Of course there's an implied promise. If platform operators openly stated "anything you develop can and will be used against you - by us" then no one in their right minds would develop on those platforms.
As you note, the trust involved is invariably abused, but always within limits - vaguely defined as they may be. What Dalton Caldwell is observing is that these parameters appear to be shifting.
That's FB's prerogative, and I don't object to their exercising it. But there were two ways to handle the change. The first was to buy Dalton out, shut down the project, but provide jobs for his team. The second was to do what they did with Instagram, preserving the product and the team, and letting things evolve from there.
It seems like going the latter route would actually encourage people who really care about products to build on FB's platform. The former route also provides incentives, but of a much lower grade, and only appealing to people who don't really care about the products they develop, and see the compensated shut-down / aqui-hire as their highest goal.