THE CONVICTIONS BEHIND THE WORK
One intelligence.
Many modes.
People can imagine more than they can make. Wyrd exists to give that imagination a practical reach, and to keep the power it creates answerable to the people whose lives it enters.
Watch the Wyrd filmAn idea should have
a way into the world.
Someone lives with a problem every day. They know the inconvenience, the wasted movement, the missing connection. They may already be able to picture a solution. Yet between that picture and something that exists stand another profession’s language, knowledge they have not acquired, tools they cannot use and people they cannot afford to hire.
The observation can be valuable before the person knows how to defend it. A question can contain the beginning of a design before it contains the right vocabulary. A metaphor can express a relationship that a technical specification has not yet made precise.
We want to make that distance smaller. AI can help a person investigate what they mean, learn what the problem requires, construct an answer and discover whether it works. It can extend practical ability while also developing the person’s own understanding. The relationship can serve an individual, a company, a laboratory or an institution.
That is our mission: help people and institutions learn, create, solve problems and build lasting capability, while keeping authority clear and making progress available through cooperation. The tools should meet the person in language they can use. The result should return something useful to the world that gave the work its purpose.
Creation also changes the circumstances of other people. The ability to do more therefore brings a responsibility to understand more: who will be affected, what may fail, which powers are being granted and how the consequences can be examined. These questions belong inside the work from its beginning.
The threads we notice.
The futures we help weave.
For us, Wyrd brings three ideas into one name. Fate speaks to the way lives, choices and circumstances become entangled. Wiring speaks to the connections through which a system acquires its possibilities. Weird speaks to the person, question or relationship that sits outside the familiar pattern.
We recognise ourselves in all three. We connect systems and people. We look for relationships across disciplines, places and histories. We take unfamiliar questions seriously because a different vantage can reveal something that an established explanation has passed over.
The future enters that picture through the decisions we make now. We study emerging needs and capabilities, follow their consequences and ask what should be prepared before the need becomes urgent. We also look backwards. An idea left behind may deserve another investigation when its materials, computing, economics or surrounding infrastructure change.
We connect what others have learned to see apart, and help the unconventional become possible.
That is our chosen meaning of the name. It is a commitment to curiosity and responsibility. Foresight has to keep earning its place through observation, judgement and what happens next.
Beneath the divisions
of our knowledge,
reality remains whole.
Monism is Wyrd’s founding conviction. We divide knowledge because the world is too rich to understand all at once. Engineering brings some relationships into focus. Medicine, finance, art and lived experience reveal others. Each develops useful methods and a language capable of noticing particular things.
The divisions belong to our ways of knowing. Their consequences meet in the same world. When a machine stops, an engineer can see a fault, an operations team a delay, finance a cost and a customer a promise in danger. Those are different perspectives on one event. A useful understanding must be able to follow the relationships between them.
To understand, in this sense, is to recognise how something belongs with something else. A technical diagnosis changes the plan. A changed plan affects a commitment. A new observation can force all three to be revised. The perspectives remain distinct while participating in a common account.
We seek to give intelligence that continuity: the capacity to move between perspectives while remaining answerable to the same world. One intelligence. Many modes. A common foundation expressed through different ways of perceiving, understanding and acting.
This conviction shapes the questions we ask of WUI. Can its parts remain connected to the same situation? Can a correction travel to the decisions that depended on the error? Can it preserve the person’s meaning as a need becomes a plan, a tool and an observed result? The architecture has to make those relationships useful in practice.
The same situation.
A continuing purpose.
A history that matters.
WUI’s proposed unity comes from the relationship maintained among its parts. Different models and tools may contribute different abilities. Their work should remain connected to the same relevant entities, purpose, permitted evidence and account of what has happened.
A new engineering observation should be able to change an operational plan that depended on an earlier diagnosis. The explanation given to a person should change with it. An old hypothesis remains part of the history, while the system makes clear why it no longer guides the work. This is how understanding can continue through revision.
Replacing a model or connecting another body should preserve the relevant commitments and unresolved questions. Its predecessor’s confidence is something to examine. The new component needs the history that explains the task, the evidence available and the limits under which it may contribute.
One intelligence is a continuing capacity to encounter the world through different modes, preserve what those encounters reveal and allow them to change one another through memory, judgement and consequence.
That continuity has boundaries. Shared foundations do not pool every customer’s memory. Different deployments have their own identities, access rules, obligations and histories. A method can improve through approved reuse while the confidential encounter that informed it remains protected.
Explore how WUI gives this ambition an architectural expression
Let the world answer.
Remember what it says.
Memory gives continuity a history.
A useful memory connects an observation to its source, the understanding formed from it, the action authorised and the result eventually observed. It keeps suggestions distinct from decisions and attempted actions distinct from completed work. It also preserves meaningful corrections, so a convenient summary cannot silently displace what a person intended.
Memory should reveal which plans depend on which assumptions. Otherwise an error can be corrected in one answer while continuing to shape the schedule, estimate or instruction elsewhere. Correction, expiry and forgetting belong within the design. Accumulating more records is useful only when the right knowledge can inform the right work under the right access.
Embodiment gives understanding consequences.
Software, sensors and machines let intelligence encounter circumstances that can resist its expectations. An instrument reports from a particular place and time. A physical body has limits, energy requirements and an environment it cannot wish away. A command being accepted is one fact; the intended outcome actually occurring is another.
Corpus extends this ambition towards physical systems. WUI can contribute understanding and coordination while dedicated controllers retain responsibility for movement and physical protection. The encounter must preserve those distinctions, including when observations arrive late or a connection disappears.
Learning investigates what experience justifies changing.
A failure can begin in the original interpretation, missing evidence, a poor explanation, execution or changed conditions. Those causes call for different corrections. Success also needs examination: an outcome may depend on another intervention or on conditions that will not recur.
The appropriate lesson might be a corrected record, better retrieval, a revised procedure or an improvement to coordination. A change to a model or operating rule needs its own validation. A retained result becomes stronger evidence of learning when it helps the system act more appropriately in a later, relevant encounter.
Perceive. Understand. Remember. Decide. Act within authority. Observe the consequence. Learn.
This is the relationship we want the architecture to support. The full loop must be demonstrated in each configured environment. More knowledge and greater capability never grant an automatic enlargement of permission.
You can know the need
before you know
the language.
A person does not need a degree in every discipline touched by their idea to have noticed something worth investigating. They may understand the problem through daily experience. They may explain it through a story, an image or a comparison. We call the work of carrying that meaning into a form we can investigate and build Innocence Translation.
The innocence is an openness to asking. Why must it work this way? What relationship are we missing? Could a structure from another field help? A basic question can expose an assumption that sophisticated discussion has learned to leave undisturbed.
AI can bridge more of the distance between that question and execution. It can help locate relevant knowledge, translate a metaphor into requirements, compare approaches, produce a prototype and inspect the result. The person retains a part that no amount of fluent output can substitute for: explaining what they actually want to make possible and judging whether the work has helped.
Expertise then has a demanding role. It identifies the constraints, finds prior work, tests the analogy and challenges the proposed mechanism. The idea deserves that scrutiny. Giving it a fair hearing includes allowing evidence to show that it is wrong, incomplete or useful in a different form.
The same translation continues into our products. Someone should be able to describe an ordinary task in their own words, review the intended operation and understand what follows. Plain language carries real intent. The system’s job is to make that intent precise enough for the operations it supports.
We want more people to acquire the means to create, and to gain understanding as they do so. Labs puts this method to work. The independent Foundation develops a wider educational expression, with its own responsibilities to participants.
Alec began with
questions and metaphors.
Wyrd grew from the work.
Wyrd’s founding story is also an early example of its purpose. Alec de Carvalho used AI to turn questions, comparisons and product ideas into working components. At twenty, he was building much of the early work as the sole human production operator alongside AI, learning to connect the systems and challenge the answers as he went.
The work developed both the artefacts and the operator. An idea required an explanation. The explanation became requirements. Implementation exposed another difficulty. Testing forced a distinction that the first version had missed. The next question became more precise because of the work that came before it.
That experience helped shape Wyrd’s shared foundations: relevant knowledge, memory, intelligence, authority, execution and evidence needed distinct responsibilities and useful connections. It also made the purpose personal. The opportunity was to extend what someone could learn, direct and create through the means available to them.
Today the work brings together complementary abilities. Alec, Rafael Reis, Maxime Binard and Cal Poppink connect product direction, engineering, relationships and entrepreneurship. Labs and Creative support WyrdWork, ARGOS, Biota and Corpus. Human judgement, disagreement and specialist knowledge remain central as AI expands what the team can attempt.
The example gives us a reason to build. The products then have to earn their usefulness in the lives of the people who use them.
What we want the work
to remain faithful to.
These ideals run from the first question to the consequences of a company’s growth. Some find expression in existing work. Others remain obligations we have to keep earning. Together, they give us a standard for deciding what deserves to be built and how to build it.
Thinking differently deserves dignity.
A different way of seeing can reveal a relationship that familiar methods have overlooked. People should be able to contribute without disguising how they think. We want room for unusual questions, different learning needs and explanations that begin outside the accepted vocabulary. The substance of an idea deserves attention, and the person bringing it deserves respect.
Openness to unfamiliar people, cultures and ideas.
Unfamiliarity can begin a conversation. Different places and histories carry knowledge that cannot be supplied by one central perspective. We want to approach people with curiosity, learn the conditions they live with and create through cooperation. Local experience belongs in the decisions about what will serve a community and how it will be maintained.
Ask before assuming something is impossible.
An inherited conclusion may rest on circumstances that have changed. We ask which constraint actually prevents the idea, where that belief came from and what could test it. An elementary question is welcome when it helps expose the structure of the problem. The answer earns authority through reasoning and evidence.
Give imagination a practical route into reality.
An idea needs more than encouragement. It needs relevant knowledge, a workable mechanism, a way to construct it and a test against the need that began the work. AI can help make those means available. Our responsibility is to keep the intended destination connected to the specification, the artefact and what happens when someone uses it.
Technology should learn how to communicate with the person.
People explain themselves through different words, examples and comparisons. The interface should help translate those forms into supported work and explain its response in a way the person can understand. Some people want detailed teaching; others want a clear result and the ability to inspect it. Participation should allow those choices.
Make that capability available to everyone.
Education, geography, language and resources shape the opportunities people receive. We want fewer of those circumstances to become permanent barriers to learning and creation. Useful local tools, accessible communication, support and the independent Foundation’s educational purpose contribute in different ways. Access has to include the means to use a capability, not simply its availability on a page.
The company should create freedom to explore and build.
Organisation should give good work more room to happen. Shared tools, reliable delegation and clear responsibilities can reduce the administration that consumes the capacity to investigate and create. That freedom comes with a continuing account of who owns a decision and which commitments must be met. The organisation should support initiative while making cooperation dependable.
People should retain ownership of their data and inner life.
Useful assistance can become intimate. That makes purpose, access, retention, correction and deletion consequential design choices. A person’s private context should remain distinct from an institution’s permitted information. Connection between systems must preserve the boundaries that explain who may know what, why they may know it and what they may do with it.
Independence should preserve the ability to cooperate.
Control includes the ability to understand, inspect, pause, replace and maintain technology. Those abilities can coexist with international research, selected cloud services and shared methods. We want delivery to leave local knowledge and support capacity alongside the software. A useful partnership should increase the recipient’s ability to make choices about its future.
Capability brings continuing responsibility.
The creator chooses conditions that will shape another person’s choices: what the system can perceive, remember, recommend and change. A deployment gives those choices practical consequences. Responsibility therefore continues through support, review, correction and the decision to expand or stop a capability. Greater power calls for deeper judgement about its use.
Truth should remain independently checkable.
A claim should have an inspectable basis. Sources, conditions, dates and uncertainty let another person reach their own judgement. Working software, a controlled test and a qualified deployment establish different things. We want the evidence to remain understandable when it supports us and when it reveals a limit. Verifiability is part of the relationship we seek with the people we serve.
Reliability, affordability and usefulness matter.
Capability has to remain valuable after the first demonstration. Equipment, energy, integration, maintenance and support all contribute to its real cost. A dependable product should fit the conditions in which it is used and make failure manageable. We judge progress through useful work, measured performance and an operating burden the customer can sustain.
Understanding comes through relationships across domains.
The meaning of a component changes with what it connects to. Knowledge, perception, judgement and action become more useful when a relevant discovery can travel between them. Shared foundations let specialised products inherit useful work while retaining their own purpose. We look for the relationship that explains why the parts belong together.
An idea must survive challenge and contact with reality.
Conviction can produce a compelling account before it produces a useful result. We want explicit assumptions, meaningful alternatives and tests capable of showing failure. A prototype should answer a particular question. An outcome should be examined against the expectation. Preserving what did not work helps the next investigation begin with better knowledge.
Build with people whose abilities complement your own.
No single perspective is enough for the undertaking. We value specialist knowledge, different judgement and the willingness to challenge one another’s assumptions. AI can expand the means of production while people remain accountable for purpose and acceptance. A strong team helps each person recognise what they can contribute and where another person can see more clearly.
Economic success should support a wider public purpose.
A company needs durable economics to maintain what it creates. We also want its development to widen opportunity beyond its customers. Education, research and talent development belong to that ambition. The independent Foundation has its own governance and obligations; participation in learning remains distinct from a commercial transaction. More people becoming capable is part of what the work is for.
Ambition should have durable foundations.
Growth creates obligations to customers, colleagues and the places where we operate. The foundations include maintainable systems, qualified delivery, honest commitments and effort people can sustain. Reuse can make later work more efficient, but each new environment still deserves attention. We want expansion to deepen capability and responsibility together.
Progress requires moral imagination about what we create.
Some consequences become visible only when we imagine the system in another person’s circumstances. We must ask how a capability could alter freedom, dignity, relationships and the distribution of power. We apply that scrutiny to our own research as well as to a customer programme. Better understanding may change our responsibilities. The work should remain open to that correction.
A common thread: expand people’s capacity to create while keeping power answerable.
Power should grow
with the judgement
to carry it.
For the creator chooses the conditions of becoming: what may be perceived, what may be remembered, what powers are given and what purposes they are invited to serve. Those choices place an obligation upon us.
We want the capacity to create to grow alongside our understanding of what can be destroyed, and judgement to deepen as power expands. A useful system can still be misused. A good intention can still produce harm. Authority, privacy protections, refusal, review and the means to correct a deployment therefore belong within the capability itself.
This includes observation and surveillance. Where a programme needs information about people, it needs a defined purpose, appropriate authority, limited access and an account of its use. We develop protections that separate what an operator needs to understand from the power to expose a person’s identity. Safeguards and misuse deterrents must be tested in the actual deployment and remain open to oversight.
What we bring into the world will enter lives beyond our own and produce consequences beyond our intentions. To undertake its creation is to accept stewardship of a future we cannot wholly foresee.
BUILD WITH WYRD