The author gives a brief overview of general design mixed models (MM) with emphasis on the use of the mixed model framework to fit and make inference for a wide variety of semiparametric regression models. It is demonstrated that the MM approach to smoothing has several advantages. The reason is that most models involving smoothing can be expressed as a mixed model and hence enjoy the benefit of the growing area of methodology and software for general mixed model analysis. Connection of the MM methodology with random intercept models, scatterplot smoothing, penalized regression, additive models, varying coefficient models, and multivariate smoothing is described. The author also covers such topics as: extension to generalized responses, hazard estimation, measurement errors, missing data and outliers, inference, model selection and diagnostics, Bayesian approaches and appropriate software.