Creating hybrids involves crossing to varietals to produce a third varietal, hoping for an improvement. It would be expected for the new strain to have characteristics that combine those of the two parents in some way, but it's not guaranteed.
Now, we put that on steroids. I would expect a random mish-mash of new characteristics along with some or all of the characteristics of the parents. That's what the steroids analogy does for me.
Genetic modification, in contrast, usually involves isolating the gene sequence for a specific trait that is desired in the new strain, then splicing the genes into an existing genome. Much more precise.
I propose a different analogy: if hybridization efforts were surgery, they would be done with a spoon. GM would then be like SURGERY WITH A SCALPEL!!! Horrors.