And as I said before in this topic the biggest problem here is similar to the issue we had with Ordinals Attack, it is the fact that nodes do not have a say in what they relay. A handful of developers decide! That's moving toward centralization.
Miners are ignoring standardness rules, what "developers decide" does *not* limit the junk getting included. Efforts have been made to convince
themthese miners otherwise, but surprise the like the hundreds of millions of dollars in fees they've gotten by mining what people pay them to mine.
Bitcoin is designed from the get go to be censorship resistant and the result of this is that no one gets to decide how other people can use Bitcoin generally, because of this filtering stuff is generally a losing proposition.
... and if there were a way to make it effective then it would just be a mechanism that could equally used to impose government blacklists or similar.But ordinals whatever doesn't have anything to do with the topic at hand, it won't be changed by this.
Instead, there is some traffic stuffing data in outputs its not possible to stop that. Instead, it could be opreturn which would at least keep it prunable. But even if op_return were worse, though I don't see how, the limit is ineffectual for anyone willing to pass the transactions directly to miners (promoting centeralization along the way and encouraging miners to bypass standardness).
So there is a limit. It doesn't stop people. It does encourage more harmful encodings that cannot be blocked. And because there is a gap between what miners mine and what nodes relay this hurts Bitcoin's decentralization by encouraging direct miner relationships and slowing block propagation (which benefits large miners over small ones). So the limit is no longer useful and causes some harms-- why shouldn't it be removed?
The fact that you also hold that not enough is being done about jpeg spam isn't relevant to this particular question, since it's not something the jpeg spammers care about or will care about.