Who would you rather be? The Software developer that has to rewrite the software of your protocol to fork away from the ASIC, or the Hardware manufacturer who changes the Algo to hash on the new forked protocol? I'll take the latter....every time.
This is a separate topic on it's own but i would pick the former, software developer, any time. It's simply much more profitable to do so.
Imagine how much power and money Vitalik has, with a click of his fingers he could influence the entire ETH dev team to fork and destroy the E3. He could also short and destroy mining or the entire ETH network by uploading some bad code. He knows exactly when updates are going to happen beforehand and could even use this opportunity to buy ETH before it pumps. It is so profitable to be Vitalik Buterin.
This is a completely different scope than what we are referring to. The topic at hand is the reactive response by Software Engineers to ASIC's coming onto their pre-existing networks. Then, once they react one time, the ASIC manufacturers can just reprogram to the new "resistant" software. ASIC = Offense, Software Engineer is on defensive. Besides, they have to take the entire protocol and the users into account before constantly changing the program. It's just not as easy as it sounds. ASIC offensive redesigns are fairly simple and fast. Eventually if a team mishandles the changing of programs to circumvent ASIC, they run a very real risk of losing their Coin to the fork and to the ASIC community.
I think you can agree that Buterin or "little dracula" or whatever his name is... is quite the exception and not the rule.
For Ethereum the software came first and ASIC is only now catching up. So when software changes the mining requirement the ASIC will need to catch up. But if that cycle keeps on repeating, does it even matter which is the cause and which is the effect of the latest revision? Each change becomes the effect of the previous cycle and the cause for the next cycle.