The righttoown

Discussion of AI often gets stuck between two fairground stalls. One sells the end of the world: “everyone will be replaced.” The other sells permanent prosperity: “technology has always created new jobs.” Both formulas spare us the effort of looking at actual work.

Work consists of tasks. Some are easy to describe, check and repeat. Others require context, trust, responsibility and the right to decide. There is no concrete wall between them. Today a model drafts; tomorrow it checks a hypothesis; the day after, it directs a chain of tools. Meanwhile, a person remains responsible for meaning, consequences and the unpleasant moment when the system's confident answer turns out to be confident nonsense.

The International Labour Organization's 2025 research assessed exposure at task level, rather than drawing up a list of future redundancies. It found that roughly one in four workers is in an occupation with some exposure to generative AI, while transformation remains more likely than full replacement. In 2026, the ILO explicitly reiterated that exposure indices indicate what technology could do; they do not predict economic feasibility, adoption speed or eventual employment.

This qualification matters. It offers no reassurance, but brings the conversation back to reality.

If half a role's tasks become cheaper, the role may survive. Its price, composition and place in the organisation may still change. Someone who can only produce a standard artefact is vulnerable. Someone who connects a business problem, solution architecture, a team and responsibility for the outcome becomes more needed. Not invulnerable; more needed. Employment contracts still do not include immortality.

The first principle is therefore simple:

Rather than defending an old list of duties, strengthen your ability to understand a problem, make a decision and answer for the whole route to a result.

I use AI in personal projects every day. It helps me research, design, write code, find contradictions and keep several streams of work in view. But between “the model suggested it” and “the product works” lies everything a demonstration usually leaves out: boundaries, data, security, tests, observability, recovery, the cost of mistakes and the person who gets called at night.

Skip that part, and AI does not shorten the delivery cycle. It increases the rate at which unverified decisions appear.

A well-designed process works differently. A person sets the direction and criteria; the system speeds up exploration and production; automated checks catch some errors; responsibility remains with someone who understands the context. Productivity rises. Then comes a question usually kept offstage: what happens to the result of that increase?

In a typical employment arrangement, the organisation finances the work, bears business risk and receives the main economic return. The employee is entitled to salary and other payments provided for by the applicable rules. Rights in a particular product, however, are determined by law and contract: under Russian law, the exclusive right to a work created within employment duties normally belongs to the employer, while authorship and statutory author remuneration are separate matters. Employment alone does not give someone an equity stake or an owner’s vote.

The problem begins when the only agreement available to a person is presented as a law of nature.

If a tool lets one specialist do the work of a former team, the gain can be used in different ways. Reduce staff. Build more product. Shorten the working week. Improve quality and tackle tasks that previously did not pay. Technology does not choose. Rules and those with a voice do.

The second principle:

When discussing AI efficiency, ask not only “how much faster?” but also “who gains, who bears the risk, and who decides?”

I have been fortunate to work with people with whom disagreement does not mean war. The business and I are not on opposite sides of a barricade. We have different professional perspectives, so the same deadline may look reasonable to one and fantastical to the other. A manager's task is to find a shared solution rather than crown the winner of the argument.

That is why I distrust the convenient caricature of a heartless owner above and a noble, suffering engineer below. In reality, those at the top also make mistakes, feel fear, answer for money and depend on incentives. Engineers can also defend a favourite technology instead of a customer's need, hide uncertainty and mistake complexity for depth.

People may be wonderful. Good intentions still do not replace organisational design.

A manager may share information, give the team autonomy and protect it from pointless decisions. When priorities, ownership or economic conditions change, that arrangement may disappear with the manager. Not necessarily through betrayal: informal rights simply have a shorter life than formal ones.

The third principle:

Relationships sustain culture. Organisational arrangements sustain a voice in decisions. Neither replaces the other.

In IT, “cooperative” can sound as though a sack of potatoes and an accountant in sleeve protectors are about to appear. That is unfair to the potatoes, the accountant and the idea itself.

The International Cooperative Alliance defines a cooperative as a voluntary association meeting shared economic and social needs through a jointly owned, democratically controlled enterprise. The key words are jointly owned and controlled, rather than “everyone is equal in everything.”

There are functioning technology examples. Spain's Igalia contributes to browser engines, graphics infrastructure and other systems used around the world; it describes itself as worker-owned and worker-managed. Britain's CoTech network brings technology cooperatives together so they can undertake work beyond an individual team's scale. Mondragon shows that a cooperative group can grow far beyond a small workshop. It also reminds us that size does not abolish crises, conflicts or failures.

These examples guarantee nothing. A cooperative can be slow, poor, closed, unfair and badly managed. One person, one vote does not turn a bad product hypothesis into a good one. Shared ownership does not fix weak sales. Democracy does not remove the need for competence.

A co-owner also carries some of what an employer shields an employee from: risk, uncertain income, the need to contribute capital and make unpleasant decisions. This is not a higher stage of moral development. It is a different exchange.

The fourth principle:

Joint ownership makes sense only with responsibility, transparent rules and a product an outside customer is willing to pay for.

I use digital artel to mean a group of people who gather around useful work, agree on contributions and own the result under agreed terms. A federation connects several such organisations through shared resources and working agreements.

One cooperative could operate a computing service for members and outside businesses. Another could develop products using that capacity. A third could organise procurement and maintenance. A shared market for work helps find customers and assemble teams. We combine the roles of producers, consumers and co-owners.

External sales pay for labour, tools, support and obligations. Some of the result can support new products and common reserves. Internal payments show who uses a resource but do not create additional external revenue. The enterprise must earn its income through useful work.

This proposal draws on SoDelo's existing analytical foundation: the concept, economic scenarios, federation design and user journeys. The platform is under development. The calculations and agreements still need practical testing; the organisational and financial arrangements remain working hypotheses until then.

We can begin with a small team, a clear customer or a computing pilot. We limit the time and budget, assign responsibility and test the result. If the chosen step does not work, we change it or stop it. This preserves resources for the next attempt.

The fifth principle:

Build with the larger purpose in view, and test every step against real tasks, costs and relationships between people.

In large-scale development, I have seen two equally expensive extremes.

In the first, one manager decides everything. The team moves quickly while that person's view matches reality. When it no longer does, the system moves just as quickly in the wrong direction.

In the second, nobody has the right to make the final call. Every decision is discussed to a state of moral purity. Speed falls, responsibility dissolves, and the most patient participant acquires an informal veto.

Joint ownership does not require votes on libraries, button colours or every customer discount. It requires agreement on decisions that cannot be delegated without oversight: amending the charter within the law, distributing the financial result under the applicable rules, admitting and excluding members, setting settlement procedures on withdrawal and selling a key asset.

The rest should be delegated to roles. A role has a decision domain, a metric, a term and a recall procedure. Dissent is recorded. New evidence can justify reconsideration; disliking the outcome cannot justify replaying the decision forever.

AI is useful here for preparation, memory and checking: assembling context, exposing contradictions, modelling scenarios and recalling past decisions. But it should not receive political authority. The voice belongs to people because the consequences belong to them.

The sixth principle:

Democracy defines the boundaries of power. Professional roles provide speed within them.

You may read all this and decide a cooperative is not for you. That is a perfectly reasonable result. Some people have a strong team and work they do not want to change. Some are building a startup and deliberately choose venture capital. Some value predictable income more. Some do not want to discuss a charter after a day's work. None of those choices makes a person less free.

Agency begins with the habit of seeing the whole agreement, rather than with a registration form.

What do I create? How is my contribution measured? What am I learning? Which dependency am I increasing? Can I leave? What remains after the project beyond a line on my CV? Where am I willing to exchange control for stability, and where am I no longer willing?

Employment can be a deliberate choice. A product of your own can become an expensive form of self-exploitation. A cooperative can give you a voice, or a calendar full of meetings. The model's name does not spare you the work of thinking.

I know how to assemble teams and take complex changes into production. I now want to apply that experience to a shared enterprise in which members accumulate useful assets and have real rights to the results.

I care about the stability of people and their families. Shared tools, products we own, several customers and mutual support could give people a stronger footing. This requires earned income and working agreements. I am ready to calculate, test and build them with other participants.

We need leaders for specific areas: computing, development, sales, procurement and organising work. We need specialists willing to see a chosen task through. At the beginning there is more opportunity to shape the enterprise and more uncertainty. Additional rights are discussed in return for specific contributions, without promising income simply for joining early.

I continue to work in a large company and see businesses as potential customers and partners. Cooperative projects are kept separate from an employer's data, resources and obligations. Participation begins with a manageable step and a voluntary agreement.

If you would like to join the IT cooperative, help build SoDelo or have a strong idea for a shared product, email me or write to me on Telegram. A few lines about yourself, your idea or what you would like to work on are enough. I will get in touch, explain what we have built so far, and we can discuss where to start together.

There is no group chat yet — that will come a little later. Cooperative membership is not open either; for now, we are getting to know each other and agreeing on the first tasks. Please feel free to write. If I have not replied within a week, send me a reminder: your message may have got lost, and a follow-up will help.

I want my sons to have more choices and more people with whom they can build something of their own. That is why I am ready to begin this work with you now.

Vladimir Lovtsov
Yours, DPUPP

---