Marco argues that while this behavior is bad for incumbents, it's good for new entrants, but that doesn't make any sense; if you're a new entrant, your goal over the long term is to become an incumbent. Blowing everything up may benefit you today, but if you survive long enough to see the next demolition spree, then you'll be the one getting 'sploded. It's like arguing that living next to a volcano is good for development because it periodically clears out old building stock.
A good platform is one you can build a business on. Sudden, dramatic change that flushes your investment to date down the toilet is bad for business.
The platform was designed the iPhone platform between 2005 and 2007, and while it served its purpose very well I don't think that we should limit innovation for the sake of supporting old apps.
Also - he was highlighting that this will suck somewhat for incumbents (which, having sold Instapaper, and "The Magazine" he is not), but that it offers great opportunities for people willing to fully commit to the new iOS 7 worldview (who in turn may be wiped out in another 8 years - but hey, 8 years is a long time.
Regarding, "If you're going to invest time and money in a platform, the last thing you want is to find out after that investment has been sunk that the platform steward likes to periodically smash everything."
Heck, based on a small sampling of anecdotes from iOS developers I know - I bet 90% of all iOS developers make the majority of their income on an app in in less than six months, fewer than 10% experience "significant income" over a full year, and less than 1% have multi-year incomes from an app that is enough to have a comfortable six-figure salary. The nature of iOS application development is different from Desktop Application Development, and, for most (excluding the top 5,000 Apps) - the real money is in new applications.
So - yes, this may be jarring for that 1% - but the other 99% have a great opportunity opening for the next six - nine months.
There are better ways to allocate development time than reimplementing an already-fine OS.
I'm still flabbergasted that Apple did their big overhaul (apparently) without addressing inter-application communication, default apps, Siri APIs or most of the other "power user" issues that would have helped them grow their computing paradigm. To me, increasing the scope and power of iOS devices while maintaining the predictability and control that are at the heart of Apple's take on mobile/touch-based computing is the problem Apple has to solve over the next few years. Judging from what I've heard about iOS 7, it sounds like Apple couldn't disagree more.