I agree with Dingo. Runaway warming would imply that the equilibrium climate sensitivity would be much higher than the accepted mainstream values between 1.5 and 4.5 K/doubling. Positive feedbacks eventually reach an equilibrium. Each additional "loop" becomes weaker than the positive feedback loop before it. If an initial increase of Carbon Dioxide caused 1.5 K of warming, and positive feedbacks from that initial warming caused an additional 0.8 K, then those additional positive feedbacks from the warming would cause an additional 0.4 K, and those would cause additional warming of 0.2 K, then 0.1 K, 0.05 K, 0.025 K etc. So even though net feedbacks are weakly to strongly positive in the climate system, runaway warming is unlikely. This is why we get a value between 1.5 to 4.5 K for ECS, and not infinity Kelvin.
And there are additional positive feedbacks that we have yet to uncover, like the permafrost feedback. But those would work in the same way as described above.