We are already running a massive experiment on the earth; we are pumping huge amounts of green house gases into the atmosphere and we are not really sure what will happen. If things go bad really quickly I would like to have some options (even if high risk) rather than be left with nothing.
"Even more experimenting" is a bad idea. The current situation is bad and we should change its causes. (Otherwise it's like healing tooth with painkillers.) Here I would guess that the average value of risk is highly negative (as in any case with dealing with complex systems).
And there are many imaginable consequences, which are far worse than the current pace of global worming. E.g. poisoning biosphere, accelerating global warming, etc.
And, no, global warming is not an asteroid impact (i.e. killing everyone and everything), so "all or nothing" is a poor mentality (and deadly dangerous!).
I agree 100% that we need to do far more to reduce our emissions, but right now not enough is being done. If we don't get our act together in time, or things on the warming front move faster than everyone hopes, then we need alternatives. We only have one world.
Given that we have record of that stability it seems likely that there are forces which push the climate towards that stability.
Of course it's impossible to know for sure either way until it happens and we witness it and write it down.
Feedbacks evident in the record include ice melt, methane hydrate destabilization, permafrost melt releasing CO2 and methane, forest fires caused by drought, and CO2 release from dried-out topsoil and peat bogs.
A good introduction to all this is in Hansen's book Storms of My Grandchildren.
It's true that the past couple million years have been relatively stable. For at least the past 800,000 years CO2 never went higher than 280ppm. But the last time CO2 was as high as now, the planet was 3 to 6 degrees C warmer, and sea levels 25 to 40 meters higher: http://www.sciencemag.org/content/326/5958/1394.abstract
It's common to use the 'insurance analogy' in these cases, or to invoke the 'precautionary principle'.
The insurance analogy breaks down when you're paying 1/4 the cost of your house every 4 years to have it replaced in case of a possibility. We can't calculate the risk of an asteroid strike so we don't bother trying to avoid it or insure for it. Trying to solve an equation with one unknown variable is pointless, because there is only one correct value, and infinite incorrect values. Therefore the probability of wasting resources is extremely high.
The precautionary principle is a recursive conundrum that eliminates itself and shouldn't be used in serious conversations.
The facts point to a wait and see approach as the only rational response, given the availability of low cost approaches as in the original post here. This of course is not a popular response, but then most religions are based on the investment of time and money up front to deal with the unknown factor of what happens in the future. Unscrupulous people then use that fear to gather power and control over others. I'll always reject that thinking and use a wait and see approach.
The insurance analogy is not the best one to use. A better analogy is if you had cancer and someone offered you a magic pill that cure you. How much would you pay for this drug? I know in my case it would be everything I own, as well as what I could earn in the future, just up to the point where I would cause excessive hardship to my family.
When in comes to avoiding the risks of catastrophic climate change all the facts point towards doing lots of research to develop options given then cost of the research is relatively low and the risks with not having options very high.