And that brings up another problem. Even if the price does go way up and the correlation between price/fee value isn't great, but it's still there.
The problem I show in my extrapolation is still very much present.
Regardless of BTC price, higher fees act as a disincentive to make transactions. So any transaction of value around the range of the fee becomes prohibitive (i.e. if you have to pay 5$ to transact, you won't want to transact anything near 5$).
Arguably, that lowers bitcoin's utility as a payment tool, which is the problem I'm focusing on here.
For the given dataset, if the data is accurate, then your point is valid.
However, the bigger issue right now is about controlling the fees. To be able to fix fees to a reasonable level, you'll have to increase the network capacity to a fairly large extent, way more than the current capacity (whereby it might not be as viable to host a node). If you don't increase the capacity enough, then there's a risk for the mempool to get stagnant at a fixed max fee rate whereby a price ceiling is implemented to control the fees. Since you can't let miners select the fees, then you're making everyone wait longer which still decreases the utility of crypto.
If you want to control the fees, you'll have to be able to increase the capacity first.