Policymakers play the hand they're dealt; disliking a policy because you dislike the conditions it's predicated on is a valid thing to feel, but not an argument that the the policy is wrong or suboptimal. It would be better if the healthcare system were less burdened, but 'dramatically increase capacity of the healthcare system' was not a feasible response to the acute crisis of the last 18 months. Of course we should work toward that (and in particular toward making a more elastic healthcare system) but that's a long term solution, and under-provisioned in a crunch to over-provisioned during slack is two problems, not none.

The naturally immunity thing is fraught. Accepting for argument that naturally immunity is exactly as good as the best vaccine, it's reasonable to equate vaccinated and recovered at a point in time. Covid has done all the damage it's going to do to the recovered group, and both are as protected as the other going forward (per assumption).

However, offering 'vaccine or recovery' as an option going forward is more harmful if there are any remaining unvaccinated/uninfected people. Some _will_ choose the 'recovery' option, at immense personal (illness, possible long-term damage, possible death) and societal (health-care over-burden, transmission effects, etc.) expense. Splitting the currently-recovered from uninfected/unvaccinated for policy-making purposes would require either faith in self-reporting or significant invasion into health records. Neither seem tenable in the US, for obvious reasons of culture and law. IIRC other countries are taking that route.

Which leaves you, I think, where your comment started. Are the external costs of assuming the risk large enough that policy should proscribe that option and mandate vaccines?