I think that these types of products aren't worth being attached to in the long run. 1. they aren't real IDE's 2. they age as the technology rapidly evolves around them what their simplicity accounts for today won't capture the platform of tomorrow and 3. these types of tools are meant to be tastes that get you excited in much more robust systems that are out there. I think that if you used any type of platform like this for more that a year or two without trying to move on from it then you didn't really learn anything.
This was exactly my thought, too. Upon investigating Project Spark, it seems like something more akin to a pumped-up LittleBigPlanet than an actual game development framework. I'm sure folks will have fun playing around with it, but not much more than that.
The interesting thing to do is to look at the progression. Project spark is a huge improvement over little big planet 1 both in graphics and game variety , and maybe programming model/power. Why won't this kind of improvement continue , or maybe there will be some collaborative model between spark and some framework ?