This is a fascinating question. AIR and the new mobile app frameworks are great for prototyping and getting an app out there really quickly, and perhaps then you can make the decision to either build native apps for your biggest platforms; or not.
Alex is obviously pointing to the verbal minority (of which he is a part), because for every person out there that speaks out against, say, the Tweetdeck AIR app; there are probably millions that are satisfied with it and in fact prefer it to the native alternatives (that don't have columns, even though they might be native).
The problem with listening to the verbal minority is that you're then building for the verbal minority, and not the bulk of your true user base.
Startups are all about priorities, right? And logically you should prioritize things that will materially affect your business (such as stability) over things that aren't real problems but get written up anyway. EDIT: I mean, he did leave Campfire because of a stability issue, even though there's a native app for it ;-)
You can dismiss the "verbal minority", or you could treat them as leading indicators. Personally, I'll err on the side of building something that's up to my standards rather than seeing if I can get away with mediocrity.
It's entirely possible to focus on both stability and, say, a quality desktop application. If you're under-resourced to accomplish both goals, maybe you didn't plan correctly.
It is also possible that both moving to a native app and achieving five nines of uptime (from four nines) will do less for the business than changing the call to action on the pricing page. Luckily, God gave us AB testing.
It is an argument about competing priorities when resource constrained: try the cheap stuff prior to dropping six figures of dev time on a feature to rescue business which all available evidence says is one account worth $20 a month. (You could go try to find evidence that users actually want to download native apps. For example, put a download button on your page, then see if anyone clicks it. If nobody does, send me half of your six figure dev budget and you're now strictly better off. It is downright disturbing how many very expensive features fail to change user behavior because business processes optimize for the satisfaction of people other than the people who actually pay money for stuff.)
I agree with your last sentence, but not the one before it. Alex isn't arguing that users consciously prefer native apps; most don't even know what "native app" means. He's arguing that native apps feel better to users and that this feeling is worth something. How much it's worth is hard to quantify. The test you describe wouldn't answer that. As for this:
which all available evidence says is one account worth $20 a month
... surely you haven't reviewed all available evidence on this question? Or did God give it to us when he gave us AB testing? I'm joking, but I think you're trivializing a nontrivial question. I'd like to know how to find "all available evidence"; my experience is that this kind of thing can be very hard to measure.
Edit: it's worth remembering that Steve Jobs made almost an identical argument to Alex's in his critique of Flash last year, even emphasizing that it was his most important point: "We know from painful experience that letting a third party layer of software come between the platform and the developer ultimately results in sub-standard apps" (http://www.apple.com/hotnews/thoughts-on-flash/). Sub-standard apps are exactly what the OP is about.
I'm not sure I follow the connection from being "verbal minority" to being a "tastemaker". You can definitely be both, but just being one doesn't mean you're the other.
Alex is obviously pointing to the verbal minority (of which he is a part), because for every person out there that speaks out against, say, the Tweetdeck AIR app; there are probably millions that are satisfied with it and in fact prefer it to the native alternatives (that don't have columns, even though they might be native).
The problem with listening to the verbal minority is that you're then building for the verbal minority, and not the bulk of your true user base.
Startups are all about priorities, right? And logically you should prioritize things that will materially affect your business (such as stability) over things that aren't real problems but get written up anyway. EDIT: I mean, he did leave Campfire because of a stability issue, even though there's a native app for it ;-)