While i agree that this reasoning is sound, we have to first define what Bitcoin is, is it a high value settlements instrument or do we wish to use it as actual money.
Yes, this is the area where intents (what we want to do with it) and game theory (how people will take selfish advantage of it) meet.
I see the current reward/fee structure as being only sustainable if fees rise significantly, which will limit adoption long before technological limits are reached. That's without bringing any kind of anonymity, privacy, fungibility or other *y terms into the debate.
Initially I thought block size increase could alleviate that, but block size is already limited by propagation and the risk of orphans, so larger block size, even if allow, would likely remain oddities.
In terms of raw technological scaling, a single blockchain is just not scalable for widespread micro-transactions: it would already take 10 MB-sized blocks to handle Starbucks sales f.i. (2 billion $ annual revenue, $4 per sell, that's 10k tx every 10 minutes if I did the math right), and that's just Starbucks...
That also gives another angle for the "how to reward full node operators?" debate, as nodes (or lack of, as in centralization) have a role wrt propagation times.
