Gradient boosting: the illustrated regression

Suppose we are working with a fragment of the California housing dataset, using the Median Income and the Average bedroom count as the features:

What are the values of $r_1$ (the first gradient)? Assume the loss function is $L = \frac{1}{2}(y^{(i)} - \hat{y})^2$.

Enter the values, with each value separated by a space.