The question isn’t whether the heuristic is simple or complex. It’s who decides what’s monetary vs non monetary data. OP_RETURN limits made sense when they were consensus rules everyone agreed to. Post Taproot, inscription data pays full freight in fees and follows consensus rules.
The “arbitrary data burden” argument cuts both ways: if someone pays market rate for block space, that’s the fee market working. Configurable filters are fine for individual nodes, but marketing them as “spam protection”implies some uses are more
legitimate than others.
Bitcoin’s neutrality is its superpower. Once we encode “this data is worthy, that data isn’t” into standard policy, we’re making subjective calls about what belongs on a permissionless ledger.