I'm pretty swamped with the Clements update at the moment - would be good to get a better system for bird hybrid observations in place - I'm not convinced adding new hybrid taxa to the taxonomy for every combination of closely related birds is the right path

something like "ID to genus and fill out the hybrid observation field" or something might be a better approach. We just had to modify the computer vision training flow to ignore bird hybrids because it was causing lots of weird behavior in the training process and model prediction output