That makes it an even bigger problem, because its an insanely large advantage for a few very early adopters that won't be possible later (i.e. instamine).

I'm not sure about this whole, "it's all sand" argument about ASICs. NRE matters. You have to get a big enough advantage from the design over its useful lifetime for the chip to ever get designed and manufactured. A small edge over more general purpose hardware that are manufactured in greater quantity with much larger R&D budgets won't cut it.