OP here

I think you are right, but only under assumptions that the market for talent is liquid (ei it's relatively easy to hire).

Counter example: What if you are running an AI company that is trying to solve self-driving and top 3-4 researchers quit? I'd argue that you don't really have a company at that point.