The ideas in the book have evolved since that first edition.
I do not necessarily disagree, I still plan on doing FF, and putting all my cycling sessions into Xert. I would rather do the Sufferfest workouts than the ones in Xert. I trust the current state of the Sufferfest training plans more than the Xert advisor.
You always have to test a model, but eventually the cycling models will improve. At some point, Xert may have enough data about me. What I do not know is how good the model is, and what their improvement plans are. Nonetheless, compare sports analytics now with what they were even 10 years ago.
Of course there are things that analytics will not be able to model in the foreseeable future: motivation, pain (physical and mental) tolerance, etc.