Ryan’s original whitepaper, he has two, is as good a start as any. His invention from 2003 is a a natural evolution of the mutual credit concept, and as “decentralized” as it can get. Then there is an implementation of it that runs on a “central ledger”, like a public blockchain. It was technically an improvement to the first proof-of-concept that ran on Ryan’s server, but his system does not actually need any central ledger or consensus, at all. All processes are only ever between two people. This “central ledger” version of his system was called Ripple.com, and then there was trustlines.network too on Ethereum (effectively the same thing, his system + central ledger. ) Even Holochain is too much actually, no public evidence is necessary at all, no asymmetric signatures, you can just use thousands of years old symmetric key concept.
I guess I’m failing to understand
Maybe a better question is: if I go 5 credits into debt with you, and 10 credits into debt with person C, how does person D know my credit history to determine how much they’re willing to lend me?
how is this possible? I’m just curious the mechanism
this appears at first glance to be contradictory
I will give the whitepaper a reread. Do you have a link to the 2nd one?
What would be the best way to summarize how this is different from our current banking system of debit/credit accounting and tracking agreed upon credit score amongst participants? What prevents Sybil attacks?
what should I do if I don’t trust my government and/or national banks?
Unfortunately at the current moment, I’m not convinced they have my best interest at heart 
I’m very open to ideas and suggestions.
Another way to look at it, how can you know yourself to determine how much you’d want your friends to trust you with.
The mechanism is that all exchanges only ever happen over one hop.
Yes because the statements are about two different systems.
It takes trust down to a lower scale (to the lowest possible scale actually. ) This prevents sybil attacks as well.
Maybe
a) Tear down the system and regress to the middle ages
b) Learn to accept the limitations of the system as they relate to human condition
c) Flee the system,
d) Improve the system
I have worked mostly on d), makes me better at b) too. Don’t want either c) or a) although I did live by c) for a while.
But I thought it was multi-hop? Where I can I find more information about the multi hop mechanism that you described earlier?
thank you for the link
I do not see more detail about this in either whitepaper. There is nothing about KYC or ID verification. If I create a fake account with false information how does the intermediary double check this with verifiable records? Who is in charge of auditing participants? Where does user data get stored in this system?
“Servers might allow anonymous accounts identified only by a node ID on the social network, and possibly also by nickname identifiers, much like email addresses: node_nickname@server.com.”
Is this not a Sybil vulnerability?
Yes it is multi-hop. Trust is always only over one hop. So any oversight only needs to be per hop. Ryan’s whitepapers are both really good.
He might not have phrased the question exactly as you did. The point is that it takes trust down to the smallest scale, a relationship between two people.
RIght there is no KYC or ID verification in the traditional sense. You are assumed to know people you lend credit to but the terms KYC or ID verification are for systems with trust at higher scale.
It has no impact.
You “audit” the people you give credit to.
You store your own data ideally. In first implementation Ryan made, it was on his server, then there were the “central ledger” versions too, but those are all non-ideal. They have been good at promoting the concept though.
That’s on the topic of privacy from "8. How private? " Ryan Fugger is downplaying it all also. And, no, it is not a sybil vulnerability on the credit network. It might be on the data processing end of it but that would then be a non-ideal way to implement it. As I said a few times, you can just keep your own record of your own credit relationships, and pass on requests back and forth by just one hop. Nothing needs to be public. Just like how actual human relationships work.
In terms of money, I only lend what I’m willing to never get back. I use trust as a means of risk assessment.
I would tell my friends to determine it based on our relationship. If they loan me cash in person, then they know for sure who I am, but in the digital world this is much more challenging. When my friends know where I live, work, places I visit, etc then they can easily find me. They could ask other people who’ve I borrowed from whether or not I paid them back and how fast. If they get in my car as a passenger, I can show them my driving record. I can take a breathalyzer. Each subsequent ride they can see how I drive…
This is just my opinion, but irl (in real life) trust is much easier to navigate. In the online realm it is more scary. How do I know a social media profile is verifiable? Could someone not just download one of my photos and pretend to be me without security measures in place? Many online services use KYC linked to an authenticated database.
Trust is delicate. I could simply act honest for a certain period of time and then decide to use that against people. Personally its not my style, I prefer to act with integrity, but I can vouch for the effectiveness of this adversarial approach since I have been tricked by it many times. “Fool me once, shame on you, fool me twice, shame on me” lol
I agree with you that is impossible and unnecessary to completely eliminate trust. All systems contain avenues of attack. It’s an ongoing battle for sure! Just like the human immune system, it requires constant and consistent attention to health.
Yes these are interesting options to ponder. I was also leaning toward c for a while, but think d might be more helpful to others. I’m glad to see people working on it!
thank you 
Your lending is also your income. Since when you issue an IOU to someone else, it will only clear if they issue IOUs to others that end up forming links to people issuing IOUs to you.
It is probably good if you don’t trust people who are not in the real world.
That is not addressed by multi-hop mutual credit. A good way to know that though is the same way you know a irl profile is verifiable. By actually knowing the person.
Not addressed by multi hop mutual credit. Might be good to not spin off in thousand directions but keep it to mutual credit as this topic is about.
Depends on the person that you want to be trusted by. Humanity has survived so far. People are exceptional at reading manipulation, except when a society is entirely built on it.
Society currently favours Machiavellian intelligence but not all types of social organization do.
I think the nation-state/Ethereum approach is good too. It doesn’t eliminate trust, but it produces a very trustworthy central infrastructure that people can use. But I still like multi-hop mutual credit and similar grassroots systems more.
Multi-hop mutual credit is not that easy to attack actually. This is one thing I like about it, besides all the other things. It does not rely on publicly verifiable digital signatures and “computational hardness” assumptions those have, at all. It isn’t in the “crypto” (how people use the term “crypto” i.e., meaning publicly verifiable proof and not traditional cryptography) paradigm at all. Much less so than Holochain that still relies on public signatures. It is an ideal system, cryptographically.
definitely not my intention, only trying to use helpful analogies
yep I pulled that exact quote from section 8
this part is confusing
So I would store all my data on my private server using client-side encryption and then choose whether / who to share it with?
Does this mean other participants don’t know who I’ve transacted with and for how much?
Thanks for the answers, I think I’m getting a better macro view
Where could I find more information about this? Not sure that I agree.
I really like the Holochain design. It feels like a good fit for me personally, and I love the flexibility and freedom for customization. Plus the community is fantastic!
Well in multi-hop mutual credit you’re supposed to know the people you interact with directly. So it does not have the problems more trust-centralized systems have. Same with social media though, people are meant to know the people there too. Social links and credit links are same type of thing.
yes
[quote=“pqcdev, post:54, topic:807”]
The “trust lines” are between people who know and trust one another. Creating fictional people has no impact because they are not real and do not have any real human relationships.
Exactly how to do the computer architecture is probably something people have different ideas about. I think you could do your own bookkeeping on some machine somewhere you control. You can even do it in a physical book but passing on requests is then not automated (unless you have a mechanical turk working for you) so probably easier with a computer.
“Other participants” being people you do not have a direct 1-to-1 relationship with, I’d prefer a design where the social links and records are not public since there is no reason for them to be. If you then want to make something public to some group you could do so, but it isn’t necessary. This is because the accounting only matters on a person-to-person scale, no oversight is needed.
You are welcome.
Asymmetric cryptography was invented in the 1970s, partly to solve how to exchange encryption keys when telecommunication was becoming so widespread. The mathematics of it also allowed for the equivalent of a written signature, a public signature. But in multi-hop mutual credit, the signature does not need to be verified publicly, it only needs to be verified by a single person recieving the transaction. For this, public signatures are not needed. Might need what is called a message authentication code to avoid the very tiny attack vector of mutating some bit of the message but that is a bit perfectionist, just in general there is no need for any public signature.
You can use public signatures too but they aren’t needed. Public signatures rely on computational hardness assumptions that while they might be strong enough, a perfect cryptographic system is even stronger because you cannot know if you guessed right, you cannot brute-force it without spamming an actual other person with 2^bits messages.
I’m definitely understanding better, appreciate the comprehensive response 
Happy to hear. I invented a system on top of multi-hop mutual credit in 2012. This is one reason I care about it. It uses the credit links as “wires” to redistribute wealth person to person for guaranteed basic income. An incredible invention, based on just a single rule that scales organically to raise all people out of poverty. But this is also besides the topic of this thread, I find that moderators on different forums can sometimes get pissed if I stray from their focus so…
Thats great! It would be incredible to achieve a world without poverty and where everyone has their basic needs met.
Sure, definitely best to keep on topic. Thanks for the convo
-All the best
In terms of engaging with the existing system, the one I like most is a variant of C. Bucky Fuller said it best:
In order to change an existing paradigm you do not struggle to try and change the problematic model. You create a new model and make the old one obsolete.
Letting the new grow up amongst the old until it replaces it outright. This is exceedingly difficult, because the old system is always demanding its tribute. But I do believe that we can carve out small pockets of regenerative, non-adversarial economic activity and hopefully grow them over time. For my own part, I’ve been lucky enough to have lovely in-laws who are letting us stay at their place for pretty good rent. It’s allowed us to transfer a fairly large part of our economic lives from the dollar economy into the squishy realm of gift/reciprocity/trust. Cooperative gardening, shared meals, occasional babysitting. I know not everyone else is in this position, but I hope that our relative privilege will allow us to be in a better position to offer the same gifts to other people by virtue of the freedom we’re gaining.
Anyhow, we’re talking about mutual credit, so I’ll stay on topc. Just felt like there were ties to the human economy and my own experiences in it.
One thing I’m still not sold on is the whole ‘trust’ things. There’s no way for me to trust my immediate counterparties immediately; trust is like a multidimensional spectrum where I can trust someone 95% in some cases and 3% in others. The thing I struggle with re: trustlines is that I hope my friend/neighbour/aunt is being honest with me, but FWIU each of their trust pairing has its own ledger. This prevents me from having good visibility into their overall credit risk, which is an important factor in determining how much risk I personally want to take on by extending credit to them. While I agree that I shouldn’t extend any more credit than I’m willing to lose, the fact of the matter is that I’m also basing my decisions on the assumption that I usually won’t lose that credit in the end. And that requires me to have a broader picture of their level of risk.
That’s why I like the idea of each person having a single ledger – their own ledger – that’s publicly visible and protected from tampering by a ‘cloud of witnesses’. I do have a grave concern about this – privacy – although I think this is partially solved by @matslats’ Credit Commons idea. Small mutual credit networks, where members connect to people in other networks via ‘steward’ accounts operating in networks one level of hierarchy up. It’s mutual credit + multi-hop, something halfway between the polar opposites of a global mutual credit network and the atomised network of Ripple.
gotta love Bucky
and that’s eloquently stated Paul, I’d have to agree
Exactly! I only meant lose on an individual basis, while incorporating the proper risk of ruin. Should have clarified that my reference was to the Kelly Criterion. It’s best to diversify risk over a large enough sample size, which can only be calculated by having access to all the necessary information. The more available metrics the better, imo
EV = SUM i [ net((% chance)(amt gain)) - net((% chance)(amt loss)) ]
No worries; I understood you as being concerned about lack of access to good data about other people in a trustlines-based network, and was agreeing/continuing on with a comparison to community mutual credit.
I think what I’m seeing here is that there is no perfect silver-bullet solution. Each solution optimises for some things at the expense of others. Trustlines are wonderfully elegant but don’t offer much built-in support for collective intentions, of the sort Matthew Slater describes. (I don’t say it prevents it; it’s just that it doesn’t have that principle baked in, so you have to do extra work.) Community mutual credit, on the other hand, is more fiddly but embodies the idea of the collective in its very design. And that fiddliness pretty much necessitates a bit of extra complexity such as a centralised server or some sort of peer-to-peer witnessing protocol (Holochain or blockchain).
Each one for its purpose; I have strong philosophical attraction to collective, bottom-up action that’s more intentional and overt than the ‘invisible hand of the market’, so I’m sure you can figure out which one I lean towards 