SWIM has attempted a couple of times to reduce ketone with ammonium acetate to the product. It seems the NaBH4 is too reactive and reduces the ketone to the alcohol. Very poor desired product yields were seen (< 10%).
Ammonia is not methylamine! The N-methylimine from methylamine is much more stable than the imine from ammonia. When using methylamine, all of the ketone gets converted to methylimine before the borohydride is added and so the borohydride reduces the methylimine to the N-methylisopropylamine. When using ammonia, the reaction equlibrium does not favour the imine formation so there is a lot of unconverted ketone when the borohydride is added and the borohydride reduces this to alcohol.
SWIM has tried making conditions absolutely favorable (silica gel, proper pH adjustment, etc....). As far as I can see there is nothing different from your procedure and SWIM's except for the Methylamine.
Yes, that is the difference. The difference between methylamine and ammonia.
Is there something SWIM is missing?
Yes, that there is a big difference between using methylamine and using ammonia.
Does this method actually work well?
Yes, it does work well, very well. It works very well when using methylamine. It hardly works at all when using ammonia.
If so, why all the fuss on using cyanoboro(which, BTW, works very well)?
Because cyanoborohydride will work with ammonia to produce the amphetamine.