As a prospective developer and mac enthusiast, I believe Apple's conflict of interest is crucial. I do believe Apple is also in the unenviable position of needing to support their own platform (and end users with apps) while also trying to (in good faith) support established and budding developers on the platform. Until they get critical mass (and growing market share), the conflict will be touchy.
Office software is probably the most crucial app(s) for Apple. Plenty of rumors suggest Apple has a killer office app in the works. Given the Safari success and IEMac demise and the vulnerable position Apple is in with regards to MSOffice, I think there's a reasonable argument that Apple would be irresponsible (to shareholders et al.) if it didn't have an Office substitute (seriously better than AppleWorks) as a contingency plan. Whether they should release it until necessary or dire circumstances... that's another question.
The most obvious recent example of reckless endangerment or negligence on Apple's part would be the Sherlock/Watson debacle from 2002. I'm not familar with the details, but I gather that Watson was an award-winning app for the Mac, but Apple folded very similar web services features into their bundled, revised Sherlock. As I recall, Apple says they had the feature in the works before Watson's release, and I have no reason to disbelieve them. However, I *do* believe Apple should have a developer friendly strategy to deal with such catch-22's. The obvious one would be to 'buy-out the app' of the victim developer as a sign of 1. goodwill, 2. user compatibility (if appropriate), and 3. the common sense to promote the best of the best in the developer community. Apple could think of the buy-out as seeding a proven developer with $1-4m dollars; perhaps there's a string attached that the monies be used for OSX development. I'm sure the Watson group would gladly have fired up for more development.
This may sound gratuitous, but I know an small insurance company which bought out its competitors in just such a way as it was sucking away clients, and the goodwill was tremendous. With a proper approach, such a technique could be a winning technique for driving app development in a friendly coopetitive fashion.
-=- Devil's Advocacy -=-
On the other hand (and extreme, e.g. MSOffice & Watson, examples aside), I also think Apple developers have it pretty good. The new Apple WebKit looks gorgeous. Granted, browsers are free, but that technology is strong for many applications and is a huge leg up to small developers.
Furthermore, Apple's apps are not bullet proof. Apple creates good first generation apps, but the freebie apps are often not industrial strength; little bugs often go unfixed across multiple releases, larger data sets become unwieldy. There is plenty of room for improvement at least for niche needs. I'm not sure how that adds up to building a thriving company on niches Apple already develops, but there's definitely *some* room even in the Apple fulfilled categories (e.g., Apple's FCPro is serious upgrade to iMovie for example; other app categories could be filled by 3P).
This is my 2c so far, but I'd love to read further insights or perspectives on this important subject.
P.s., here's to hoping the new G5 machines lead to market share growth for Apple and Mac community.