Tim Oren has a great thread on Reed's Law that shouldn't be missed.
I agree: a naive interpretation of Reed's Law is one of the things behind 'delusional valuations'. If you have no idea what this means, imagine multiplying the future revenues of a firm by the unbounded exponential growth of a network.
But that's why VC's absolutely need to talk to economists more: generally speaking, when econ models incorporate networks, there is always
a negative term to saturate and limit the otherwise monotonically increasing relationship between network size and utility. This creates a U-shaped relationship implying an 'optimal' network size. You can grab some of Economides' papers for a demonstration.
If I think about it, Reed's law isn't even true to biology - where bounded logistic functions are used to model the growth of populations, epidemics, etc. These obviously have limiting terms as well.
So, IMHO, Reed's law is a new interpretation of old concepts in science - and the older concepts is truer to the complex dynamics of the real world if ya ask me.