There are diminishing returns to exogeny, especially in species with a high degree of genetic redundancy (like dogs, cf https://pubmed.ncbi.nlm.nih.gov/17437958/), whereas with guided endogeny you can achieve better outcomes than you would via random cross-breeding (on task performance, health, or whatever else the breeders are optimizing for).
How is it still possible to give this as a suggestion rather than a rock solid fact? Every piece of evidence and theory screams of it.