With powerful software, like SAS, applying machine/statistical learning methods to the analysis of data is really no more difficult than carrying out multiple regression with variable selection, logistic regression, discriminant analysis. 2. The emphasis when applying machine/statistical learning methods is generally more on prediction rather than explanation and inference. Indeed, with very large datasets traditional hypothesis testing and constructing of confidence intervals becomes meaningless, and the emphasis is necessarily on prediction and identification of variables that are important for obtaining accurate predictions. 3. In some problems machine/statistical learning methods are much more accurate predictors than traditional methods. For the mullein data from Lava Beds National Monument a classification tree (albeit a big one) provided substantially more accurate predictions than logistic regression models and Random Forests, a more sophisticated ensemble classifier, yielded spectacularly accurate predictions. 4. Machine/statistical learning approaches, including the simpler ones such as classification trees, can provide some different but very interesting insights into the structure of the data. The example of the lichen species Lobaria oregana in the Pacific Northwest was a case in which a very simple tree with 4 leaves fit the data well and was easily interpreted in the context of the data.