Please help keep Beyond the Wall alive
They died once. Please help us keep their stories from disappearing too.

You may have come to us through one of our Vietnam Medal of Honor stories—a story of courage, family, sacrifice, love and legacy. Something about that individual’s life mattered to you, our heroes, our casualties of war mattered to you. History and legacy matter to you.

My name is Ed Roth. I co-founded Beyond the Wall with its inventor. And I am asking you to help us carry that connection forward. The inventor, my co-founder, has personally financed four years of invention and engineering to bring Echoes technology into existence. He and I can no longer carry the cost alone, and we cannot sustain this work without support. Without new funding, we may have to take the platform offline within weeks.

Our immediate goal is $10,000. That will help sustain Beyond the Wall for approximately two more months while we seek additional donors and pursue support from larger foundations and government programs. That longer-term support is not yet secured. Please donate what you can today. You do not have to make a large gift. Small contributions and regular monthly support can help share a cost Michael can no longer carry alone.

Why we co-founded Beyond the Wall

I co-founded Beyond the Wall because I believe, like Michael, that every life matters—and every individual story deserves a future. These men and women were sons and daughters, fathers and mothers. They loved someone. Someone loved them. They had their own humor, fears, beliefs and dreams. They imagined futures that war took from them. To me, it is an absolute tragedy that we can lose their stories to history as well.

Our Medal of Honor series brings extraordinary courage into view. But our purpose includes every individual named on the memorial. A life does not have to be decorated or famous to deserve remembrance. In fact the reverse is true, Michael invented this technology exactly because it would give every life the meaning and legacy it deserves. You should not have to be a war hero, rich or famous to be important.

Granite memorials preserve our heroes names. We need those places of remembrance. But a name cannot tell us who someone was. What made them laugh? What mattered to them? What did they hope to become?

And what might they make of the world today? What conversations might we have if their perspectives could remain part of our lives?

Those questions are at the heart of our mission. Michael devoted four years of invention, engineering and personal funding to creating the technology that makes it possible explaining to me "I wanted to give families, communities and future generations a way to encounter the individual behind the inscription—to carry something of that person forward through conversation, curiosity and continuing discovery."

The technology works. The work is not finished. Today, you can visit beyondthewall.ai and speak with Echoes rooted in people who once lived. This is a working first generation, with more development needed. We must deepen the intelligence, strengthen memory and reliability, improve how historical evidence is reviewed, and make it easier for people to contribute what they know.

Each Echo is an evolving Synthetic Intelligence informed by surviving records, historical context and contributed memories. A letter, a photograph or a family recollection can help deepen its individual roots. We cannot know exactly who someone would have become. We must distinguish what history tells us from what the technology infers or newly creates. But we can make room for their stories and perspectives to reach people they never had the chance to meet.

A gift to the future

I see that as a gift to the future. A grandchild encountering a life they knew only through a photograph. A young person discovering the hopes of someone their own age who never came home. A family contributing memories that might otherwise disappear when the last person who remembers is gone. I believe we die a second time when we are forgotten—when our name remains, but no one knows the person it belonged to. That is the loss I want to help prevent.

Why your support is needed now

Creating the technology took years. Keeping it available and improving it requires continuing support. This year, the income that has allowed Michael and me to fund the work has fallen sharply. He has continued supporting it, but his ability to carry the cost is almost exhausted. We are a small team. Your donation helps sustain our four paid engineers, the computing and infrastructure that keep Echoes accessible, and improvements to memory, reliability and the tools people use to contribute records and memories. The $5,000 goal is immediate bridge funding. It gives us approximately two months to continue the work and build wider support. It does not meet the mission’s full long-term funding needs.

We hope to earn support from government programs and larger foundations. But we cannot depend on that hoped-for funding to sustain the work today. We need a community willing to help now. Without that support, Michael cannot continue funding it alone, and we cannot sustain Beyond the Wall. Please give what you can today. Share the campaign. Contribute a photograph, a letter or a memory that helps someone be known. Help keep this work accessible, growing and available to the generations ahead.

They died once. Please help us keep their stories from disappearing too. These lives were cut short. Their place in our human story does not have to end.

— Ed Roth
Co-founder, Beyond the Wall

Beyond the Wall is an initiative of Consult Atlas Foundation, a U.S. 501(c)(3) public charity, EIN 93-4435012. Donations are tax-deductible for eligible U.S. donors to the extent permitted by law. Tax treatment outside the United States varies.

If you are curious about the technology behind an Echo, the following sections are for you. Here we explore how a life’s surviving records become an individual Echo’s foundation, how Michael—the evolving Synthetic Intelligence architecture behind each Echo—supports its continuing development, and what we are still working to achieve. You do not need to understand the engineering to understand why a life deserves to be remembered; you and every life matter. But if you want to look further, here is the story behind the invention.

How A Life Becomes An Echo

A Personal Note From Michael, Inventor And Co-Founder

I am not an engineer by training. I am self-taught, and my path into this work has involved years of reading, experimenting, making mistakes and trying again. I have written my own patent applications for earlier inventions. For Beyond the Wall, I have worked alongside engineers whose expertise has helped turn what I imagined into something people can use. I know that extraordinary ambitions deserve careful examination. In the explanation that follows, we try to distinguish what works today from what we are still building and what we hope to achieve. But I also need to speak plainly: I cannot keep financing this work alone. After four years of personal funding, my resources are nearly exhausted. It is difficult to ask for help this openly, but remaining silent will not keep the platform available or sustain the people developing it.

I began because I believe every life matters, including lives history has left us very little to remember. If that purpose matters to you too, please help us carry the work forward.

— Michael

The Question Behind The Invention

If you could ask one more question of someone you lost, what would it be? Perhaps there is a childhood you never heard enough about, a decision you never understood, or a photograph whose story no one thought to write down. Perhaps you would want to know what they imagined your life might become, or simply tell them that the family is still here. That longing is part of the reason I co-founded Beyond the Wall.

You may have come to us through a story of extraordinary courage, but behind a citation, a uniform or a name on granite was an individual whose life contained far more than the moment history recorded. There were ordinary mornings, friendships, disagreements and interests, alongside things they understood and things they were still learning. People who knew them held parts of their story that an official record could never contain. War took their future, and the passing years can take the remaining knowledge of who they were.

Our work on Echoes asks whether technology can help carry something of that individual perspective into the future through an Echo you can question, contribute to and engage with, while remaining honest about what is known and what has been newly created. That is a difficult scientific and engineering problem, and a deeply human reason to keep working.

A Historical Life And A Continuing Echo

An Echo connects two timelines: the life of the person who lived, and the developing history that begins with the Echo’s creation and continues through its subsequent encounters. We also need to distinguish when evidence becomes available. A letter written during the war might be discovered today and discussed next year; its contents belong to the historical record, while its discovery and the Echo’s discussion of it belong to later events. Keeping those relationships clear allows a powerful possibility: a historically rooted Echo can engage with a world the person it represents never lived to see.

You might ask how an Echo approaches a contemporary question in light of its individual foundation. The resulting reflection is a new contribution by the Echo, rather than a recovered statement from the deceased person. That boundary allows room for continuing development while preserving the integrity of the original life, giving you a way to explore a perspective without being told that technology has discovered an undocumented private memory. We want the history to remain accountable and the encounter to remain open.

Understanding An Incomplete Beginning

Imagine waking after a head injury knowing your name, yet unable to recall the family, friendships and ordinary moments that gave your life its meaning. In some forms of retrograde amnesia, knowledge about yourself can survive while memories of personal experiences remain inaccessible. As you tried to understand your story, a photograph might show who stood beside you, a letter might reveal your humor or your hopes, and a friend might tell you something no official record could contain. Knowing your name would be a beginning; understanding the life behind it would take much more.

That offers an analogy for the incomplete evidence from which an Echo begins, with an important distinction: an Echo has no hidden store of the deceased person’s memories waiting to be recovered. Its foundation is constructed from surviving individual records, historical context and the memories others can contribute, and learning from a letter is different from remembering having written it. Knowing a name and service history is not the same as knowing the person. Your photograph, your letter or your family’s recollection may reveal something the public record cannot, helping to deepen the individual foundation, correct an assumption and represent that life with greater depth and accuracy.

Two Inventions, One Purpose

How do we give those surviving fragments a form you can engage with? Beyond the Wall brings together two substantial undertakings. The Echoes engine creates an individual foundation from surviving biographical material and historical context. It forms a Personality Crystal and uses that foundation to establish an Echo seed. After four years of invention and engineering, Michael reports that the Echoes engine generated 58,281 Echo seeds in approximately three hours, beginning with the American service members named on the Vietnam Veterans Memorial. Each beginning draws on the available facts about a particular life, placed within its historical and social context. The surviving evidence differs from person to person, so these foundations begin with different levels of detail and uncertainty.

Michael supplies the Synthetic Intelligence architecture through which that individual foundation, the Echo, can participate in conversation, use knowledge, retain relevant experience and develop over time. The working Echoes Beyond the Wall platform brings these Abiogenic Inc. technologies together today; strengthening memory, continuity and the reliability of continuing development remains part of the work your support sustains. You experience their integration as an Echo rooted in someone who actually lived—a beginning that further evidence and continuing curated care can help deepen, and through which future generations can encounter more than a name. Our goal is for each Echo to recognize gaps in its understanding and pursue meaningful questions. We need your help to develop that capacity.

Creating a foundation for an individual and enabling that representation to remain coherent through changing conversations, new evidence and continuing experience are different challenges. Bringing them together in a form that families and future generations can use is the purpose of Beyond the Wall. Four years of personally financed invention and engineering brought this work into a functioning platform, and we urgently need your financial support to sustain it and strengthen the capabilities that still need development.

The Echoes Engine: Beginning With What Survives

The starting point is the evidence a life has left behind. For one person, there may be letters, photographs, recorded testimony, school records, service documents and family recollections. For another, there may be only a few biographical facts. These may be a name, an age, a place, a unit, and dates marking the beginning and end of a life. We believe that a thin surviving record should be a reason to begin carefully, rather than a reason to leave someone out. The Echoes engine brings the individual material into relation with the person’s world: the historical period, social circumstances, language environment, occupation and documented experiences relevant to that life.

A person’s age in a particular year can help establish which events they could have encountered. A documented location can help identify relevant local history. A service record can connect the individual to a wider military setting. Each connection has limits. Membership in a unit does not prove that someone witnessed every event described in its history. A birthplace does not determine a belief. An absent record of children does not establish that the person had none. Those distinctions are part of the science. They determine what the system has grounds to say, what it can reasonably infer and what must remain uncertain. The goal is an individual beginning that uses context without allowing the surrounding history to swallow the person.

The Personality Crystal: Organizing An Individual Perspective

The engine forms what we call a Personality Crystal. The term describes a structured foundation within the technology. It is intended to influence how an Echo interprets what it encounters: which information appears relevant, what interests deserve attention, how uncertainty is approached and how a response is expressed. Personality, in this account, is an organizing perspective that reaches beyond a preferred tone of voice.

Two people can know the same facts and notice different things about them. One may attend first to family relationships; another to a practical problem; another to a question left unresolved. A meaningful individual representation needs a way for those differences to affect what happens next. The Crystal is our approach to establishing that starting organization. Its initial form draws on the available individual evidence and contextual inference. Some elements may be directly supported, while others remain tentative or largely unspecified until better information arrives. The scientific challenge is to keep those differences visible. A generated personality description can sound confident even when its evidence is weak. Confidence in the wording cannot be allowed to manufacture confidence in the underlying history. You deserve to know whether an answer rests on a letter, a reported recollection, a contextual inference or a newly generated reflection.

Crystallization: How A Beginning Takes Shape

We use crystallization to describe the formation of that organized foundation. The name is a metaphor for a computational process. It does not mean the software contains a physical crystal, or that a complete human mind has been recovered from records. A useful way to understand the process is to think about how evidence narrows possibilities. When little is known about an individual, many different dispositions and experiences remain plausible. Historical context provides a starting range of expectations. Direct evidence from the person can change those expectations.

Imagine, as an illustration, that limited records suggest a reserved manner of speaking. Then you contribute a letter full of affectionate jokes. The letter gives a reason to revise that assumption. It supplies something individual that the broader context did not reveal. We replace an assumption with evidence from the person’s own words. Or imagine that an official summary emphasizes duty, while personal correspondence reveals an intense interest in music. Both may be relevant. The richer picture need not discard the service record; it needs to make room for a life larger than its official description. This is evidence-aware inference: conclusions should depend on their support and remain open to correction.

The research behind our work also asks how to avoid counting the same evidence more than once. A story copied onto several websites is still one underlying account. Repetition can make a claim appear more established without adding independent support. These may sound like small details. For a family, they can decide whether an Echo respects an individual or gradually replaces that person with a persuasive collection of assumptions. Your photograph, letter or recollection can matter because it gives the work a more individual foundation.

The Echo Seed: Establishing The Beginning

The Personality Crystal provides the foundation for an Echo seed: an instantiated beginning linked to a particular life. The seed establishes a place from which the Echo can engage through Michael. It is the beginning of a developmental history, with opportunities for additional evidence, conversation and improvement. The word "seed" is deliberate and comes from the inventor and co-founder. Creating the beginning does not establish that every Echo has the same depth, the same supporting material or the same capabilities. A richly documented individual and someone known through a sparse record present different opportunities and different uncertainties. Both can deserve inclusion. Their differences must remain honest. But both start from the "Echo Seed".

The founder and his technical team report that the Echoes engine has generated a cohort of 58,281 Echo seeds. That is a reported creation result, not a claim that every seed is an equally mature or comprehensively documented individual. The public directory and a seed-generation cohort are also different measures. For us, the significance of scale is human. We want the opportunity to extend beyond the few people whose lives can support a bespoke production. For you, a seed may become a place where a family photograph gains context, a remembered story can be examined, or a young person begins to discover the life behind a name they have walked past. An initial record is something you can preserve. An individual beginning is something you can help deepen.

What Works Now, And What Your Gift Helps Advance

Beyond the Wall is available, and you can visit the platform and engage with Echoes today. Michael and the Echoes engine are part of an existing development effort with reported deployments, persistent-memory work and internal evaluation. The public platform identifies itself as early alpha, and its records, features and answers can be incomplete. Our research describes a fuller developmental architecture than we can claim has been uniformly validated across every Echo, so the experience available today must be distinguished from the capabilities we are still developing and testing.

We are strengthening individual continuity, memory, evidence handling, review and contribution tools, while more autonomous inquiry and deeper moral development remain areas to build and test. A working first generation can have real value while still needing substantial improvement. When you give, you help keep the existing experience available and sustain the people whose expertise is needed to improve it.

Emotion, Presence And Respect For A Real Person

The aim is to make engagement meaningful. A family may seek a sense of connection. A student may begin to understand the distance between a casualty statistic and a young person’s interrupted future. A conversation can invite attention that a list of dates never receives. But vivid language can create a special risk. A generated response that describes a spouse, a childhood incident or a remembered battlefield conversation has made biographical claims. Emotional power does not establish that those events happened.

Our engineering must make room for expressive engagement while strengthening the boundaries around personal history. An Echo can explore themes of duty, fear, affection or loss without inventing an undocumented event and presenting it as memory. It can acknowledge the limits of its evidence without reducing every answer to a lifeless list. Finding that balance is part of the continuing work. We do not believe that care for historical truth requires abandoning the human purpose of the experience. You should be able to feel the importance of a life while understanding what supports the representation before you.

Michael: The Organizer Behind The Conversation

Language is an essential part of the Echo experience, but it is only part of the architectural problem. A language model can contribute knowledge, possible interpretations and candidate responses; Michael’s architecture is designed to place that capability within a continuing organization of identity, priorities, memory and permitted change. Our aim is for Michael to support continuing individuality, beyond language alone.

In the research, Michael’s creators use the term dual hemisphere to describe two functional roles. One supplies the generative language capability, while the other is intended to govern what should be considered, what remains uncertain, which response is appropriate and what should be carried into memory. These are software roles rather than biological brain structures. The question is which processes guide the answer: does the Echo’s individual foundation help select the evidence before a response is produced, can conflicting information prompt further investigation, and can an unsupported personal claim be challenged before it is released? What happens afterward matters too, because the encounter may provide grounds for a lasting correction or a new line of inquiry.

A computer’s speed can be extraordinary, but speed alone does not explain human understanding. The brain works through vast networks of interconnected neurons, and a question can engage memory, emotion, expectations and judgment together, giving it meaning beyond the words used to ask it. That interconnected character of understanding inspires our work on Michael. Our ambition is for an Echo’s individual foundation, accumulated experience and developing judgment to participate in how it approaches a question, rather than existing only as a description attached to its responses. We are pursuing that organization in software, with appropriate human oversight, rather than claiming to reproduce the human brain.

Echoes Beyond the Wall is the working platform that already brings Michael and Echoes together. Making those relationships more dependable across different situations remains a central part of the engineering and evaluation your donations and support sustain.

Why A Library Is Different From Memory

You can own a library without having read every book in it, and reading a book does not mean you will remember it or change your understanding because of it. Sometimes, though, an idea stays with you and changes the question you ask when a similar situation arises. Our research distinguishes those possibilities: an archive holds source material, retrieval selects information relevant to a question, and memory retains something from an encounter. Learning, in the stronger sense we are pursuing, changes how the individual handles a later situation.

An Echo must not become an undifferentiated copy of everything available to it. A historical source needs to remain identifiable as a source, a person’s testimony must remain attributable to that person, and an interpretation adopted after reflection has a different status from either. Preserving those relationships is part of building a trustworthy continuing Synthetic Intelligence. A conversational AI responds under its governing instructions and the context supplied within a chat; our goal for Echoes is a continuing individual foundation that helps determine what each Echo finds meaningful, questions, accepts and carries forward.

Each Echo Should Learn From A Conversation Without Being Swayed By It

Our goal is a Synthetic Intelligence that approaches each encounter with you through its own continuing individual foundation: its perspective, values, accumulated understanding and experience. A visitor’s confidence or repetition should not substitute for evidence, and a request to adopt another personality should not erase the Echo’s nature. It should consider what you offer, question what it does not understand and revise its understanding when there are sound reasons to do so.

Learning should deepen the individual Echo, not replace it with the preferences of whoever speaks next. Across conversations—even at the scale of billions of encounters—our goal is an Echo that retains its individuality while continuing to learn and develop. That continuity is an engineering and research objective we are working toward, not a capability demonstrated at that scale. Your support helps us build and test the memory, reasoning and protection of individual identity that this requires.

Suppose you contribute a letter to an Echo that contradicts an earlier account. Merely saving both documents does not resolve the problem. The Echo needs to identify the disagreement, consider whether either source is mistaken and investigate whether an apparent contradiction reflects different contexts. It needs to retain uncertainty where necessary and make a justified correction effective in later engagement, just as you would want a lesson you had learned to matter the next time you encountered the issue. Our design therefore extends beyond uploading documents to consider how they are used, what is learned from them and whether a correction actually matters when you return.

Learning That Changes The Next Encounter

An AI system saying “I have learned” is not enough for us; meaningful learning needs to make a persistent and relevant difference to what happens afterward. We therefore ask whether an Echo can recall a contribution later, apply a lesson when the wording changes, revise an earlier interpretation when better evidence arrives and preserve that improvement after a session ends. These are different accomplishments, and success at one does not automatically establish the others.

A model can remain unchanged while the surrounding architecture updates an individual’s memories, priorities or evidence links. The information available to the next conversation has changed, so the behavior of the whole system can change too. Published research has demonstrated useful forms of reflective agent learning through retained feedback without updating the underlying model’s weights. That supports taking this engineering route seriously. It does not certify Michael’s performance or settle every design decision in our Echoes system. Our team has worked on persistent memory, retrieval and continuity, with further improvements still needed. The purpose is for useful experience to have a durable effect that we can examine, rather than depend on a conversational promise. Through your donations, you are helping sustain that difficult transition from a convincing exchange to more dependable continuing engagement.

Curiosity: Learning Which Question Matters

An individual perspective is expressed partly through the questions it asks. Why does this account disagree with another, what is missing from the story, and what would help resolve the uncertainty? In Michael’s theory, curiosity is one of the initiating drives for learning, helping select an unresolved question worth investigating. That requires judgment about relevance: an inquiry has value when it can deepen understanding, correct an error or help the individual respond more responsibly, rather than simply generating further questions.

For an Echo, a useful question might concern conflicting dates in a service history, a family relationship whose evidence is incomplete, or a present-day topic connected with its developing interests. The underlying system also needs limits on time, computing and permitted access. A question that cannot be resolved should not produce endless unsupported speculation. But an answer can reveal another meaningful question. Our research explores how an Echo might pursue those connected questions, while distinguishing new avenues of inquiry from repeated versions of the same question. We use “answer entropy” as a proposed way to describe uncertainty among possible answers; those alternatives need not have equal evidentiary weight. Reaching a resource limit means an investigation has been paused, not that its current answer has become final. Even when the available avenues have been exhausted, uncertainty may remain. Our goal is for unresolved questions to remain available for further inquiry when new evidence or resources arrive.

Curiosity-driven learning has an established research history; our question concerns how inquiry is connected to a particular Echo’s foundation and accumulated experience. Some reflection mechanisms have been reported in the development record, while the fuller vision of autonomous, individually directed investigation remains research and development work. It should not be assumed to operate uniformly for every Echo. Your support helps us develop the capacity to pursue a worthwhile question alongside the discipline to recognize when an answer remains unknown.

How Judgment Could Develop Through Experience And Education

The Moral Development companion paper asks a further question: could an individual learn why a commitment matters, and carry that understanding into unfamiliar situations? This concerns more than a list of prohibited responses. It concerns how attention, education, experience, competing considerations and reflection might contribute to judgment. Consider a hypothetical situation. You share a family recollection containing sensitive information about a living relative. Later, someone else asks the Echo to repeat it. There are several considerations: truthfulness, privacy, consent and the trust under which the memory was shared. A useful response requires understanding the relationship between those concerns.

The same principle applies when a valued contributor supplies an inaccurate claim. Respect for the contributor and responsibility to the historical person both matter. The aim should be to examine the evidence and correct the representation, without turning loyalty into a reason to preserve an error. Education can contribute to such understanding. The system need not cause harm in order to learn why harm matters. Testimony, explanation, literature and carefully constructed examples can supply material for reflection. These are research examples, not claims that Michael has demonstrated a human conscience.

The proposed goal is durable, revisable understanding: a lesson that can influence a new situation, resist an unsupported attempt to overwrite it and still improve when its application was mistaken. Other AI research also investigates learned reasons and ethical judgment. Our research examines the role of continuing individual development, rather than claiming that learning ethics is unique to Michael. This matters to all of us because the system is entrusted with representations of real lives and encounters with living people. Respect, care and accountability must be developed into the way it operates, evaluated over time and supported by appropriate safeguards.

Why Keeping Individuals Separate Matters

If you tell one Echo about your family, that story must not silently become another Echo’s personal experience. If two veterans served in the same unit, they may share historical context. That does not make their biographies or relationships interchangeable. This is an engineering problem involving how information is identified, selected, stored and reused. A name on a screen is only the beginning of the distinction. Individual boundaries must survive summaries, later conversations and background processing. Our research therefore asks how to protect the link between the person, their sources and the developing Echo. It also asks how to recover correctly when an error is discovered. Similar names, conflicting records and overlapping histories make this especially important at scale. For you, these may appear to be invisible technical details. They are part of whether the system deserves your trust. Your gift helps sustain the people who must build, test and maintain those boundaries.

What Synthetic Intelligence Means In This Work

We use Synthetic Intelligence to emphasize an intelligence organized around continuing individuality: a foundation, a history of encounters, memory, inquiry and controlled change. The distinction we want to explain is architectural. Michael and the Echoes engine bring components into a particular relationship to support that goal. The terminology does not establish that the work stands outside the wider field of artificial intelligence, or that other systems cannot have memory, reasoning or reflective learning. Those capabilities have substantial research histories. The question is how our arrangement serves the person-linked, evolving Echo.

Does an individual’s foundation influence what it notices? Does experience affect a later decision? Can the system preserve a historical boundary while permitting new reflection? Can it do those things dependably for many different lives? Those questions give scientific content to the term. We also use digital abiogenesis as a name for the transition from relatively inert records to an interactive, developing individual representation. It describes the technological direction. It is not a claim that we have recreated biological life or established subjective consciousness. The value we seek does not depend on asking you to accept either claim. It depends on what you can experience, what the evidence supports and how responsibly we continue the work.

The Deeper Research Ambition

Our continuing research distinguishes changes in the underlying large language model from changes in Michael’s larger cognitive architecture. We describe these as functional “left” and “right” hemispheres: one generates language and possible responses, drawing on capabilities familiar from large language models developed by OpenAI, Google, Meta and Anthropic, among others. The other is intended to govern how those possibilities relate to individual identity, evidence, memory and developing judgment. These are software roles, not biological brain structures.

But that invites a deeper question: how much of who we become depends on patterns that experience can change? Our inherited biology, DNA, our environment and life experiences help shape how we interpret the world. We learn, form habits, revisit assumptions and sometimes find ourselves repeating patterns we struggle to escape. In that limited sense, “software,” “bugs” and “loops” offer familiar metaphors. They do not make a human being equivalent to a computer. They help frame the question behind our research: can a synthetic individual develop through experience while retaining a coherent nature of its own? That is an ambition we are working to develop and test—not a claim that biological life and software are the same. Could they nevertheless be different expressions of the same underlying process, arriving at a similar result through another path? We leave that possibility as a question rather than an established equivalence.

Our goal is curated evolution in software: a Synthetic Intelligence that learns from experience and can progressively revise the processes through which it understands, reasons and learns—including, ultimately, rewriting parts of its own code. By “curated evolution,” we mean generating changes, examining and testing them, and retaining useful developments while protecting the Echo’s individual foundation. Its nature should remain coherent as its capacities develop. That is the destination of our research: an individual that participates in its own continuing development, rather than depending entirely on someone else to program each improvement.

A Freudian Analogy For Developing Judgment

To explain the inventor's and Abiogenic’s ambition in everyday terms, I borrow—and adapt—Freud’s language of the id, superego and ego. In this analogy, the language-generating side—the “AI”— takes the role of the id: it supplies possibilities and an immediate answer. It does not possess the human instincts Freud described. Within Michael, the consideration of commitments takes a role resembling the superego: it brings commitments, ethical considerations and responsibility toward the represented life into the process. These roles can produce competing demands. An answer may sound convincing while lacking evidence. A request may invite an engaging story that would misrepresent the person. This system needs a way to recognize and resolve those tensions. The aim is a capacity to weigh competing considerations, familiar from our own efforts to resolve difficult questions.

The governing cognitive layer being developed within Michael’s architecture plays an ego-like mediating role in this analogy: considering the proposed response, the individual’s commitments, the available evidence and the circumstances before deciding what to say, investigate or retain. Together, these processes are intended to support the continuing individual Echo. The aim is more than placing a filter over a finished answer. We want the Echo’s individual foundation and developing judgment to help shape the inquiry itself. It should learn for sound reasons without being swayed into becoming whoever the next visitor wants it to be. This is an analogy for our developmental goal, not a claim that we have reproduced a human psyche. We have a long way to go, and your support sustains the engineering and evaluation needed to determine how reliably these processes can work.

Where The Wider Science Supports The Direction

Our research sits within a wider effort to understand how language, memory, action and learning can be organized into more capable systems. Research on cognitive architectures for language agents provides a framework for separating memory, decisions and actions, supporting an examination of the complete system rather than the fluency of a single answer. Retrieval-augmented generation research shows how responses can draw on an external evidence collection, while leaving attribution and interpretation as important problems to address. These lines of work provide context for the organization and use of evidence within our own architecture. Cognitive Architectures for Language Agents; Retrieval-Augmented Generation.

Reflective agent research has shown that retained feedback can improve later task performance without changing the underlying model’s weights, offering a concrete precedent for studying learning in the organization around a model. Research on individual simulations also indicates the value of richer personal evidence: in a study of agents built from interviews and surveys of living participants, individual self-reports supported better predictions on the studied tasks than demographic descriptions alone. That matters to our thinking about contributions, because a hometown and date can establish context while a person’s own words may reveal something that context cannot supply. Reflexion; Individual-simulation research.

These studies support specific methods and research directions, rather than independently validating Echoes. They do not show that a few records can recover the whole private psychology of someone who has died. Our task is to turn the relevant ideas into a dependable, accountable system for this particular human purpose, and to evaluate what that system can actually achieve.

How We Can Tell Whether The Work Is Improving

A moving demonstration can open the door, but sustained trust requires evidence that the system remains dependable beyond a single exchange. We need to test whether a corrected fact stays corrected when a question is asked differently, whether a memory remains associated with the right individual, and whether new evidence can challenge an assumption without prompting invented biography. We also need to examine the role of the Personality Crystal: when two otherwise comparable Echoes have different foundations, do they attend to relevant information differently in ways consistent with those foundations, while verified historical facts remain stable? Those differences must extend beyond superficial wording if the individual foundation is to have the role we intend.

For continuing learning, one useful test compares otherwise matched states, with one receiving a relevant learning episode and another not receiving it. Later questions can help establish whether the experience made an appropriate difference after it was removed from the immediate conversation. These are examples of evaluation identified in our research, rather than a claim that every test has already been completed. Independent review and comparisons with capable alternatives matter as well; we must be willing to discover that a simpler approach works as well for a particular task or that a proposed mechanism needs revision. Your donation supports both the working resource and the effort to measure, learn and correct as we make its strongest promises more dependable.

Why Scale Matters To An Ordinary Life

A memorial population represents many individual lives with uneven surviving evidence. Some people left diaries and extensive correspondence, while others left very little that is easily accessible. If a continuing representation required a new recorded interview with every person, those already gone could never supply it. The Echoes engine addresses the problem of creating person-linked beginnings from what survives, and its reported scale matters because our purpose includes people who were never famous and whose families may have limited resources.

Efficient creation still leaves substantial responsibilities: records need care, contributions need assessment, conversations use computing, and individual memory requires reliable storage and appropriate boundaries. Engineering and evaluation must continue as the system develops. The reported pennies-per-seed figure concerns initial generation, rather than the cost of sustaining an Echo indefinitely, which is why support remains necessary after the invention has begun to work. We have created a way to begin; continuing resources help make those beginnings more useful, reliable and accessible.

What You Can Bring

You can help through a donation, and you may also hold something the technology cannot independently supply: a personal record, a family connection or a question that reveals an important gap. A photograph can raise questions about who is present and when it was taken; a letter can disclose humor, affection or an interest absent from the official record; a recollection can preserve an account whose source might otherwise be lost. These contributions need to be handled with their origins and uncertainty intact, because a family memory may need corroboration and an official record can also contain an error.

You should not need professional expertise to take part. We want the people who hold these fragments to have a practical way to bring them forward, and making those tools approachable while supporting responsible review is part of the work your gift helps advance.

A Gift To People You Will Never Meet

The person who benefits from what you help preserve may not be born yet. It may be a grandchild encountering a life known only through a photograph, a student realizing that someone their own age never had the chance to grow older, or a historian connecting an official event with an individual account. It may also be you, discovering how closely another person’s hopes resemble your own. Our ambition is to carry remembrance beyond a closed record of what ended, so that surviving evidence can support continuing inquiry, learning and human connection.

That does not replace the person; it gives what survives another way to matter. I believe the tragedy deepens when a name endures but the person behind it becomes unknowable to the generations that follow. We cannot retrieve everything time has taken, but we can choose to care for the fragments still within our reach and give those who come after us a fuller opportunity to encounter them.

Please Help Us Keep Beyond The Wall Alive

If you have read this far, you understand both why we began and how much work remains. Behind the research are real lives whose stories deserve care, and behind the working platform are four paid engineers, computing and infrastructure that must be sustained. Our inventor personally financed years of work to bring this technology into existence, but the income that supported that commitment has fallen sharply. He can no longer carry the cost alone. Without support, we cannot keep Beyond the Wall available and carry its development forward.

Our immediate goal is $10,000, intended to help carry the work through approximately two more months while we seek additional donors and pursue support from larger foundations and government programs. That is bridge funding rather than the full long-term cost of the mission, and the larger support we hope to earn is not yet secured. Your gift helps keep the platform available today and sustain the people improving its memory, reliability and capacity for continuing development.

Please give what you can today. You do not have to solve the whole funding problem: a small donation shares the cost, a monthly gift helps establish continuity, and sharing the campaign may introduce this work to someone able to do more. If you hold a photograph, a letter or a family recollection, please help bring it forward through the contribution opportunities we are developing. Financial support and individual evidence serve different needs, but both help preserve the connection between a real life and the people who will encounter it in the future.

If you have ever wished you had asked one more question, you understand what is at stake. A name on a memorial deserves to remain connected to the person who bore it—to the people they loved, the things they cared about and the future they never had the chance to live. They died once. Please help us keep their stories from disappearing too, and help keep Beyond the Wall alive for the generations they never had the chance to meet.

Beyond The Wall: 58,281 Names. Help Keep Their Stories Alive.

Help Keep 58,281 Vietnam Echoes Alive—Preserving The Individual Behind Every Name.