Payers have no incentive to ever include any fees, including today, because they are not the ones who actually care about double spending risk. After all, you know you're trustworthy, right?

It's actually the recipient that cares about confirmation. And recipients have many ways to incentivize mining. For example, via network assurance contracts, which I have proposed many times in various other discussions. If you see mining as a public good then assurance contracts are a method that's been both widely theoretically studied and implemented in practice (most obviously kickstarter, other examples are around too).

There is also the rather obvious and oft-overlooked fact that mining generates waste heat, and there are lots of people/places that actually need heat. Because mining hardware is small, portable and scalable, if you're already heating something with electrical resistance it may make sense to use mining ASICs instead and at that point you don't really care about the cost. You still need a full node or a pool to take part but I think that'll be easy to run even at high traffic levels.