But from my simple understanding of FPGA's, wouldn't saying that if one could hit 300m# as was implied . . it wouldn't be much of a leap to say that a comparable ASIC would reach well over 1G#?
Perhaps. But if FPGAs do the job cheaply (2k per board as he calculates it), then there is not much incentive for ASIC manufacturers unless they can offer much better price for performance. If the FPGA costs 2k (and is sold 3k on retail) and the ASIC goes for 1gh at 10k retail price, it cant be sold. People would rather buy multiple FPGAs. It'll have no serious cost/hash advantage over the FPGA and FPGAs are reprogrammable/reusable as far as I know, which allow them to be resold when they are obsolete as x11 miners (something that can't happen with ASICs, limiting their life).
This seems very unrealistic, based on the typical GPU hashing speeds I see. Unless the miner were to be so inefficient that it makes these numbers non-scalable I can't see an FPGA coming to 300m#.
I have no idea how they perform really, but I'll say this which is somewhat "weird": Three days ago I saw a dream that FPGAs could go 52x over GPUs on X11 (that would indicate ~150Mh). I do not know if its prophetic or subconscious junk, but it was too specific to ignore - which usually indicates that it's not subconscious junk. Last time I saw something about cryptocurrencies, it was about Litecoin going from 16$ to 25$ (and it happened within the next 2 days).
Well, if they must come then I'd hope them to be more in line with your dreams than the estimation.
I guess I'll be doing some digging on FPGA's, to get a better understanding on how fast they could implement x11.