Pretty cool! Though I think, purely from a technical point of view, this does nothing to 'cancel out' atmospheric turbulence like AO would. However it does reduce the area any atmospheric turbulence can 'smear' photons to a degree (at the expense of total light input - bad for DSOs, not an issue for planetary).
To be honest, in bad seeing, for general use, you would probably get much better mileage from a Deconvolution plug-in, using a Gaussian point spread function to 'de-blur' your stacked image.
However, if you're doing planetary, an apodising mask may be beneficial as you're stacking lots of short exposures (does depend on your stacking algorithm!), not giving the atmosphere enough time to smear out over a larger area (as in a long exposure). In the latter case (again depends on your stacking algorithm) the atmospheric wobble is not random enough but clearly shows a direction/distortion ('the heat haze effect').
The whole point of using a Gaussian point spread function when deconvolving is that, over time (e.g. in a long exposure), the atmosphere's smearing approximates a 'random smear in all directions' distribution (which matches a Gaussian distribution). That's why a Gaussian 'de-blur' works so well in this type of scenario.
Understanding the point spread function (e.g. how a point light is diffracted into these patterns) really helps understanding why images look the way they look and how you can improve them. Once you understand that, you will understand mis-collimation and all sorts of aberrations.
Plus, there's all kinds of neat tricks you can do (including deconvolution, apodising masks) and problems you can solve.
. What kind of price tag would we be looking at I wonder .....but then again I mite be dreaming!!
time to save.
Malcolm & Marc for the info 
