So, spam filters work, but not completely. And the op_return debate is still reasonable to have? I'm still puzzled about what would happen in the case of continuously orphaned blocks.
Discussion
Spam filters 'work'... but they don't work well.
Yes they 'work' but not equally... and in their current state, especially with mempool 'filters' they only work on smaller miners. Large miners just create custom APIs and the interface to accept spam out-of-band to receive the high fees associated with arbitrary data.
Therefore the current 'filters' are broken... and in their current state tend to centralize hash-power rather than decentralize it.
In my opinion people saying that miners will 'decide' out of 'goodness' not to mine the spam is a pipe dream - they will mine whatever pays the most provided it is not a DoS attack (like unverifiable transactions).
Plus... decentralizing 'block construction' does NOT guarantee that miners "won't mine spam" if the spam pays higher fees. How does that make any sense? For Bitcoiners to *really* decrease the on-chain spam they need to make it more expensive to spam... and the only technically, ethically, and economically sound way to do that is *use* Bitcoin the money as money.
The more people use Bitcoin 'the money' the healthier it is overall.