[DennisP]

Based on geological history, the negative feedbacks seem predominant for small perturbations, but if you push things a little too hard the positive feedbacks take over. It's happened before several times; a little initial heating from orbital variations, which leads to greenhouse gas emissions, which leads to the global average temperature going up several degrees. A good explanation of the evidence for this is in James Hansen's Storms of My Grandchildren.

People have long thought the threshold to be around 2 degrees C, but now some are thinking it's more like 1.5. The CO2 level everybody agreed was safe was 350 ppm.

[mlyle]

(and then obviously negative feedback takes back over, as we don't run away to infinite or zero temperature).

[easymodex]

I suppose, but before that happens the environment could get very, very unpleasant for humans (and animals in general).

[mlyle]

Yah, the point is if we draw a set of hills with bumps and stable equilibria for a ball to roll down-- there is resistance to perturbation near an initial point, and then a sufficient "nudge" to get over a hill and roll somewhere else... but the ball always ends up in a stable region at the bottom of a hill.

How different a place that new equilibrium is ... is open to debate and not really known. What climate scientists have shown is that it can be, and likely is, pretty bad.