The article raises critical questions about the use of polynomial features in function fitting, a common practice in machine learning. It discusses the potential pitfalls of overfitting due to high-dimensional polynomial features and introduces the idea of orthogonalization and regularization as potential solutions. The commentary highlights the informative nature of the article and references related scholarly work, suggesting that the discussion is grounded in established research. Readers express curiosity about extending these concepts to other series like Fourier series and propose visualizing regression models in more complex dimensions. The interaction suggests that while polynomial features can be problematic, they offer opportunities for exploration in model fitting and regularization techniques that could mitigate risks of overfitting.