The fact that filters work at all when mining is as centralized as it is today (which is going to be fixed quite soon thanks to @npub1qtvl...7dze 's DATUM protocol and @jack 's work in the hardware/firmware space) means filters going to become insanely powerful in the future. This is why there is this push so hard to rip filters out now, because once spammers making out-of-band "transactions" must negotiate with hundreds of miners instead of a single pool, they'll struggle to get these gamified attacks against Bitcoin going as easily and they will remain something we can tolerate rather than horrible, runaway attacks like inscriptions. Individual miners adding spam is tolerable, that's what affords Bitcoin its censorship resistance and will always be possible regardless of what nodes may choose to relay. Right now however, you can evade a filter by convincing one giant miner (such as Antpool) to ignore it. Then it does so on behalf of tens of thousands of hashers pointing their machines to it. People don't understand that "sub 1 s/vb summer" is a product of mining being almost completely centralized. It doesn't prove filters are useless, it just proves how censorship-prone Bitcoin is because you can make massive, sweeping decrees about what ends up in the blockchain by dealing with one single entity responsible for 30% of block templates. And again, even with mining in this state, MARA still elected to abandon routing around a filter because the delay in block propagation was not worth it. They are a single miner and probably the biggest miner in the world - only outsized by pools like Foundry and Antmain who must soon be broken up into their constituent parts if their hashers stop being complacent and Bitcoin finally abandons centralized mining. They are economically incentivized to do this because FPPS is usually a waste of money for a hasher. It is significantly harder to spam Bitcoin in an environment where mining is decentralized - which it obviously needs to be. It's only possible to get persistent, gamified spam attacks if nodes agree to help the spammers out which they have no reason to do if the concern about spam filters causing mining centralization is a lie - and it is. Because the reason mining is centralized is not spam filters - so removing them isn't going to fix anything. The reason is that Foundry, Antpool and others are acting as giant miners on behalf of their hashers due to the mistake that is Stratum V1. We are going to get past this horrific mistake and it makes sense to plan accordingly, not to snatch defeat from the jaws of victory by tearing out filters in resignation to mining remaining centralized when it isn't going to much longer. Spam filters are a product of a network run by and for the benefit of nodes. Nodes having to relay spam is an admission of defeat - that pools will always remain centralized arbiters of what goes in blocks and thus can always be bribed into filling the chain up with crap that the nodes rejected from their mempools. It then takes this further and uses that as a rationalization for coercing nodes into trying to at least make the spam available to a few more miners to "level the playing field" among maybe a dozen miners instead of just a couple. Big whoop. The chain ends up trashed, nodes don't run for their own benefit any more and are now tacit free service providers for miners undermining their own investment. That is not a Bitcoin with a future and it's a completely unnecessary compromise. The purpose of spam filters is self evident for anyone running a node, even for a mining node. Excuses being made to justify their removal are shameful and insist that we can only ever have a Bitcoin where the Bitcoin miners of the world are mere hashers using the Stratum V1 protocol to evade their role in choosing what goes in the blockchain and that thusly, spam filters can surely be routed around easily enough that nodes must acquiesce. To reiterate: Spam filters have nothing to do with the centralization of mining, it is the poor design of the Stratum protocol which turns inevitably centralized reward split coordinators (aka "pools") into *miners* for the purposes of deciding what goes into the chain. That was never sustainable, and it's being fixed. If that gets fixed, then what nodes relay - and pretty much only what nodes relay - is what ends up in blocks as nothing out-of-band is sustainable if they can't find a miner large enough to grant them the blockspace necessary for their attack. If you only have a small group of entities deciding what goes into the chain and what doesn't, you're not decentralized or censorship resistant. But yes, you can route around spam filters trivially. Again, no one is proving what they think they are by paying a giant pool to evade a filter. The more decentralized mining gets, the more filters work. And I repeat: The fact that they work at all today with mining as centralized as it is and that Core must aggressively break stuff like datacarriersize for BitVM guys to have a shot at using OP_RETURN reliably instead of fake pubkeys should be incontrovertible proof that filters work as we say they do.