There is a line of denier argument that I often hear, it goes something like: the extra warmth means more ice is lost allowing that heat to be radiated to space thus cooling things down. To which I can't help but respond with: so warming leads to cooling? But its still warmer.
That kind of situation sounds self-contradictory, but it's actually quite common in environmental systems.
Say you have one unit of warming that is externally forced.
A negative feedback will partially counter that forcing. Maybe the negative feedback is a multiplier of -0.2. That reduces the warming to 0.8. But this lower warming makes the negative feedback weaker, so it adds back 0.04 (-0.2 * -0.2) of warming. Another iteration of the negative feedback reduces it by -0.008 (-0.2 * -0.2 * -0.2). After an infinite number of iterations, the initial warming settles in at 0.833333 units:
So the negative feedback has reduced the amount of warming, but it hasn't caused actual cooling, just a reduction in the warming.
A positive feedback will have the opposite effect. It will amplify the initial warming, and will also amplify itself. For example, a positive feedback that is a multiplier of +0.2 will increase the initial warming to 1.2 units, then will amplify itself to add another +0.04 units, and another +0.008 units, etc. Eventually, the warming will approach a limit of +1.25 units:
So the negative feedback reduced the warming, and the positive feedback amplified it.
People often make the mistake of thinking that the negative feedback must turn warming into cooling (reverse the sign of the original effect), or that positive feedback must cause the warming to spiral out of control. Neither of those happens as long as the absolute value of the multiplier is less than one (i.e., it's within the range [-0.9999, +0.9999]).
If the multiplier is greater than +1, or less than -1, the system will run away on either the positive or negative side, respectively. That is rare in most environmental systems (fortunately).