I love this! Two fantastic key points:

1. Only use the platform to accelerate growth, don’t depend on it. If you completely depend on them and they decide to change (in a big way), you’re dead. Zynga is another great example of how horribly wrong things can go.
2. Don’t build a product that is only intended to fill a platform’s gaps. When they decide to fill that gap themselves, you’re dead. I saw this happen when I was at Apple. There would be a cool little software company that addressed some deficiency in the Mac OS. We would add that feature/fix that issue in the next OS release, and that little company died overnight.

These mega-platforms are really just distribution channels. You should always diversify your channels and plan for independent growth to minimize your exposure.