You are aware that the Banal sequences all but obliterate that point, right? But I guess in your world, the Banal sequences are not real, or spill-back, or whatever tale you'd like to believe.

You know, I worked a little bit in directed evolution back in the day. It was for catalytic enzyme optimization using a secondary visual readout assay in a yeast system. Somewhat 'easy' because we could at least use the metabolite from the reaction and build the readout assay on top. Still took months and months to develop it before we could start screening, and the few residue changes we gained in the end.

All Childsplay compared to what you are suggesting they at the WiV did.

Viral infectivity assays are not straight forward; coupling them to set up a screening platform capable of performing at the scale you are imagining might be beyond what is currently possible. It is highly unlikely that the WIV had the expertise, resources, time and equippment to set up such an experiment; and why again would that not have been done on like a WIV1 backbone?