How we want to work

We want ASUME to operate on trust rather than continuous supervision. People should have enough context to understand the problem they own and enough autonomy to make reasonable decisions about it.

Responsibility means moving the work forward, making progress visible, raising uncertainty early, and asking for help when the problem requires it. We care more about what improves than about where somebody works, when they appear online, or how much activity they produce. Autonomy is not the absence of accountability. It works because ownership is clear.

ASUME is small today, and we do not want to introduce organisational structure before the work requires it.

As the company grows, we want responsibility to remain close to the problem. A person or small group should have enough context and authority to investigate, build, test, speak with users, inspect evidence, and improve the work without unnecessary handoffs or approval layers. The exact size and shape of a team should follow the problem rather than a fixed organisational model. Some problems may need one clear owner. Others may need product, engineering, inference, design, or domain knowledge working closely together.

We will add structure when it improves ownership, coordination, quality, or safety. We should not add it simply because companies of a certain size are expected to have it. The goal is not to preserve a particular team size. It is to preserve short learning loops, clear ownership, and direct connection to the problem.

People working on a problem should remain close to the evidence that can prove them wrong. For product work, that means users and actual workflows. For inference work, it means evidence, evaluations, failures, and human review. For engineering, it means how the system behaves in production. For commercial decisions, it means what customers actually do rather than what we hope they will do.

We do not want layers of communication to gradually separate the people making decisions from the reality those decisions affect. Direct contact shortens the learning loop.

As ASUME grows, this may become harder to preserve. That makes it more important, not less.

Much of the work at ASUME requires sustained concentration. Research, engineering, inference design, evaluation, writing, and understanding customers all become worse when attention is constantly fragmented.

We therefore prefer asynchronous work where it makes sense, clear written context, and meetings that exist for a reason. Flexibility is part of this. We care about the quality and responsibility of the work rather than enforcing routines simply because companies traditionally have them.

As ASUME grows, we want to preserve people's ability to think deeply rather than gradually replacing concentration with coordination.

Important context should not exist only in someone's head. Decisions, assumptions, trade-offs, specifications, problems, and important changes should be written down when doing so helps another person understand or challenge them.

Writing is not bureaucracy. Used well, it reduces bureaucracy because fewer decisions have to be reconstructed through meetings and repeated conversations.

We do not want documentation for its own sake. We want enough written context that the company can learn from itself.

We want disagreement to be direct, specific, and respectful. If the reasoning is weak, say why. If an assumption looks wrong, challenge it. If a product decision does not make sense, make the disagreement visible while it can still change the outcome.

The same applies when something works well. Understanding why good work succeeded is useful evidence too. We do not want people optimising for agreement with founders, managers, or the loudest person in the room.

A company trying to understand reality has to make it safe to point out when reality disagrees with us.

We do not want to confuse the appearance of a company with the substance of building one. More meetings do not automatically mean better coordination. More dashboards do not automatically mean greater understanding. More processes do not automatically mean greater maturity. More activity does not automatically mean greater progress.

A process should exist because it solves a problem. A metric should exist because it helps us understand something important. A policy should exist because it makes behaviour or responsibility clearer. When those conditions disappear, we should be willing to remove the process too.

As ASUME grows, credentials may tell us something about a person, but they should never become a substitute for understanding how that person actually works.

We care about reasoning, curiosity, ownership, craft, learning rate, judgment, and the ability to work constructively with other people. Degrees, titles, prestigious employers, and years of experience can provide useful evidence. They are not conclusions.

The same principle we apply to companies should apply to people: look beyond the surface information.

We do not know exactly what ASUME will look like at twenty, one hundred, or one thousand people. Trying to design all of those organisations today would mostly produce assumptions disguised as policy.

Instead, we want to preserve the principles that help us learn: make reasoning visible, give responsibility to capable people, protect focus, challenge weak assumptions, avoid unnecessary structure, and change our ways of working when evidence tells us they are no longer effective.

We want ASUME itself to remain revisable.