It shows it doesn't matter, thus you might as well get rid of it to reduce the incentive to bypass the public mempool via centralized APIs to private mempools.
Sure, but that begs the question again since 83 byte op return as consensus rule this would solve the same problem just a well. Bringing us back to - what is the benefit we get from 100Kb op return vs what is the cost/risk of keeping or removing it?
🤙1