One thing for sure, it will raise the bar for entry for opening channels user-facing, definitely a boost for LSP adoption since 0 conf channels provide immediate liquidity without waiting for however many blocks to confirm. LSPs have the advantage that they can aggregate channel opens when they come in quick succession.

I think in the long run the mempool will need to do some new neat tricks to suppress congestion, the nature of which I am clueless at this point. Seems like it would need some more depth to the p2p/mempool protocol, the introduction of full RBF certainly makes it more tricky.

On the other hand, as has been discussed in recent Breez blog posts, aggregating liquidity in LSPs does sorta seem logical and mirror the way that network connections have specialists too (ISPs). It seems reasonable to compare the things together since the dynamics of the problem of creating connection paths are very much similar.