I'm sorry but this isn't true. Platforms such as NOLOH, Cappuccino and many others allows you to develop sophisticated web applications without having to worry about JS, CSS, HTML5, or browser differences, among many other things.
You say Developing for iPhone, you write the app once and it will work the same on everyone's hardware, similarly developing with NOLOH or Cappuccino for example allows you to write your code once and have it work automatically on all browsers, and operating systems.
Furthermore, you definitely wouldn't have to spend millions of dollars to develop products like Google's with these technologies.
As a matter of fact we've had complete novices use NOLOH and create applications significantly more advanced than their counterparts applications from Google, Microsoft, and others.
The problem is that as you say developers aren't using the tools that are available, it's not that these tools don't exist. In my interactions with developers the problem usually stems from not wanting to use these tools that are available. For whatever reason when it comes to the web many developers don't feel like they should be using frameworks, but rather deal with the hell that is the current web themselves.
However, when it comes to building applications for iPhone, iPad, and similar devices they have no problems using the frameworks provided to them.
You can have the best tools around but if nobody uses them it doesn't matter, but saying that the web is a bad development platform or that it takes many millions of dollars of resources just isn't true any more. All it takes is an open mind.
I think you meant, "it will work the same on everyone's iPhones".
Also so far most mobile browsers seem to be based on WebKit, and you can assume a specific browser, so it should be easier than on the open www.
The biggest problem though is a lack of a well-organized app store: centralized distribution, the arguable benefits of a review process, and most significantly the monetization framework that it provides.
Also, GWT.
Probably doesn't go far enough for "easily make user interfaces" though.
Voting on HN has gone to the dogs a bit recently with thoughtless (and often mass) downvoting. There is often no real "reason" except some dumb fellow got up oon the wrong side of the bed today.
My suggestion, don't worry about it, just say what you want to. By and large, good comments are upvoted sooner or later. (upvoted you btw)
1. Luckily for Apple, it's precisely the class of applications that makes them money. I can't imagine Steve cares very much if GMail is a native or web application.
Ironically, a main disadvantage it has is extreme dependence and incompatibility among different browsers, and dealing with this can easily swamp the development cycle. Even performance differences across different versions of JS can make massive differences to an application. This situation shows not even a hint that it could improve in the future. Developing for iPhone, you write the app once and it will work the same on everyone's hardware.
I've been modded down on forums for a couple of years consistently for pointing this out, yet you sure don't see many developers actually building complex apps using JavaScript platforms, despite the lip service given to it.