The building and launching approach makes intuitive sense, but it runs against the feudal nature of large tech companies. This point is perhaps clearer if we t

The building and launching approach makes intuitive sense, but it runs against the feudal nature of large tech companies. This point is perhaps clearer if we take Google as an example instead. Mid-level engineers attach themselves to a project or initiate it, then guide it past a certain point. Once promotion is secured, the project is dropped by its initiators. Staffed with lower-status engineers who were enrolled without the prospect of promotion, the project begins a gradual rot and the Google graveyard deepens. The key concept here is that the projects are launched and fueled at the start explicitly for the chance of promotion. In the strategy you describe, with Facebook higher-ups mandating the top-down creation of various projects as part of the global strategy of the company, it would become risky for mid-level engineers to ever get involved. They would not have their names connected with the success of the projects as higher-ups would have imprinted themselves on it beforehand, but would be connected to its failure, whereas in the Google scene the 'moonshot' culture is well-established and limits this sort of blame. Due to massive opportunity costs, Facebook employees would be better off trying to get promoted through the traditional channels, or to quit and pursue the promising ideas themselves and hope to get bought out eventually.