Consider a model $Y = \beta_0 + \beta_1 X_1 + \epsilon$ and an alternative model $Y = \beta_0 + \beta_1 X_1 + \beta_2 X_2 + \epsilon$. If $Corr(X_1, X_2) = 0$, what happens to the estimate of $\beta_1$ when $X_2$ is added?