The article discusses different methods for estimating parameters in nonlinear regression models with errors. It explores least squares estimation and robust M-estimation techniques, especially useful when error distributions have certain characteristics. The researchers show that robust M-estimation can be more efficient than least squares in some cases. They also introduce a weighted robust M-estimation method that doesn't require specific assumptions about the distributions of errors and regressors. The study demonstrates that these estimation methods are consistent on a range of scaling parameters.