I figured once banks were free of the regulatory nightmare surrounding bitcoin they would naturally become LSPs. They have the vendor, and intrabank connections and they LOVE taking transaction fees every five minutes so, it's a match made in financial heaven. The one down side is they have similar (though not complete) control over transactional freedom like they have now. The ability to close channels on people they deem undesirable suck but you could always go one hop away and connect through a friend who has a channel with them.

I am more of an economic nerd than the cypher punk variety but I do have an affinity for code and privacy. But it feels like more of these coder nerds need to read some Rothbard and Mises before commenting on the monetary aspects. (Not so much you but the "Needs more features" crowd)

Reply to this note

Please Login to reply.

Discussion

yeah, i read Human Action twice lol... does not need more features, just needs more routing algorithms, IMO... right now we have simple straight routes, we have keysend and AMP, it's lightning, where is the fan-out-fan-in multipath redundant routing!!!!???

to me, after examining all the info about the protocol that seemed like the direction that would be most important to explore, how to do redundant paths, and the other part, improving network state data propagataion, because the other problem is congestion - clients pick routes randomly from what they know was last known to be operational, but this whole thing of hiding channel state is a real problem, there needs to be some way to reveal enough channel state to help nodes making invoices avoid currently congested paths, this is almost the main problem, nodes being offline or unreachable is less of a problem than nodes getting too much traffic in one direction at the same time that they can't route

i like that... FOFI i think that would be the thing

if i were to contribute this as a routing algorithm to LND, which is a tire fire of buggy shit, but if i was, it would have to be based off the AMP, but AMP only solves the capacity problem, by parallelizing channels in a path, it doesn't solve the problem of route failure

it uses a cryptographic technique that enables the atomicity of the transmission, and the same thing should be possible to leverage to do multi-path REDUNDANT routing and thus enable failure resistance

this would kill the main weakness in the protocol that its haters point at the most

hell, i know exactly where to look to add this feature to LND, which is the main codebase deployed, but fuck those fuckers and especially fuck roasbeef... i tried to commit like 3 different PRs to get myself in the commit history fixing some legit problem that was in the issues, nope, nada, nix, not one got actioned, by the time someone noticed i was trying to help 3 more commits had clobbered parts of my commit and i needed to update it again... went through that cycle like 4 times with a simple one that isolated the generated protobuf protocols and that was enough for me, 3 times i am being really persistent and they fucked me the fourth time

that blackpilled me on the whole project

the only way i would even try to contribute to lightning at this point would be to be in charge of a LND fork and we would blow them away and within 2 years everyone would be running our node instead, even CLN users

Reading Man, Economy and State for the second time. Only like 500 more pages than human action it feels less poetic, if I can use that term.

Honestly from my perspective it always made more sense to me that lightning would have a more federated model. Small group of channels with the wealthiest member (or a pooled fund for outbound funds) connecting to each adjacent economic group. Then larger actors would connect adjacent towns, Then adjacent cities, and so on enveloping the world. Routing would nearly always go [Individual-Group-Town-State-State-Town-Group-Individual] for the longest hops. Maybe I just think to much like an econ guy and not enough in the practical limits of the code.

i tried to read ME&S but it felt like he was trying to just distill HA for dumb people, i think i got through about 1/4 of it before giving up

but yes, lightning is an open federated design, i can imagine building a redundancy protocol so that multiple providers could aggregate and fan out their infra to provide highly available big channel infra... it can already be done with plain simple network failover protocols that already exist, just a matter of tailoring one to do it with LN... the other thing it needs is a shared, highly consistent shared state data system so the multiple channels don't clobber each other's movements when a channel gets to the edge of capacity in one direction

see, that's the thing, i can see the network tech and the econ at the same time, really, it's just a matter of me finding someone with the money who is satisfied that i have the capacity to engineer it, and it would be my job for a decade to make this shit rock solid

but instead i'm here building a shitty middleware for a shitcoin gaming social platform, such is life

Maybe channel states could work like counting cards? Max capacity, minus X% equals +1 get over 5-6 and you become an "Outbound preferred" channel? I dunno if that is possible I'm not a code guy buy channel mode preference might be smart.

that's along the kind of lines i'm thinking, the thing is currently the protocol has near zero propagation of anything about channel state, the best you could do with a node right now is for it to literally just pretend to be offline but then it can't rebalance, as i see it, it just has to say what direction it's better to send payments through it

giving capacity info is probably a harder sell for the LN devs, but at least a sign of preferred direction would be great

Yeah, that's why I thought percentage would be more apt. 30% spend capacity could mean 30k or 3mil but it would signal to the network that it prefers to receive while routing. Again, I'm not a dev but this feels too obvious to have not been thought of or at least considered.

yeah, i'm not sure why there isn't a simple binary directional bias info propagated in the p2p node data

they really do hide the capacity though, and it's really lame because routing payments REQUIRES some estimate of at least the direction to send them over a path

Is this a feature that can be client side through LND or Core? Or would it have to be a whole BOLT iteration?

no, it's a peer to peer network thing

lightning is based on Sphinx https://cypherpunks.ca/~iang/pubs/Sphinx_Oakland09.pdf

it uses a source-routed onion encryption layering scheme

https://github.com/indra-labs/indra was my attempt to turn that into a whole network transport protocol, i did that work from late 2022 to 2023

one of the things that i figured out during that research and development work i was doing was that in order to make this into a network transport, it needed a very fast converging, but decentralised mechanism for propagating network state... congestion is the biggest risk, nodes being offline is the obvious but congestion is the real danger, and this is exactly the same dynamic with Lightning

people don't know what i'm doing, or where i come from, or what i do, it almost seems like i have the curse of Cassandra

Most people are just surviving. It's tough to focus on very beneficial tech if the benefit isn't immediately tangible. Cypherpunks have a bias toward the privacy and anonymity value. Most people see that (incorrectly) as a shady want to hide malfeasance. As an anarchist, I understand your feeling completely. So many problems that if people listened to my distrust for the government, would have been avoided.