Until a few months ago, no one talked about satoshis per byte. For years now transactions have been so cheap that you could have addresses with lots of tiny inputs, and it would still cost only 0.0001 or 0.00005 BTC to send. Since I was just a user sitting behind the driving wheel, clueless as to what an alternator or a distributor was, I paid absolutely no attention to the size amounts coming into addresses. Now suddenly, the truth comes forward, and it can now cost tens of dollars to move out smallish amount of bitcoin in an address. It can also cost a still reasonable amount, if you only have one input in an address that you're spending.
Anyway, now we all know better, Satoshis/byte is a real world thing now, and smallish amounts of BTC will be sent less and less frequently, sending bitcoin on chain will not be for tiny amounts anymore, fees will be more reasonable once this turnover of small inputs finishes, and we'll all adjust our usage habits.
everything you said is due to CORE removing many fee control mechanisms.
the cost can actually be eleviated by bringing in a NEW 'priority fee formulae'
which the fee rises not by how much you spend. but more so how soon you want it confirmed.
take this one as an example
here is one example - not perfect. but think about it
imagine that we decided its acceptable that people should have a way to get priority if they have a lean tx and signal that they only want to spend funds once a day. (reasonable expectation)
where if they want to spend more often costs rise, if they want bloated tx, costs rise..
obviously those wanting to spend many times a day can benefit from LN
which then allows those that just pay their rent once a month or buys groceries every couple days to be ok using onchain bitcoin.. and where the costs of trying to spam the network (every block) becomes expensive where by they would be better off using LN. (for things like faucet raiding/day trading every 1-10 minutes)
so lets think about a priority fee thats not about rich vs poor(like the old one was) but about reducing respend spam and bloat.
lets imagine we actually use the tx age combined with CLTV to signal the network that a user is willing to add some maturity time if their tx age is under a day, to signal they want it confirmed but allowing themselves to be locked out of spending again for an average of 24 hours.(thats what CLTV does) to win lower fee
and where the bloat of the tx vs the blocksize has some impact too... rather than the old formulae with was more about the value of the tx
as you can see its not about tx value. its about bloat and age.
this way
those not wanting to spend more than once a day and dont bloat the blocks get preferential treatment onchain ($0.01).
if you are willing to wait a day but your taking up 1% of the blockspace. you pay more ($0.44)if you want to be a spammer spending every block. you pay the price($1.44)and if you want to be a total ass-hat and be both bloated and respending EVERY BLOCK you pay the ultimate price($63.72)note this is not perfect. but think about it