There are no cases of "infinite" positive feedback. Most such systems get to a point where breakdown occurs. Not many get to a point where the mechanism decides to simply ramp itself gently back down again.

If your governor is broken, you can always shut off the engine before it breaks down. Something turned those spikes around on a cyclical basis, and until we know exactly what, we don't really know what's going on.

A mechanism where increased temperature drives more CO2 out of "storage" is plausible, and I have no argument with the basic idea. That methane frozen at the bottom of the sea is something that needs watching.

The point is, when you get to the top of this cycle, theoretically it should level off, or at best decrease slightly over a long period of time as uptake mechanisms slowly work. Yet in all the graphs, it peaks and almost immediately (geologically speaking) starts to drop back down again. Why?

Note that all of the papers on this assume temperature changes lead CO2 initially.