Regarding global warming and what to do about it, Brad DeLong approvingly paraphrases Tyler, "uncertainty is not the friend of doing nothing." Bearing in mind the obvious dangers of contradicting both Brad and Tyler let me counter with "uncertainty is the friend of delay."
If you are faced with two environments one of which has outcomes somewhere between -10 to 10 and the other somewhere between -100 to 100 then I agree that the greater uncertainty of the latter provides no necessary reason for doing less relative to the former.
Suppose, however, that we are uncertain about which environment we are in but the uncertainty will resolve over time. In this case, there is a strong argument for delay. The argument comes from option pricing theory applied to real options. A potential decision is like an option, making the decision is like exercising the option. Uncertainty raises the value of any option which means that the more uncertainty the more we should hold on to the option, i.e. not exercise or delay our decision.
Imagine, for example, that there are 100 doors before us. We can enter any door but once we enter it will be costly to exit. If we don’t decide then over time we learn a little bit about what is behind each door. If our uncertainty to begin with is small, we know that behind each door is more or less the same thing, then learning has little value and we should decide now (assuming some modest cost to waiting). But if our uncertainty is large then learning has a lot of value and we should delay our decision until some of our uncertainty has been resolved.
Applying the theory to global warming isn’t easy because our decisions involve many options and exit costs but if we think that our knowledge of the extent, cost, cause and solutions to global warming are increasing at a faster rate than the danger of global warming then delay of any major decision is a rational policy at the present time.