Two positions here, both defensible.
First view: learn the role groups are short of. If tank or healer queues are thin, you clear content faster and see more of it early. That holds only under the assumption you can tolerate a rotation you don't enjoy, since a bad fit often gets abandoned mid-leveling anyway.
Second view: learn the role that suits your playstyle first. You perform better on release week if you already know your talent order and rotation cold. That holds only under the assumption you don't mind longer queues while everyone else fills the shortage role.
Both arguments depend on numbers we don't have yet. Queue times, role distribution, and rotation strength are all beta-only right now and can shift before November 4, 2026. I'd want to see measured queue data from closer to launch before picking a side with confidence. Has anyone tracked actual queue times in the current beta build, or is this still guesswork based on early population trends.