Is the claim that IQs don't (even roughly?) measure intelligence or that intelligence doesn't vary between people?

Why would almost any intelligence test we derive seem to converge -- people who score high on one score high on all others and same with low/low results?