plan that is reversible, not causing permanent harm.
I think this lends argument to plans such as getting ready to put soot into the atmosphere to make it more opaque— but for waiting to do so until the bad effects are noticeable. If we were to be restricted to CO2-emission-reducing policies, it makes for a complicated and interesting tradeoff between restricting now, possibly uselessly, and restricting later, when adjustment costs are higher.