Leia o fragmento a seguir, extraído de um relatório técnico sobre sistemas de recomendação baseados em aprendizado de máquina:
"While early recommender systems relied on collaborative filtering, the shift toward deep learning has not simply replaced one algorithm with another; rather, it has reconfigured the very notion of relevance. Where collaborative filtering inferred relevance from aggregate user behavior, neural approaches often infer it from latent representations that are opaque even to their designers. This opacity is not a bug to be fixed but a structural condition of the technique, which complicates accountability regimes that presuppose explainability."
Considerando o fragmento e as estratégias argumentativas nele empregadas, a alternativa que melhor sintetiza a tese central do autor é:
Selecione uma alternativa pra criar sua conta e salvar o progresso.