It's not annoying. Users are free to run any software they like. People can run Knots, then some more Knots, then more Knots and so on and so forth...

It doesn't stop the spam from being propagated around the network, and it doesn't stop the spam from reaching miners.

Some miners won't mine it out of principle. Some will, at least to maximize fee revenue.

The argument I do hear from knots-advocates... is that they need the entire network to run Knots or at least similar filters to them. Just 'some people' running Knots isn't sufficient for them.

So when an insufficient % of people run knots (let's say below 90% of the network) they complain. And complain a lot.

And even if somehow over 90% of the network were to 'switch to knots' it's not convincing to me it would stop the spam. The spammers can and would just use their own nodes with default settings and the transactions would end up getting to miners anyway. Just like the sub 1-sat/vbyte transactions... are 'getting through' despite the vast majority of the network filtering them by default.

So what happens then? Knots advocates will want a soft fork or hard fork to 'fix the spam'. They won't get consensus for it, and there will be a chain split.

Bitcoin Core (Bitcoin as it is today) vs Bitcoin Knots (Bitcoin mempools purified).

This stuff isn't that hard to predict imo