I still don’t think I would have done that, lnurl was a much more established way of fetching invoices and it made onboarding to services like WoS much easier. There is some fantasy scenario where not using lnurl would have improved non-custodial adoption, but I think it would have only hindered adoption and made it more difficult to onboard wallets who could care less about running a nostr relay.
Discussion
I don't think it's fantasy at all. Node package managers could easily have offered receiving capable NWC software (when they cannot do it for lnurl).
Wallets could have implemented the ability to pay to nprofiles directly.
Even mobile wallets (such as mutiny) can ran builtin NWC for sending (so why not for receiving).
Would it take more time? Yes. But doing things in the name of fast adoption is my definition of rushing it.
It’s only because of Zaps being demonstrated in practice that there was enough excitement and momentum for building other better tooling and for people to spend the extra time to try more complex setup.
Hindsight forgets the state of mind of people back then.
I recognize there was a demand for social micro tipping. Just disagree on the approach.
This is not just hindsight, check the initial discussions and contemporary proposals.
Even with hindsight, it's still important to reflect back on what could have been better.