I wholeheartedly agree with your first sentence. However, the thought experiment feels vulnerable to bias and other factors unrelated to performance, (just like the salary negotiation process you're comparing it to).
The thought experiment indeed highlights the same dynamics: it's not about whether there are effective differences in quality in people (and thus different "valuations"), but whether the noise in the system is larger than the signal.
