One thing I think people have a hard time grokking is that even a moderate amount of hashrate filtering some transactions they don’t like won’t actually drive up the fees those transactions pay, as long as those transactions are relatively time-insensitive.
The market for less-time-sensitive block space is all one market. If one miner mines only a subset of the transactions available, a later miner will pick up the slack and normalize the ratio because they just take whatever pays the most (and those happen to now be the filtered transactions).
You can see this (usually) in the inverse - even though OCEAN filters a ton of transactions (including all kinds of weirdly specific rules that might impact normal transacting), they don’t actually make all that much less in fees - it’s all one market so the filtered transactions bid up the price because the non-filtered transactions don’t know who’s gonna mine the next few blocks.
This changes dramatically when there’s a huge demand spike for transactions they wish to filter (e.g. around the halving) where they may get paid materially less, but we’ve only really seen it the once.
There’s also the case of current OP_RETURNs paying more, but that’s mostly a side effect of one transactor (the OP_RETURN bot) submitting only to Slipstream, which charges more than other options (mostly F2Pool). This isn’t sustainable if there’s sustained demand, as other pools will want in, though there’s little reason to think it will be sustained once the drama dies down (all the people complaining about OP_RETURN are really driving a lot of demand for them lol).