I try to focus on making good decisions, decisions that don't lead to fatal outcomes even when the predictions are off. So allow me to modify your questions slightly before answering.

Can monetary demand for blockspace remain persistently low? Yes. Even in a world on the Bitcoin standard, sidechains can pick up enough load such that JPEG inserters remain a problem. We should be ready for a scenario in which demand remains low enough for long enough, even if we don't believe this is very likely.

For your second question, if enough node operators wake up then yes. (After all, where is the economic incentive for node operators to put up with spam?) Otherwise, there would be a centralizing effect due to increasing hardware requirements. Node-level filtering has basically been highly successful in keeping spam out for Bitcoin's entire existence minus the last two years. It should return to being so.