The fact that these techniques do work fairly well for semiconductor reverse engineering is a reason that recent research on "stealthy dopant-level hardware Trojans" was scary.
as well as some other microscopy techniques that might conceivably detect these differences, but that it's more difficult and expensive overall compared to optical imaging.
https://duckduckgo.com/?q=semiconductor+decapsulation
The fact that these techniques do work fairly well for semiconductor reverse engineering is a reason that recent research on "stealthy dopant-level hardware Trojans" was scary.
http://www.iacr.org/workshops/ches/ches2013/presentations/CH...
Though maybe that's what your reference to the inability to "seriously" inspect chips refers to. :-)
This research produces optically indistinguishable ICs with different electrical properties, which hinders optical reverse engineering.
My impression is that there's still a semiconductor device reverse engineering technique which likely defeats this measure (FIB imaging)
https://en.wikipedia.org/wiki/Focused_ion_beam
as well as some other microscopy techniques that might conceivably detect these differences, but that it's more difficult and expensive overall compared to optical imaging.