As for the 6-8$ of electricity, naturally, tx/s, fees and btc price will rise to compensate. In 4.5 years (post halving for 6+ btc per block) we may be doing 20-50 tx/s instead of 3, and in 8.5 years (post the 3+ btc per block) we may be upwards of 50-100tx/s, on-chain.
This wasn't exactly your main point, but there's no certainty that fees will rise to compensate for the loss of subsidy. Miners can also simply go out of business due to unprofitability. The question really being discussed now is: "What kind of usage of the blockchain will lead to the best income for miners, and how should we decide what usage we aim for?"
This because ultimately the miners decide what kind of network they secure.
"What kind of usage of the blockchain will lead to the best income for miners" i think the answer is crazy simple,
infinite usage , the more the usage the better.
pretty sure no matter how to slice it, even if miners require 1GB/s internet connections and high quality hardware, because of huge blocks, the more usage there is on the network they more money they will make.
pricing the fees is relatively easy to, put the low enough fee as to include 90% of the TX but high enough and exclude 10%, you want the fee high enoght as lose that 10% of TX's so that you know the other 90% are paying as much as you can get them to.