Rather than subsidize "white" miners in this convoluted way, why not have users donate to ALL miners by stuffing data into op returns at a fee rate high enough and with a mempool backlog large enough to deter non op return data utxo bloat (aka spam).

if spammers bid more aggressively, greater good minded users can use cpfp or other mechanisms to increase their own bid.

I doubt many users would want to use wallets that forced to donate to an anti spam fund like this.

But they might equally not want to donate to the anti spam fund the other way. And at least this way doesn't introduce a state attack vector.

Adam back wants to shift utxo bloat control (aka anti spam aka censorship) to users / wallets, who would subsidize miners who block spam.

Blocking spam sounds like a good thing. But can we really tell what is spam? Won't the jpegs just keep getting sneakier and sneakier. and it becomes a spy vs spy red queens race and opens a censorship attack vector to boot?

I agree utxo bloat is a problem but idk if this is a solution.

If wallet subsidies push data into prunable op_returns, maybe that's a good thing.

But a) will users pay the extra fee, even if it's tiny? b) will wallet maintainers be able to resist state address and mpubkey whitelists when this is inevitably attempted?

maybe accepting centralization risk of utxo bloat is better than centralization risk of opening wallet filters as a state attack vector.