>So averages and social dynamics map exactly to individuals?

I don't think anyone claims that but patterns are significant enough that it is useful in mapping to individuals but within context.

When I'm interacting with an Indian, I can reasonably guess what their values are, what their personalities are and what they find important.

Same when I'm interacting with a woman.

Same when I'm interacting with an older person.

Of course, with context I'll learn more about the person, but I will not dismiss patterns that do exist and are largely true.

If you consciously dismiss these patterns you are just lying to yourself and robbing yourself out of meaningful starting points.

The correct word here is cultures, not races. South Asians of European ancestry (like one side of my family) are still culturally south Asian (albeit a particular subculture in our case). I know lots of people of non white ancestry who are culturally British (I mostly am myself).

I think culture is also a lot less dehumanising than race because it is complex and therefore less prone to grouping people in the same stereotyped way. Not all South Asian cultures are the same - there is a huge difference between, for example, an urban Sri Lankan (as I was born) and a rural Pakistani. Much as there are huge cultural differences within Europe (e.g. between Wales and Albania).

I do agree that stereotyping leads to dehumanisation, but I also think cultural differences are worth talking about. I also strongly agree that thinking in terms of race rather than culture is very damaging and encourages racism.

there are obviously drawbacks, but generalisations and assumptions are literally how we naturally store knowledge. its the quality of the information that is the problem not the generalisation.