Very poorly. The only way it's usable is by relying on third parties, at which point you might as well use standard services. Or even actual cryptocurrencies for that matter.
My understanding of lightning is, that unless you run your own node (which requires a lot of funds and needs to be online 24/7), you are using a third-party node for these payments. The positive would be, that there are (I presume, don’t know) far more nodes/services available to choose from than there are credit-card firms.
This is just what I remember from reading into lightning a few months ago.
I'm curious about what made you reach that understanding. My guess is that you either tried or heard from someone who tried to open channels with some LN giant like Bitrefill. You shouldn't do that. Big channel operators don't want to dilute their liquidity with unreliable small channels. You should instead establish a channel to a smaller node that may be directly or indirectly connected to that merchant. LN is all about routing.
Also in the thread's context we are talking about Visa and Mastercard banning merchants right? In a scenario where the LN is widely adopted that merchant can setup and run BTCPayServer (https://btcpayserver.org) to maintain his own independent LN or on chain payment process system.
As for users, there has been work to facilitate nodeless operation on software wallets. Probably based on https://github.com/lightninglabs/neutrino. I know Blue Wallet is building something based on a lightweight LN node so people can establish channels from their personal wallets (https://github.com/BlueWallet/rn-ldk). There is an Android exclusive wallet that (https://lightning-wallet.com) can do that and Muun (https://muun.com) is doing some smart tricks to provide nodeless and custodial LN experience.
You don't need a lot of funds, or to always be online. You can run the software quite happily on a cellphone (I have previously), the requirement to be "online" is only in the scale of occasionally (once a day would be more than sufficient), and even that can be delegated (dork "watchtowers") with little to no risk.
I'm in a similar, semi educated situation as you are. But my interpretation of what I have read so far is different:
As far as I understand it: When a channel you have open is closed by the other side of the channel your wallet needs to know this within a certain time (24 hours? Can't remember).
So to avoid having to read the blockchain updates once a day, you would subscribe to some service that tells you when one of your channels is closed.
I also think you only have to do that if you distrust the other side of the channel. If you select a channel partner (a friend, a well known institution, a bank?) and only keep a small amount in that channel, I think you can just trust them and not sign up for any "watchtower" to watch if they try to cheat on you.
PS: Looking up the growth of the Bitcoin blockhain it seems that even if you do the watching yourself, that would only mean to read less than 1MB per day? Here is the data:
That provide liquidity, open or manage channels for you, watch your channel so you don't lose your money or even hold your keys (as is the case with some popular wallets people recommend).
And the network itself centralizes around large hubs (otherwise route finding doesn't scale).
The repeated comments about "routing not scaling" are sort of weird, there's no real technical hurdle to overcome. Approximations of route finding are really all that's necessary, there's no strong centralizing force to be found there.
Uh, this kind of route finding is still an open problem, and there's not even an idea of how to address it.
Approximations don't scale, as routes will start failing more and more as the network grows(due to liquidity issues). Therefore routes will go through well-connected and well-funded nodes, which is the centralizing force you're looking for.
