I've been following the discussion on the x11 thread (
https://bitcointalk.org/index.php?topic=556277.180 ) and there were two interesting highlights - at least interesting to me:
1. A guy estimates that it'd take a few months and a 50k cost to roll out ~300MH FPGAs for x11, with 2k per board cost:
https://bitcointalk.org/index.php?topic=556277.msg6077250#msg6077250If this happens, diff will make block reward go to 5 pretty quickly (and reducing the number of coins perhaps?)
2. Scrypt-N is already ASICable with an ASIC miner that seems to be able to handle N-factors up to 35 years ahead:
Also from their faq:
Can Neon mine for Bitcoins?
No, unfortunately Scryptr is dedicated and optimized for Scrypt mining and does not have a dual-mode
Can Neon support mining for Scrypt Jane or Darkcoin?
Scryptr is designed and optimized for Scrypt mining and not a general purpose mining chip. For Scrypt Jane, only a small subset that uses a combination of SHA256 and salsa20 mix, but in general no. Darkcoin definitely not.
Years of experience with manufacturing processes, specifically new process development, force me to double if not triple every number that that guy said . . assuming he's telling the truth.
So that's a few months x 2.5, 50k development x 2.5 (probably more, but sticking with consistency), 300M# x 2.5, 2k x 2.5 seems more reasonable to me.
If that few months x 2.5 is even close to a year then that's not nearly as bad.
Plus getting that kind of hash out of my GPU's would easily cost me above $50k in just computer parts . . so in comparison thats a far cry more reasonable than the scrypt asic field.