This is kind of alarming to me. A fine should scale with the the greater of the harm done and the amount needed to deter the behavior. But that's not what's happening here; instead, people are proposing that the fine scale with Facebook's total revenue. Even Facebook's most ardent critics probably don't believe they derive a majority (or really even a significant amount) of revenue from the lax API controls companies…
Shouldn't it be the lower of the two? If deter < fine < harm, then they wont do it. And if harm < fine < deter, then they will do it but everyone comes out ahead (assuming the fine money is distributed properly).