Minimizing the RSS involves making numerous iterations in attempt at approximating the best possible values of p1 and p2.
Here is our first iteration. The residuals, or differences between the data and the predicted values of the regression curve are large- as can be expected in a first attempt. Many of the residual values are dramatically less than 0, the value they should approach.
In our fifth attempt, some of the values apporach the regression line, there is a better mixture of positive and negative residual values. However, there are some values still too distant from the target value of 0.
What is our final iteration?