Chrome is a pig for sure, but most attempts I see at measuring memory consumption are wrong. The two main ways they're typically wrong is either by misunderstanding how memory works (which is to be expected, it's subtle![1]) or by doing the measurement wrong (e.g. confusing CPU usage for RAM, or looking at the wrong process, or using a memory consuming extension in one setting but not the other). I think this post does both.
(Disclaimer: I worked on Chrome. I believe the conclusion "uses a lot of RAM" is correct, but it doesn't mean the reasoning that led to there is right.)