How to set up a self-governing business: outlining the “laws of robotics” for Hamster Marketplace

The idea for a decentralised trading platform for small and mid-sized producers of gadgets and electronics, which Denis Bulavin, producer of children’s tablet PlayPad, came up with, seems like the dream of an entrepreneur who was promised success in everything by a Genie.

Contracts with retailers too strict?

“— Well, let’s make our own marketplace, online, and with its own audience.”

This idea requires decentralising trust itself, without which it would be impossible for thousands of market users who don’t know each other to achieve self-organisation. It appeared at exactly the right moment: when blockchain had already been invented. But it’s one thing to successfully send, receive and save small packets of data. It’s another matter entirely to set up a multi-level global business within a two-dimensional system, including such powerful sectors as logistics, support and payments, while preserving its operating capacity. Could you live in a life-size sketch of a house, rather than at home? It turns out that the answer to the marketing inquiry of a project like Hamster Marketplace will be a fundamental technological solution, making it necessary to create a totally new type of management mechanism. In this case more than crypto technology is needed. All the same, you have to start working from an outline of the project on paper.

Isaac Asimov

What’s Isaac Asimov got to do with it?

What’s Isaac Asimov got to do with it?

Firstly, he’s one of our favourite science fiction authors.

Secondly, when we began to lay out the principles which a truly decentralised distributed system for managing a business should follow, it quickly became clear that we were forming our own laws of robotics. But in our case the chance of putting our ideas into practice is within sight, unlike Asimov, who created laws but didn’t live long enough to see robots appear.

As a side note, Asimov would probably have approved of the idea of a system that governs itself without outside interference.

Let’s get digging
Automatisation and decentralisation of the business management process can take root at very different levels: from the surface to the deepest level, which creates a fundamentally new way of managing a company.

The simplest method to put into practice is managing a company according to the principle of a joint-stock company, where the shareholders are the producers who are represented on the marketplace — once a year they choose the governing body of the company. The governing body in turn manages the company like a normal business, for a salary plus bonuses. In this case blockchain acts as a protection against corporate machinations with voting.

At the next level of immersion into the architecture of the business, blockchain can offer a reliable infrastructure for bookkeeping. According to the most recent EU directives, records in blockchain can be considered by the examining bodies as official valid documents. As such they can serve as an instrument for confirming honesty for the company’s shareholders, who want to be sure that the data given by management in the accounts is valid.

If the two points mentioned above are hardly new, then not a single business in the world has yet used blockchain directly to manage the company’s budget or allocate resources. Blockchain is necessary here not so much as a mechanism for transactions (it’s not going to be possible any time soon for a legal business to function fully using crypto currencies), but as a mechanism for transferring and taking into account orders to a normal bank. Orders in the form of smart-contracts allow the process to be additionally automatised and sped up on the one hand, and guarantee a new level of reliability on the other.

The fourth level, going as deep as Limbo in the film “Inception”, is management directly through blockchain. In this case the marketplace management represents something like a brain center, which accompanies the process of making suggestions. Suggestions are then accepted, becoming resolutions, or are not accepted and remain at this level forever with direct blockchain voting. This level is the most interesting, so let’s examine some of its principles in more detail.

How decisions will be taken
Decisions are the main element in any process, any matter, any system, any business. On the whole, any organised structure consists only of decisions.

A decision, in the sense of an order, decree, or command, is only one of the stages of the process which doesn’t actually begin with taking the decision.

Stages of a decision:

• Defining the problem. A decision is an answer, but in order to get an answer, you need to know the question. And a question needs foundations to be set. A solution to a non-existent problem is like multiplying by zero. Nonetheless, in any area of human activity there are always enthusiasts and even professionals at imitating the working process: they always seem to be furiously busy, working away where there is absolutely nothing to do. Any participant in the system of decentralised management will be able to set a problem, or even an outsider, if a member approves the incoming message.
• Validating the problem. In any system of direct democracy there will be come percentage of “uncertain” enquiries: “Walking along the office corridor, I saw that the walls were painted in two different colours. Is it a bug or an intended function?” Automatising the management process lowers the likelihood of such possibilities appearing, because the automaton needs an algorithm. And an algorithm requires a clear view of the problem, including criteria defining its presence and evaluation of the ensuing risk factors. If this is not done by the same person who formulated the problem, then it falls on the shoulders of the community. It goes without saying that hundreds of thousands of members do not all need to check the “guest book” together. The system can allocate it at random in a certain share — for example, 1% of participants, who in each actual case form a validation committee, to decide whether or not a problem should go forward. In this way, each actual participant in the system sees no more than 1% of incoming inquiries.
• Forming the inquiry. Very well, there is a problem: the validation committee has confirmed that it wasn’t only in your imagination. You have analysed the problem and assessed the risks. This is the moment to decide what comes next: what should we do about this problem and why? Usually the desired outcome already forms at the stage of setting the task. But this is not necessary — because it is a separate stage in the process of taking a decision, and means that in a decentralised system it will also inevitably go through a collective process. The problem finds itself open to general access, where any participant in the mechanism may:

clarify where the problem belongs (which department should resolve it);

suggest a desired result at the level of: “Well let’s repaint it! Let’s move house! Let’s demand that they lower the rent!”.

In the last case the quorum necessary for a decision to be made may already be more significant: up to 5% of the participants in the system, depending on which departments were proposed as responsible for the decision. For example, it is unlikely that for assigning the content department the task of correcting a typo on the web page, agreement would be necessary from more than 1% of system members or 2–3 votes in absolute terms.

• In most cases the department assigned to this problem should have the right to accept the problem — or dispatch it to be revised in the form of entering the necessary level — for example, the second, if it is necessary to re-validate the very fact that there is a problem, or to step three, if it’s necessary to revise the goal, having added their own commentary.

One of the departments in the decentralised system should be the department of arbitrage, which, for example, would resolve a situation where the assigned department endlessly sends the problem back.

The functions of arbitrage should include both the possibility of compulsory allocation of the problem, and impeachment of the team of a certain department with general voting.

At the same time, given that we are speaking about a distributed decentralised system, arbitrage as a mechanism does not necessarily even have to have a permanent staff set-up. For example, the most trusted system members could be invited in random order, who would give their consent to act in the “adjudicators’ role” from time to time (and at a frequency that they agreed to).

• Analysing the options. Any problem has a minimum of two possible solutions, one of which is not to solve it at all. And if it is to be solved, then the number of ways of doing it could increase to infinity. When the problem has been defined and the goal identified, the department to which it has been assigned should set about analysing the ways it can be solved. If the problem, for example, is “to buy a certain model of computer”, then the department’s duties will include choosing appropriate offers and, quite possibly, even the process of bargaining.

Multi-level problems such as, for example, hiring staff, can be split up by the department (in this case — with questions of job placements, or by the department which needs an employee) into intermediate stages with voting. For example, firstly defining the duties, then agreeing payment terms with the community.

• Taking the decision. The final element in resolving the problem. The table has been chosen, the employee found — the community has to say Yes or No, to hire or not to hire. The final decision differs from the intermediate stages of taking on a problem in that there is the option of “completing the task”. If the system member voting thinks that the decision they are voting for will resolve the problem laid out at stage one, then they can choose “complete the task”. If more than 50% of voters, like them, check the same box, then, in the case of taking their decision, the problem will also automatically be resolved.

Sounds complicated? Well, that is only the basic possibility of developing the situation. Actually, a multitude of secondary scenarios are possible. But if you divide the situation into its separate parts then taking a decision looks exactly like this for a person. It’s just that people make their own decisions though a centralised mechanism (if, of course, we’re not talking about the character from the film “Split”), so many stages of taking a decision occur immediately, almost unnoticeably.

In a decentralised mechanism for decision-making, each problem should be broken down into its simplest, most discrete constituent parts. If a question can be broken into two (to repaint the new apartment, or go off on holiday?) then it must be divided in two — otherwise it will be impossible to understand, however the community votes, which option they were actually voting for.

Therefore total decentralisation implies not a one in four referendum, but thousands and millions of mini-votes, from which the company’s decisions can be made.

However, the system members’ activity should not end up being an endless stream of micromanagement. There are several measures which overall lower the administrative load on each system member to the level not only of what is acceptable, but what is desirable.

• Different levels for resolving tasks demand different degrees of involvement from the members. The smallest tasks may need only approval from any two members. If there are, let’s say, 10,000 members in the system with the right to vote, then that means that no more than one in five thousand such small questions will fall to each system user. Various types of task-resolution do not require the work of permanent teams. Some areas of work (for example, arbitrage) may be covered as a one-off by an elected team.
• Any tasks which can be solved by algorithms should be solved by algorithms. The principle of dividing any problem into the minimum possible steps allows any question to be very clearly categorised. And if according to 100% of criteria it falls into a model for which an approved solution protocol already exists, then it will be dealt with automatically by the system, without input from members.
• The systems of ratings, evaluations and members’ self-definition allows the distribution of questions for voting to be made even more effective.
• And the system for defining the desired level of participation in management guarantees that the decentralised mechanism of management never invites members to vote on decisions which don’t interest them. There should always be the opportunity to define the category or level of importance of the task which the participant doesn’t want to miss a vote on. Members who don’t want to take part in resolving problems can totally switch off enquiries about voting participation, thus entrusting the fate of the project to the other members.
• Letting AI decide. It should not be forgotten that blockchain is not the most significant invention of the past decade. Self-teaching neural networks are a breakthrough on a much larger and more meaningful scale. Should the developers of the Hamster Marketplace decentralisation system ignore this?? Of course not. Each member of the management system will have access to the neural network, with which they will be able to delegate the solution of certain categories of questions, having tuned it, for example, to take decisions based on analysis of their own decisions or based on analysis of results of general voting.

The aggregate of these five factors — detailed specification of the problems, elaboration of the system of controls and counter-balances, tuning of the degree of individual participation in problem resolution, irreversibility of records in blockchain and use of AI — allow us to form a mechanism for self-government that is maximally effective, democratic, unobtrusive and at the same time genuinely decentralised. In a certain sense, we are talking here about creating collective reasoning, where each system member is a neurone, and each discrete enquiry is an electric impulse.

Testing, testing, and once more testing
Coming up with the principles of working with a decentralised managing mechanism is in practice not that complicated. What’s complicated is how to fine-tune it, to create a system that is ideally balanced from all sides, and which has no loose ends — no chain of actions which could come unstuck. For this you need:

• to build a prototype system;
• to wade through billions of combinations of possible development scenarios.

Here one neural net should generate scenarios of problems, and another should help to cover any vulnerabilities which are found. Thus an ideally balanced system will be protected not only from internal faults, based on the accidental or intentional exploitation of admitted vulnerabilities; to a certain extent it should also be resistant to external factors. While there are probably no algorithms that could defend against nuclear war, the system should at least be resistant enough to stay manageable and independent in the case of local cuts to lighting or the internet, or attempts at hostile take-overs by rival companies.

Because at the moment when the system of decentralised management of Hamster Marketplace is totally fine-tuned and launched and the platform comes under its control, there will be no turning back. Mince can’t be turned back into a piece of meat, and a record from blockchain cannot be wiped out just like that. But on the other hand, if the project is successful, its result — a system of decentralised management on blockchain and with elements of AI — will be much bigger and more significant that just the marketplace itself. Options for using a similar mechanism will mostly likely apply to more than just blockchain itself — from any association of people to running whole states. Fantasists, futurists, crypto enthusiasts and plain old anarchists have long dreamed about a functioning mechanism for direct democracy. But nothing like it will ever be made by the state or for the state — there is no motivation for state institutes to give up power.

But such a solution could definitely come from business, on the other hand, and even more so from trade. And the Hamster Marketplace team has every chance to be among the first — if we have your support.

Crowdfunding for the marketplace
In order to finance the creation of the marketplace as a functioning business, solving problems for small and mid-size producers of niche gadgets and electronics;

and to provide an opportunity for studying and testing mechanisms of decentralised self-management of a platform by producers who present goods on it,

Hamster Marketplace is launching a crowdfunding campaign in Fall of this year.

Selling tokens will enable us both to finance marketing expenses for promoting the marketplace and to support sales on it, and all necessary research and development to create an autonomous and reliably functioning mechanism of decentralised self-government.

Tokens, in turn, will be a part of the architecture of the platform. It will be possible to use them to pay for Hamster Marketplace’s own resources:

• vendors will be able to use them to pay for internal promo instruments and promotion, saving their fiat money;
• buyers will have access to limited shares, offers and early access both to goods, and to the platform itself ahead of its release to open beta.

Right now you don’t need to do anything. Just leave your email address on the Hamster Marketplace ICO website so you don’t miss the dates to register to take part in the ICO. You can also subscribe to our channel on Telegram or join our Telegram group in order to get first hand information.

P.S. If you are a seller or supplier of services or goods — please take part in a small survey for us.