Interesting framework to reason about lock-ins. It's not only about depending on a vendor, there are other factors to keep in mind. Now obviously I'd see the risks around cloud hosting in particular in a different way than the author... but I guess we're biased differently there.