In a multiple linear regression model $Y = X\beta + \epsilon$ with $n$ observations and $p$ predictors, the residual sum of squares is $SSR = \sum(y_i - \hat{y}_i)^2$. If we add an irrelevant predictor variable to the model, what happens to the Adjusted $R^2$?