I have also been reading this! My understanding is that people caution against using ACEs for individual clinical predictions because ACEs show population-level statistically significant effects on outcomes, but those may or may not be applicable to an individual case (since so many covariables also affect outcomes; e.g., personality, SES, education...). As I understand it, it's not so much that ACEs cannot be used as part of a holistic way of understanding a person and their possible clinical trajectory, but rather that too many folks were using ACEs as a sort of "end all, be all" for determining that a person was likely to have negative outcomes. The ACEs data have also been hijacked by the trauma cult to argue that everything and anything must be the result of childhood adversity.