They don't have to go hand in hand. Others have pointed out eLife as an example of high prestige open access. There's no reason, other than history and inertia, that the validation has to come at the expense of access. However, I'd argue that I haven't yet seen a good model where the validation isn't expensive (by some definition of expensive). So I definitely think curation/validation can be delinked from the paywall, but I'm not convinced they can be provided free or cheap (and yes, the standard argument that reviewers aren't paid is valid, but the idea that publishers don't have real costs managing that process is simply naive).
