Not really, Mpemba is about water freezing. There was a recent Nature paper discussion which looked at the non-linear effects of various climate modifiers[1]. I pointed out that some models have warming causing the onset of a glaciation event, the mechanics are warmer air carries more water vapor (itself a powerful greenhouse influence) which sets up a positive feedback loop making the air warmer and then sucking up more water from evaporation, which retains more heat which makes the air warmer. You cross the dewpoint in the stratosphere and that gets 100% solid cloud cover above the tropsphere. At which point the albedo of the planet goes way up (its a uniform white reflector at that point) and the water in the troposphere precipitates out as the mother of all snow storms and the greenhouse gas equivalent of nuclear winter.
Do I actually think this is going to happen? Not really. Are we recording unprecedented levels of greenhouse gases? Yes. Is the climate we're experiencing similar to the climate that was in effect when those levels occurred in the past? Not really, the early Pleistocene was much warmer than it is now. Moore's paper from 1980 was probably pretty influential here. He, like others, is/was really interested in why we had ice ages in the first place.
The point though is that we talk about "global warming" which implies to a lot of people who don't read the papers that our inputs into the climate will only cause increased heat, but it is more accurate to say "climate change" since we don't have good non-linear models for how the planet responds to all the forcing functions we are throwing at it.
We casually throw around terms like "hottest temperatures ever recorded" or "lowest ice levels on record" but some times lose sight of the fact that the "record" we've been keeping in terms of climate change is, according to ice cores, and other geologic carbon dating of glaciation events, perhaps 2% of a cycle. (assuming 100kyr cycle and 2kr of nominal human recordings). Remember that 2% of a year is one week. We can't take a week's worth of measurements and even imagine all of the variations we'll see that year, so we have to be careful about extrapolating with the amount of data we have. Pretty much everyone (except some really hard liners) can agree that climate can and will change, its understanding that change which we can only develop as we go along. And the latest contribution to our understand was the Nature paper that said, in essence, don't try to model this stuff with linear models.
[1] http://www.nature.com/nclimate/journal/v5/n5/full/nclimate25...
[2] https://www.researchgate.net/profile/Ted_Moore/publication/2...