But from what I learned from this topic is the purpose of the removal of OP_RETURN limits is to let dick pics and fart sounds lovers use the part of the block where the data is prunable rather than using more "annoying" data storage schemes.
Just a little correction: AFAIK it's not a different part of the block. The reason OP_RETURN data can be ignored is that, once the complete transaction including merkle tree and TXID is calculated, the data behind the OP_RETURN opcode is not longer necessary for the local nodes's validation work. It is of course necessary if you want to maintain a node which enables other nodes to query transactions and blocks.
What I've read recently is that there is currently no node software which enables you to remove OP_RETURN data instantly once the TXID is calculated. So basically there's no advantage currently of OP_RETURN to the methods which encode the data in the witness, although both can be pruned in pruned nodes. There is of course still the 1000x times mentioned advantage regarding Stampchain and other fake-pubkey methods that this data isn't included in the UTXO set (which is a separate database).