But the end result is effectively the same. If you throw in the constraints of what GP mentioned about customer retention, at the Pareto frontier it boils down to the same optimization, just that instead of manually optimizing the specific variables they become latent variables. There is no difference in the resultant enshittification.