I cant believe maxwell has the audacity to call a miner optimization an attack. Thats what ASIC manufacturers have been doing all along, optimizing miners.
Did you know that in cryptography an attack is a term describing a means to weaken the security of a cryptographic function?
It's even obvious from the title of Greg's proposal:
BIP proposal: Inhibiting a covert attack on the Bitcoin POW functionIn this case specifically this attack has more implications than simply lowering the costs for a particular miner: it also influences the transaction selection with yet unknown consequences to the security model of Bitcoin.