Sorry, I didn't want to expand because this has been discussed ad nauseam already; you can dig up the forum for other threads discussing it.

First off, I'm as sure as you are that ASICMiner won't break the network, in any plausible scenario.

Most of the concern boils down to trust in the proof-of-work concept. In other words, any singular entity (even if altruistic) holding majority hashing power nullifies the justification to use proof-of-work.

You need to only imagine what you will see in the press when this happens to get that it's bad for Bitcoin. Would you really want to be in the position where you have to convince the whole user base that everything is fine?

Besides, you only covered the possibility of a malicious owner. How about a cyber attack? How about a physical attack? There is no reason to try to imagine all scenarios and debunk them one by one. It really doesn't matter. In the end, we prefer to use precise algorithms to be immune from our own fallacy of imagination.

With this perspective, I think we can stop discussing whether we should discuss it. It's very straightforward and relatively easy to produce a solution.