Running multiple LN nodes for the past year has been very educating on the viability of LN. LN works, it works well, when it is working properly. When things go wrong its a drain of your sats. I have no doubt that it has cost me many times more in bandwidth, hosting costs, channel closures and HTLC timeouts and onchain fees than I have made routing sats. You will not make money routing sats unless you have some serious liquidity. Why do I do it? It provides utility for me for small payments. Both my nodes are <100M sat liquidity. I like the tech.

The biggest concern for me the mainnet tx fee, which is an essential component of the security model LN. LN cannot function in the scale users expect with fees >100s/vb. As fees drift up with more spammers over time, channels will only be able economically be created by Exchanges, Banks and large LN vendors.

From this I am convinced that for the next few years at least we will have to accept majority custodial users from LN. WoS works so well, I have tried getting other folks to use self-custody channel management software, but it would fail in hard to explain to the noob ways. WoS works OOB, and works well. Next best is hosted nodes for non Bitcoin (or non Technical) business, with LaaS providers with JIT liquidity.

But technology does not stand still -- as you state, Covenants will accelerate multiparty Lightning usage, and I think we will see some major fintech players realize the profits from economies of scale in routing LN traffic and decentralize custodial hosting. The ability for a phone running 24/7 always connected lightning node is currently a limitation of the OS support today, which is limited due to battery tech. This will change in time with higher efficiency silicon.

I have no doubt that you will be running a highly efficient mobile LN node, or maybe some new L3 tech for moving sats in 2034.