In nature, interspecific hybridization is an important evolutionary phenomenon that generates genetic variability in a population and fosters species diversity. With an increasing appreciation of hybridization in shaping today's biodiversity, a variety of computational methods are employed in practice to detect and infer hybridization. While some methods can perform such task accurately using genomic dataset, other methods cannot despite their high popularity, because they are not originally designed to detect hybridization. In this talk, I discuss the pitfalls of the popularly used computational tools, both network and non-network based methods, when inferring reticulate evolution. I argue many methods are misused and their outputs are misinterpreted. This practice must be halted to move forward in understanding reticulate evolution. Moreover, I suggest future research directions toward accurate inference of reticulate evolution using phylogenetic networks.