dont quote me on this but i think the new client will be asic resistant. An ASIC will only find 5 or 6 goldcoin blocks per 10 min. and by that time the block reward will have decreased. With a multi-threaded client i think timers can be set to ensure blocks keep rolling in on a 2 minute average no matter what the diff. and hashrate is at. So if a multipool or ASIC jumps on they will receive a lot of orphan blocks and when they are done blocks should still keep rolling in on a 2 minute average. Many coins are going to fall hard due to difficulty becoming locked really high. Scrypt coins that survive this should do very well. Maybe microguy or akumaburn can confirm this.
ASIC resistance has to do with the proof of work hash and how difficult/expensive it is to build such a device for a particular hash. We have heard about scrypt ASIC miners under development and for preorder. Eventually, if scrypt ASIC systems catch on, they will put GPU miners out of business just like they did for Bitcoin. There was a time last year when it was more profitable for me to mine bitcoins with my GPU than for Litecoin.
The new client for GoldCoin and its 51% defense provide certain advantages regardless of what hardware is being used to mine. If some miner started using some 5MH/s ASICS for Goldcoin, it would be the same as if that miner had build several rigs with 8x 7950 GPU's. Either way, he would get orphans, etc, as he pushed the 2 minute per block minimum limit. The difficulty would rise to the correct level for the network hashrate over the course of several hours or a day. If hundreds of ASICS were put on Goldcoin, then the CPU and GPU miners would be spending more on electricity than what was mined and give up!
If we wanted to GoldCoin to become ASIC resistant to allow the miners to continue using their current GPU hardware without the ASICS pushing the difficulty up (as it happened with bitcoin), then we would need to change the proof of work hash. One group was suggesting this for Litecoin and the main developers rejected it. Some have suggested X11. X11 is a combination of 11 ASIC friendly hashes and would only delay the time when an ASIC would be made for it. Scrypt-N wounds interesting, because as N increases, the amount of memory required to do the hash increases. More memory means more expensive ASICS. To handle these coins, the ASICs would need to have enough memory. After a certain point in time, they may not have enough memory to continue mining the coin.
Here are a few thoughts concerning Scrypt-N and ASIC resistance:
http://bitcoin.stackexchange.com/questions/22604/is-vertcoin-really-asic-resistant
