So yeah, pretty much looks like the bitcoin is already kind of settling layer.
Any major credit card provider in the US provides this with no hassle. I've had people steal my number with 4 different credit card providers and the only thing I've had to do is call them and identify the bad charges. The only inconvenience is the 3 minute phone call and waiting a week for a new card.
Higher transfer rates, faster resolution. Only do "real" bitcoin transactions a couple times a day between players. Basically how all other high-frequency markets work.
In my view it's very simple: credit is extremely cheap to produce. If a retail merchant is willing to accept Bitcoin credit in exchange for his products, the consumer surely isn't going to insist on paying the higher Blockchain fee, and widespread bitcoin credit will become a thing. And credit (a promise) will always be cheaper than trustless systems (a guarantee), so I don't think the question is "either/or", but "how much do you want to pay per transaction?".
If you're willing to pay ~$0.15 there's no reason to not use the Blockchain right now. If, however, you would like to send bitcoins to someone and pay, say, 1/100th of a cent in fees, the only solution that works right now is credit (issuing digital Bitcoin IOUs in exchange for payment channel payments).
In short: it's extremely important to have cost in mind. Sure, if we could do everything trustlessly at the same cost as using a trusted third party, why not choose trustlessness? But this is not reality. Storing a public key and an IOU in a levelDB database (we don't even need SQL!) is many orders of magnitude cheaper than any trustless solution invented so far. I expect efficiency to improve for trustless solutions, but never enough to beat a trusted third party on cost.
If you're buying a cup of coffee using Bitcoin and paying $0.15 in Blockchain fees, you've spent more in fees than if you used VISA! Is it truly a necessity that we embed your coffee purchase in a global ledger and store it on 10,000 independant nodes spread out across the globe? Or could less suffice in this particular case?
It's also important to point out that an IOU clearing system on top of bitcoin would only introduce merchants to credit risk, not consumers. Consumers send real bitcoins to issuers (transferred trustlessly via payment channels), and get back an IOU with the merchant's public key on it. The consumer receives the product after handing over this IOU to the merchant, and will not (and should not) suffer if the issuer is unable to fulfill his obligation.
So, without a doubt, it is necessary that merchants be vigilant in assessing the credit quality of their accepted issuers. But if you're a merchant, you're most likely involved in multiple credit relations already, and adding a new one - in exchange for extremely low fees - may be a risk worth the profit. After all, these IOUs are instantly redeemable, so as a merchant, you choose how often you want to ensure your income, by redeeming IOUs into bitcoins on the Blockchain.