Why don't we just quickly soft fork to disallow large OP_RETURN outputs (i.e. keep them capped to 80 bytes) only, with forward-looking activation (no retroactive block invalidation)? This would give the market a clear and simple way to decide whether such transactions are acceptable or not. It wouldn't have any of the issues mentioned so far in the discussion like creating an incentive to put CSAM or other illegal content to attempt a reorg double spend due to the complex deployment mechanism proposed in BIP-444 or like putting in jeopardy existing Taproot functionality by touching other primitives. We don't actually need to remove all the mechanisms that can be used for embedding data (which has been demonstrated to be impossible). Rather what we need is to give users the opportunity to signal whether arbitrary data is welcome or not. If the legal concerns are compelling enough to warrant consensus changes, miners should support a clean soft fork immediately. If the market quickly adopts the new consensus-level OP_RETURN restrictions it would announce sufficiently clearly to the world that arbitrary data is not welcome and deter future attempts to bring that to Bitcoin by showing the network's readiness to resist such efforts. I believe that this PR is too dramatic and needlessly complicated for what we need: letting the market decide whether to accept or reject the explicit invitation for arbitrary data storage that Core v30 has created. Technical changes should be limited to this question only and simple enough for people to make a decision without being distracted or overwhelmed by adjacent issues.