Certainly, the long-term economic outlook is extremely uncertain, but I’m not sure what “who knows!” is a better answer to cumulatively worsening climate problems than to cumulatively worsening fiscal problems.
Her point is that ignoring climate change because of model uncertainty is like ignoring the entitlement outlook because of model uncertainty. This is part of a theme of hers, which is that it is odd to worry about climate change and not about entitlements, or vice-versa. They are similar problems in that they involve the long term and uncertainty.
However, I would argue that there are differences. Perhaps the least important difference is that I have more confidence in the models that project trouble ahead for Medicare than in climate models.
The most important difference is that we can improve the entitlement outlook costlessly. The problem with entitlements is that we are promising more than we can deliver (at least if the economic projections are reasonably correct). Suppose that we raise the age of eligibility for Social Security and Medicare for people under 50 to something like 72, and then we index that age for longevity. This will change what we promise. We still have the option, down the road, of delivering more benefits to people now under 50. But lowering what we promise them helps forestall the situation in which we either renege on our promise or we raise tax rates ginormously to try to keep our promises.
To appease the economic models of entitlements, we don't have to make any sacrifices today--we just have to make more conservative promises going forward. To appease the global warming models, we have to make rather large sacrifices of output.
Suppose it were the other way around, and that there were a global warming policy that cost nothing if it subsequently turns out that the models are incorrect, while fixing entitlements would cause irreversible losses in output. In that case, I would be all for taking steps to deal with global warming and strongly against taking steps to curb entitlements.