The platform trap
Like an expert chess player thinking five moves ahead, it can be inviting to imagine yourself atop of the heap, controlling key parts of the internet by having your platform powering large industries that are wide and varied. But the problem is always in getting started.
When you rattle off a series of average ideas in succession they can sound rather compelling when grouped together. But each individually can rarely stand on their own. Answer honestly the question of why each component is better than the world's leading company who spends all their time on that one thing?
The other problem with launching as a platform is that it often takes a lot longer than if you try to be the best at something small and new. Because platforms require a lot of building the gap between hope and reality can get dangerously large.
Platforms also rarely launch as platforms. Facebook embedded a social fabric into photos before launching a platform to allow other companies to do the same to games, music and other categories.
In abstract, platforms can seem elegant and beautiful, but in reality it's very hard to launch as the best in more than one thing. Crawl before you walk and then run.