I agree. If devs said in the white-paper that their algo is ASIC resistant, they mean it and they will change the algo if necessary just to make sure no ASICs are in the game.
Ethereum probably won't bother changing their algorithm, since they're getting fairly close to POS at which point ALL mining goes "poof".
Any Bitmain ASIC for ethash algorithm better be bloody LOW in price for anyone to have a prayer of making their money back on them, especially if the estimated "still a couple months off" release date I've seen mentioned a couple places proves to be correct or on the LOW side.
I would be inclined to call this "mistake #2 Bitmain has made this year", following in the footsteps of their Cryptonight miner.
Too bad they're still making money hand-over-fist on their SHA256 and Scrypt miners and can AFFORD to make a few multi-million dollar mistakes like this.
ETH devs have always said that the PoS change won't be instant but gradual, starting with 1% of blocks being mined by stakers and going to 100% over the course of a year or longer.
The only question is - did bitmain pay off the dev team well enough for them to allow ASIC's on the network? Because ETH devs have the power to do what monero did and just fork to a modified algorithm, there is no reason to stick with the current algorithm if they get nothing out of it.