Since I've accepted the conventional view, and taught it in my classes, I thought I should check this out. The research Blow mentioned is based on the 2012 American National Election Study. I checked, and there is no relation between age and opinions about relative intelligence or hard work. However, the ANES doesn't seem to have asked the question in previous years, so there's no direct information about change (I say "seem" because the ANES is a large and complex series of surveys, and I'm only moderately familiar with them, so I may have missed something). But the General Social Survey has asked the questions pretty regularly since 1990. Specifically, they ask people to rate how intelligent [hard-working] various groups are on a scale of 1-7, and you can compute the difference between the ratings for whites and blacks. Here is a figure showing average opinions among whites concerning intelligence: zero means that blacks and whites are rated as equally intelligent, and positive numbers mean whites are rated as more intelligent.
There is a definite decline over the period. You could argue that all of the decline was between 1990 and 1996 and there's been no change since then. But if you look at the percent of people who rate blacks and whites the same, there's been a steady increase--39% in 1990, 59% in 1996, 69% in 2012. So these results support the conventional view--they also show that younger generations are less likely to think there is a difference, suggesting that the trend can be expected to continue.
The difference between the ANES and GSS patterns is puzzling, but the evidence of a downward trend in racial prejudice is still strong.