Introduction
Here we will present the basic concepts of adelphos without going too much in detail, just to highlight the key concepts that make it different from other similar systems.
The interested reader can find more details in the linked pages.
Money
In adelphos users exchange goods and services using a virtual currency based on trust and credit called AdelphosCoin AC which is not convertible to any real currency. See Adelphos Coin for details.
Trust and credit
In adelphos each user issues some money in exchange for a good, there is no central mint, nor a private mining like in cryptocurrencies.
The system, however, limits the amount of credit / debit that each user can get / can give.
Trust and credit.
The AdelphosCoin
Let’s return to the example of the library book; that example worked because there was trust between the participants and there was no money involved in the overall transaction.
But in real life?
* Would you trust to bring an object from a person you know to a
person you know?
* Would you trust to sell an object giving it to a friend knowing
that it will reach destination?
So, Adelphos is able to do the routing, but how to put it into practice? In a world where people do not trust and also real carriers with a centralized logistic have complains, losses and delays, how to make it work without a centralized brain and, most of all, without real gains (not in real money, however?)
With money things become complex because the transaction can happen if the seller can receive something concrete in exchange.
Let’s think about a used phone, for example. Alice wants to sell a used phone and she wants 80€ from it.
In the “usual” way Alice puts an ad in one of the sites that allow users to put ads; there are many of them, there is also a version in the fediverse, called flohmarket.
Bob sees the ad, the price and the shipping cost, let’s say 5.00€. He thinks that this is a deal, he contacts Alice, sends her the money and the shipping address. Alice does the packaging and sends it.
Eventually Bob receives it and, in some platforms, can leave a feedback to Alice.
This works because Alice and Bob agree on the nature and value of money and Alice knows that she can use those 80€ for something else, also outside the exchange platform.
But Adelphos is different. It wants to create a different economy, based on trust and credit, not using centralized currencies based on debt issued by Central Banks.
But experiments like this have not had too much success; I think that the main obstacle to a widespread adoption of a virtual currency is the lack of trust and the fact that they rely on an external entity to do the routing.
But the main obstacle is trust.
Suppose that Alice agrees to use Adelphos and puts the ad into it, Adelphos uses an internal currency, called AdelphosCoin (AC), she puts the ad and puts the price 80AC.
(For now we have used the simple equation 1AC = 1€, but this is not fixed, you can see in Adelphos Coin some more details about the relationship between AC and the real world currency.)
Alice puts the ad in Adelphos if she trusts that:
* she can spend later the 80AC into something else
Adelphos gives the used phone to Bob if it trusts that:
* he will later contribute to the community by putting into the
system something of equal value.
From this simple example we see that trust is bidirectional, from the user to the system (the user trusts the system to have something back for that amount of AdelphosCoin) and from the system to the user; the system, in the form of the overall set of users, trusts the user that he will repay the system (not Alice in particular) by giving something back.
Adelphos models this scenario dividing the trust in two ways: trust from the system (in-trust) and trust to the system (out-trust). Both trusts are independently updated and govern the overall behavior of the system.
But how can a computer measure an aleatory value like trust which,
like love and friendship, escapes a formal definition? Computers do
not have friends, they need an algorithm, clear data to compute, we
just cannot tell a PC Alice trusts Bob'', we need to tell ithow
much Alice trusts Bob?’’.
Let’s return to the phone example, because the key is in the credit.
Alice maybe does not trust adelphos for 80€/AC (given the equivalence 1€=1AC), at least initially, but suppose that she has a used book that she does not use any more. The book has not any real value, maybe 1 or 2€. Probably she will trust to sell it to adelphos, because the risk is less.
We can say that the trust of Alice is a number between 2 and 80.
It also depends on the person to whom Alice is selling. If she gives the phone to her brother Bob, maybe she does not want to be paid, because maybe later in time she will need Bob’s help to configure her laptop, for example, or to repair her scooter.
The logarithmic nature of the trust.
So, Alice trusts Bob for 80€. But how about a Bob’s friend? Let’s suppose that Bob has a friend, Carol, that would like to buy Alice’s phone. Will Alice give to her 80€ in credit? Probably not.
But how much? 60? 40? The problem is that escaping the circle of the family trust drops sharply, that is why usually these currencies based on credit are not widely adopted.
Yes, of course, there are fuzzy system, AI agents, etc…, but from the start I wanted to create a system that could run on a very simple laptop, with only 4 Giga of Ram, no graphic card, to be able to run it on a solar panel also in a remote area with an unstable internet connection. Current AI systems are very hungry on system resources and/or memory.
Adelphos is a system that wants to help to expand the trust of users, but to do that we need to measure trust because computer do understand only computable things (even the current AI systems transform the language into numbers called tokens, but this is another story…).
There have been in the past some experiments to create a web of trust, but they too are mathematically complex; the simple idea that I had in designing the trust computation in adelphos is to note exactly how trust feels in real life.
It is very high in the family and drops sharply; it is still high for first order friends, maybe not so much for second link friends (friends of friends) and then rapidly goes to zero.
Suppose a simple example: would you accept a friend to sleep on your couch? Alice is one of your greatest friend, she happens to pass by wants to sleep one night at your place. You probably will say yes, and you could be happy too for the visit.
A friend of friend? You know Alice very well. Alice asks you a bed for her friend Julie. You DO NOT know Julie… would you accept her? Maybe yes, maybe not.
A friend of a friend of a friend? Probably no. Unless you plan to be paid for it. In that case the absence of trust is compensated by the promise to gain some money.
This simple example shows that trust is logarithmic in scale, not linear. In physics logarithmic scales are used for example for sounds, and are called deciBel. A decibel is actually a quantity obtained from a value Q using this formula
$ q_{db} = 10 \log ( \frac{Q}{Q_{0}} ) $
Where $ Q_{0} $ is some reference value.
In this case a factor of 10x increase will translate into the addition of 10.
Adelphos uses the same principle to measure trust.
$ Q_{0} $ is 1 Adelphos Coin.
So the trust of $ 0_{Db} $ is actually a trust of 1 Adelphos Coin
This because $ 10 * \log ( \frac{1}{1} ) = 10 \log (1) = 10 * 0 = 0 $
A trust of 10 Db is a trust of 10 Adelphos Coin
A trust of 20 Db is a trust of 100 Adelphos Coin
And so on.
In-trust and out-trust
We have talked earlier about the dual nature of trust; the trust from adelphos to the user and from the user to adelphos.
In this paragraph we will define them better.
In the previous paragraph we have equated trust with credit, and in adelphos the trust limits the amount of credit / debit that a family can have.
Out trust
The out-trust is the simplest of the two: this, in fact, is usually the normal way in which we feel it: we feel the amount of trust we put in one person not that trust that the person puts in adelphos.
In a certain sense the out-trust is not really needed; if one person wants to sell an object of value without any money he should be free of doing it. But adelphos enforces it because it wants to exclude possible disappointments.
Adelphos wants that the users build slowly trust into the system and be loyal to it, instead of having a user that starts with a high trust, and feels robbed after.
In trust
The in-trust is the opposite: it is the trust of the system in us, so it limits the amount of things we can buy without selling anything.
The in trust should forbid the behaviour of a person who login into adelphos and starts to buy things without anything in return dropping in debt.
The limits and the trust.
The main rules of adelphos that link trust with credit are:
The out-trust limits the credit
The in-trust limits the debit
Every actor in adelphos has a balance, we start from level zero.
The change of trust in adelphos.
The trusts cannot be modified by the user, however; it is Adelphos who, based on a particular algorithm, will modify them after each successful transaction.
We will talk about this later when we talk about the commerce of goods and services and the feedback mechanism see here
The Brotherhood
Adelphos wants to expand the mechanism of families to upper levels, to families of families and so on.
One of the key concepts of a family is the brotherhood; well, this is in some way a definition of the family: we distinguish friends from members of our family just because we feel that members of our family are closer than friends.
Closer in what sense?
We do not want to go down in the physical and emotional sense, but we remain on a material part; husband and wife, also for the civil law, need to love each other, be loyal, but, above all, to contribute together to the common good.
The common good of the family is one of the greatest values of a family, each member is free to pursue his/her interests, hobbies and inspirations but should do all work with the common good of the family in mind.
Adelphos limits to model this aspect of the family, which is important and can be summarized as follows: in adelphos members of the same family share gains and losses.
This seems obvious in level zero, in a certain sense it is obvious that husband and wife share the same account, but adelphos is fractal in nature, so it tries to put the rules of the family also to upper levels, this means that also in upper levels there is a sharing of gains and losses.
But we have to be practical: even in a normal family there are personal goods which are private for the husband and a wife, so also adelphos remains to a practical level.
Adelphos defines a brotherhood constant called Beta $ \beta $, which defines the ratio of the gain that the family distribute to other members and the gain that goes to the seller.
Yes, in adelphos when you sell an object for 50 AC, for example, it is not true that you will receive 50 AC, you will likely receive less because a part of your gains goes to other families, maybe also to people you do not know.
But this is also contrary! If you have 10 AC in your balance, you might buy an object for 20 AC or more (it depends on the topology of the network and by other characteristics) because people you might not know put their credit to help you buy.