use even more bytes in the long run. We still pay per byte, so it's not like a larger OP_RETURN would directly cause heavier usage. Right now, I don't think anyone's worrying about it.
playing devils advocate to your devils advocate, what's wrong with stringing txs together like member does? If it comes down to overhead then that in turn comes down to cost.
Only that the overall amount of bytes used is more, & I think the main complaint about larger OP_RETURN is bloating. I suppose there's a sweet spot. Unintended Consequences.