Certain problems have an economic component that there is no way to engineer around. For example, suppose you said "I want to quickly onboard LN users and immediately provide them with balanced channel so they can instantly send/receive LN" - This is not a solvable problem at the moment given the economic component. LSPs would have to lock up huge amounts of sats to fund "onboarding the world". Unfortunately I don't think covenants will help with that class of problem.

There are certain problems that you can create a technical solution for, but results in complex UX/UI experience that may doom it to never being used. To be honest, I think many covenants fall into this category. Its not clear how its going to practically work.

Negative or flat outcomes. Any new OPCODEs will be used in ways thats hard to predict. Are we sure we are not just incentivizing some other problem somewhere else? (ie. we free up X% of blockchain space by using covenants, but enable Y% new blockchain growth by shitcoiners using new features? Is the net outcome worth it?)

Lastly, and this is the most nuanced, certain classes of problems inherently involve custodial relationships. When you give your fiat to someone to onboard into LN, you are already in a custodial relationship. You can't "atomic swap" your fiat note into LN in a trustless way....therefore if thats the case, then solutions like Liquid / Fedimint, Cashu already exist and provide a solution. In other words, the "last mile problem" is inherently custodial. Further, because of the economic realities mention in #1, its not clear how that relationship ever really becomes "non-custodial"? Yes if you onboard $1000 USD via a CEX you can convert that into non-custodial LN / mainchain money....but what about $5 USD? Point being: What are we even solving?