[Mempool spam] Should we as developers reject non-standard Taproot transactions from full nodes?

In a recent email thread on the bitcoin-dev mailing list, Luke Dashjr criticized the lack of action taken to filter out spam transactions in Bitcoin Core. Dashjr argued that spam filtration should have been implemented months ago and it was a mistake not to extend the existing filters to Taproot transactions. He suggested that a more narrow approach, such as using OP_RETURN like "Ordisrespector" does, could be tried. Dashjr also pointed out that miners are making millions of dollars from these inscription transactions, and many people will continue to run nodes that do not attempt to block them. The current flood of BRC-20 transactions embeds very little data in the chain per transaction and could easily be adapted to use OP_RETURN or any number of other data embedding schemes. If they were modified to embed no data at all, they wouldn't be much smaller, but Dashjr believes that people would still complain that they are spam. This bugfix doesn't need to wait for a major release, and Dashjr's comments highlight the ongoing debate around how best to handle spam transactions in the Bitcoin network.