Momentum is such that I would think a competing implementation will need to offer something new or interesting to attract users.
Yes, this is why I said perhaps optimized for watchtower functionality. Currently this is what you'd use Core for, but is the reference implementation the best place? Should Bitcoin Core be optimized for LN? Can you truly optimize for it if you're supposed to serve all? btcd could be that but it's insufficiently used.
If libbitcoin is able to ship their new version with some kind of really fast IBD, that might be enough to attract new users.
You only do IBD once though. So although that's been Eric's selling point, and it's super useful if you find yourself with some half-custodial wallet that suddenly doesn't work and you need a full chain to recover all your stuff, it's way more interesting for most people to know how this works in day-to-day operation. For this we need a final implementation.
Could a new node implementation compete on development process or review?
perhaps optimized for watchtower functionality. Currently this is what you'd use Core for, but is the reference implementation the best place? Should Bitcoin Core be optimized for LN? Can you truly optimize for it if you're supposed to serve all?btcdcould be that but it's insufficiently used.