Yes, a semi-research project with lots of peel offs works well. It is nice that Wolfram has a nice product with good profits that can bankroll something like Alpha. Seemingly inefficient (not directly profitable) experimentation is a great way of driving innovation and invention. Contrast with Apple, who has a razor sharp focus on profitable products: they don't make many mistakes, and they aren't taking any risks these days.

As an aside, I don't think the mostly constructed knowledge representation approach taken by Alpha will scale in the long run. You guys might want to look at playing around more with machine learning. On the other hand, there are gaps in machine learning that are best filled by DSLs and explicit construction.