A nutritionist predicts daily calorie intake ($y$) based on hours of physical activity per week (x) using the regression equation $\hat{y}=1800-50x$. The standard error of estimate is $s_y=150$, the sample size is $n=25$, the mean hours of activity is $\bar{x}=5$, and $S_{xx}=250$. Construct a $90\%$ prediction interval for a person who exercises $x_0=8$ hours per week.