First, ignore technology to start and instead look at the business model for each platform. Ask if the platform itself has a reliable means of producing revenue long term. If so, then ask if the platform presents a business model that allows a developer to make money. If your not sure about such stuff ask someone with business experience.Beyond an initial flurry of interest, nifty tech can't sustain a platform if the economic underpinnings are not there. Even if a platform prospers, it doesn't mean that small developers will.
As near as I can tell, Android isn't actually a platform but more like a loose standard. Each phone vendor can customize it to a high degree so there doesn't seem to be a means by which you can write a single app and know it will run on all Android phones. That will cause major market fragmentation so even if Android takes off big time that doesn't mean that every developer, especially small developers, will be able to sell to the entire installed base.
Long term, open platforms (like contemporary PCs) present major problems for small developers. There is no intellectual property protection so developers who don't have large institutional customers they can sue can't prevent piracy. Security will become a major issues as black hats target people's phones. There will be a huge number of crappy or actually fraudulent apps cranked out that make end users leery of buying software from a vendor they don't recognize. This means small developers will have a hard time breaking into the market.
One of my professors in college told me something that has proven true in my 20+ years in the computer industry: The major strength of every design is also its critical weakness and vice versa. The very things that make open platforms attractive to developers and customers are also the same things that will cause them major problems. The very things that turn developers off about closed platforms are the things that provide the greatest benefit to developers long term. Having a closed platform's vendor vet every app slows down acceptance and limits choice but improves overall quality, security and consumer trust. And so on...
Career wise, there is a difference in paths between running your own business and learning an API so that others will hire you. In the former, you should develop for the platform that has the best business model and the one you would most like to use as a consumer. For the latter, you should develop for the platform with the most buzz. Even if it flops, no one will find it odd that experience is on your resume. Just rough rules of thumb.
