Mostly fair comments there, however I would contest the emboldened section as I believe there are far too many people in the world or even just in individual countries to "distinguish between people based on individual characteristics". It's simply not practical to measure
everyone so we are left to taking samples of general characteristics and once again reverting to the ever useful tool of generalisation. I do have to disagree with you about it only being a tool for "humour, discrimination and laziness" however. Limiting the usefulness of generalisations in this manner seems to be a self-propogating fallicy - in that you "generalised" in a negative way that generalisations are negative... making your generalisation... negative.
Like you said - aren't semantics fun.