as everyone will most likely say, too much is just a bad thing, especially when it comes to AS5. from personal testing, ive tried numerous ways.
the first step was to clean both the cpu and heatsink with rubbing alcohol (% purity doesnt matter so much, as that will mainly only affect the rate at which it evaporates... 0% alcohol wont evaporate at all (as its mainly just water, lol), and 100% alcohol will probably evaporate too quickly to be much good at all, just as long as its dry shortly after applying it to clean the surface from fingerprint grease and whatever else, as residue does affect how well heat is dissipated.
second step was how much to put and where. the first time i ever put it on, i followed some instructions that said to put a small dot on the center of the cpu (where the cores are hottest, as it will spread from there expanding 'slowly' once pressure and heat is applied, primarily periods of heat and cool at various intervals, off and on, for the most improvement for the period of it setting in).
i tried also using some on the heatsink itself, spreading with a credit card even (like was instructed by a site, so all the divits and 'imperfections' could be filled in, resulting in 'the ideal heat transfer'), that proved to infact be a bad idea, to be honest, as the heat wasnt transferred very well; the heatsink was still fairly cold, and the cpu was running hotter than it shouldve, a lot hotter (it was just too much paste).
so, i cleaned it all off again, and set the bare heatsink ontop of the bare cpu, the heatsink got hotter, temperatures were kept relatively stable, but they still increased slowly though, along with how hot the heatsink felt... so:
i removed the heatsink again, cleaned both surfaces, and put only a small dot on the center of the cpu, and then put the bare heatsink ontop again...
heat dissipation was best that way, with the minimum amount of AC5 possible (just above none; about the size of a BB pellet), and dissipation only keeps improving after a period of up to a week, it definetly needs time, initially it will probably appear to be a horrible paste, and even a waste of money... and for anyone who says it is, probably didnt test it near long enough to find out for sure either, or didnt test different ways, or anything like that.
it does pay to experiment though, as i found you can only benefit from finding the most effective way to use it. it may not be universal for each cpu and heatsink, but the ways i listed above were what worked best for my s939 X2 3800+ stock cooler and then TT blue orb II (its not the best cpu cooler, but it works).