In short: asic producers can release a new asic version for every N as previous generation asic miners will be too power hungry anyway.
I agree. Scrypt-n is however much more expensive to implement is an asic! And changing to scrypt-n would remove the threat of the currently announced asic miners. And what is even more important: it would scare asic developers from trying again for scrypt-n. You only need to show your determination not to allow asics for your coin!
I fear that Scrypt's way to implement RAM, and now Scrypt-N's to implement way more RAM, will have a boomerang effect in trying to prevent centralization, creating more centralization as an unintended mid-term consequence.
The Litecoin devs are right in one thing: If a coin is mine-able in a GPU, it's game over. ASICs are just a matter of time from that point onwards. The thought however of making the ASICs expensive through much RAM, is something that will escalate ASIC prices and reduce the affordability of ASIC miners for the masses.
If RAM was not a part of the equation, a greater degree of decentralization could occur once a coin moves from the GPU to the ASIC cycle of mining, through affordable ASICs that normal people can actually buy without paying a fortune. So the fact that ASICs will have to be very costly in order to have much RAM, is not really good down the road.
So whats the solution?
If any...
LPC
I think the solution, in the long run, is the replacement of proof-of-work, with an artificial intelligence entity that will run the currency network. It will be a 100% reliable and trusted "third party" which is neutral and objective in its handling of transactions, solving both the proof of work and proof of stake inadequacies that are a result of dumb if-then-else algorithms. It will also solve issues related to network resilience, spam attacks, etc.
As for now: Darkcoin implements a diminishing reward formula (moore's law) that reduces the block reward as the hashrate goes up. This allowed CPU miners to make more DRKs per block than GPU miners and ASIC miners will get the least - when they launch. With CPU mining we were getting something like 75-100 coins per block if I remember correctly, with GPU we are now in a range between 5-25, but with the existing hashrate it fluctuates in the 15-20 range. If the diff goes up a lot, it will bottom out at 5 DRKs per block. Thus early miners with CPU/GPUs get a larger share of the monetary base for the mid-term.
Yet, it may solve one problem and create another (less incentive for securing the network), but all this is highly dependent on the price of each coin. If the price is high, the incentive is there. We'll have to see how this one works out.
True, there is some chance that a company will try to make an ASIC that can do N-Scrypt for any N up to some number, say, M. Put aside the arguments about how difficult that is, it's at least technically possible. They will still be hesitant to do it for fear that there's a hard fork taking vertcoin to M+1.
Contrast that with X11, where you have a similar argument that it's very difficult to implement it in an ASIC. But once it's done, game over. You have to move to something else, entirely.
X11 hashes are better for ASICs but a design for the ASIC must be built first. Scrypt design is already complete and rolled, so the circuits must have a small tweak to account for the -N and it should be closer to launch than X11 - if ASIC designers want to implement it. ASIC designers don't need to do all the work from scratch with scrypt-N. Most of the job is done already. Yet, X11 might have a bigger incentive if far more coins adopt it instead of scrypt-n, and if it also allows faster acceleration compared to GPUs (compared to the acceleration factor of scrypt-n vs gpus).
I think miners in particular are very "doomy" regarding the ASIC factor. When we say it's over, what we really mean is that "it's over for us, because we can't make money with our GPUs". The coins continue their life as normal, even with ASICs, as Bitcoin and other SHA-256 coins demonstrated. Litecoin and Dogecoin continue as usual. Actually Dogecoin (!!!) is probably in the best position of all, given that by the time of mass ASIC adoption, it will have already distributed the bulk of its monetary base to GPU miners.