If there is an “evil” person/organization, that usually means that they have one or more terminal preferences that you would think of as evil. But they also probably have non-evil terminal preferences as well. And they’re just as likely to spend money to satisfy their non-evil terminal preferences as their evil ones.
Any transaction with an evil person/organization, that satisfies their non-evil terminal preferences, should be encouraged, because you’re decreasing the resources they have available to them to do evil with, without causing evil to happen.
Facebook bought WhatsApp because it increases the size of their social graph, sure. But if I consider “making statistical associations between people” a morally-neutral act, and “actively spying on the particular actions of individuals” to be an evil act, then “feeding” Facebook the WhatsApp social graph was, from my perspective, a setback on the pace of the evil they were doing. In an alternative world where they didn’t buy WhatsApp but instead just outcompeted it with Messenger and everyone switched from WhatsApp to Messenger, Facebook would have done more evil by spying on conversations that—in this world—it didn’t get to spy on, because they occurred in the interim period where they owned WhatsApp but hadn’t fully wiretapped it yet.
Sure, it will finish digesting WhatsApp and change it into something that allows them to spy on people. In the meantime, the WhatsApp founders are free of working for WhatsApp; and so could, theoretically, just build (or, more easily, invest in the creation of) another secure social network. And then trust people to just decide to abandon WhatsApp for its Facebook-owned-ness (like they’re already abandoning Messenger for its Facebook-owned-ness) and switch to it.
It’s a cycle: Facebook will probably make every successive competitor secure-chat network company an offer they can’t refuse, and absorb them. But this results in less evil being done to people than a world where those networks never existed in the first place; and, in the process, Facebook’s total revenue and market share (even including those acquisitions!) are shrinking, because they shed huge amounts of money for each acquisition, and the time they spend “digesting” these companies is time their employees have to spend doing redundant work—making a beach-head into data-legibility of these user-bases, rather than actively exploiting the data-legibility of the users they’ve got. In other words, these acquisitions are revenue-negative for the first year or two, because they can’t immediately shove their hooks in, and so there’s nothing there to immediately sell to advertisers.
The Roman Empire collapsed because it got too big and didn’t have its distant member states under its control. It couldn’t tax them, or use them as soldiers in its wars. In theory, you could have made it collapse faster, by accelerating the rate at which it acquired these loosely-held fringe member states.
Or, for another analogy: your liver only has so much garbage it can process from your body at once. Give it too much to do—too many molecules to xenometabolize and “absorb” into safe bodily molecules—and the result is called “septic shock.”
Acquisitions are a poison. You can do one or two at a time, but even then, they’ll slow you down; slow down the rate at which you act upon the stuff you fully control. And if you do too many of them, you’ll just keel over.
Personally, I want to feed Facebook as many companies as possible. Ideally, feed it companies where the founders then leave and build more companies to feed it, with the aim of extracting unbounded amounts of resources from—and continuously poisoning—the behemoth. (Heck, these acquired companies have employees that get absorbed, too. Hopefully they’re vetted by the founders as following their ethical precepts. If they are, then the founders are essentially air-dropping good people into Eviltopia—good people who the evilcorp can’t just clean-sweep out without stalling the acquired company. That’s even more “poison” it has to deal with, and even more revenue-negative delay.)