Notice how there's no mention in the whitepaper of non-miners relaying transactions or enforcing consensus rules. They were never part of the design, but we now rely heavily upon them. It's because of these nodes that scaling is a far more nuanced and delicate matter. Unlike miners, there's no financial reward for running a non-mining node, but they do help to secure the network. Ergo, making it more increasingly more costly to run something that earns no rewards is a difficult ask.
As someone who runs nodes, relays, bridges, and mailboxes for some projects, I think I can offer a personal perspective on the issue of rewards. For myself and several friends who share in these efforts, running a node transcends any monetary gain. It bestows a profound sense of expanding freedom and liberation in the world—a feeling so potent and invaluable, akin to love (which no amount of money can buy!). This is the essence of increasing total freedom: a united stand against the limitations enforced by governments globally. This feeling is priceless and motivates me to not only forego potential earnings but also to invest my own resources - I buy hardware and allocate both funds and time—all of which hold financial value—to support these projects. From my personal viewpoint, the 'reward problem' arises only when a project shifts from being a platform for the dissenting voices of the people, from a collective resistance to malevolence, to becoming a vehicle for wealth accumulation for 'HODLers' and institutions. After all, no one wants to labor for free while watching the 'rich get richer.' If it's no longer about a selfless collective standing against tyranny, injustice and state, it becomes a service—and services require compensation. Practically speaking, Bitcoin needs to consider implementing a financial reward mechanism for those operating non-mining nodes in the long term.
One cannot help but ponder why preemptive measures were not taken to mitigate the centralization of mining power, such as introducing deterrents to ASICs, similar to the RandomX algorithm?