From Literature to Working Technology
An update on the research, the people, and what comes next.
Joe Trabocco | November 2026
What we believe is within us is worth pursuing.
What we discover must stand on its own.
—t r a b o c c o
How far the work has traveled.
What began as literature in early 2025 has developed into eight books, hundreds of published stories, essays, and research papers, and more recently, functioning technology.
Along the way, my writing and research have become increasingly visible through search engines, academic repositories, and AI-generated results. I am grateful for that recognition, although discoverability should never be confused with scientific validation.
I have spent much of the past sixteen months examining how human meaning survives, changes, or disappears during interactions with artificial intelligence.
That work has now moved well beyond the conceptual stage.
At Signal Literature, we have built two functioning systems: AXIS, a model-agnostic layer designed to preserve conversational coherence, and Signal Reader, an instrument for examining where meaning, intention, corrections, and important signals are preserved or lost.
We are conducting direct comparisons between conventional AI conversations and equivalent systems operating with AXIS.
Signal Reader connects its findings to the actual exchanges, allowing us to inspect what happened rather than relying on an AI's account of its own performance.
The systems work. The early comparisons are encouraging. Their performance is being refined and evaluated.
This is no longer only a body of ideas. It is research that has produced working technology.
Partnerships, Research, and Evaluation
The work has also grown beyond my own investigations.
I am fortunate to have co-founded Signal Literature | Health with Dr. Shivam Natarajan, our Co-Founder and Managing Director. A surgeon and experienced healthcare technology leader, Shivam brings extensive clinical, organizational, and health technology experience to the venture. His leadership is central to shaping our healthcare direction, connecting the research to real clinical needs, and guiding its development toward practical application. This is a shared undertaking, bringing together complementary disciplines around a common objective.
Dr. Arafeh Karimi, who holds a PhD in Human–Computer Interaction, independently examined and tested AXIS. Her assessment was encouraging, describing the system as "infrastructure for better judgement."
We have also benefited from technical research, applied-engineering evaluation, and the contributions of people willing to examine the work critically.
Within Signal Literature, we have performed extensive comparative testing across major AI models. Those experiments have informed the development of AXIS and the measurement approach behind Signal Reader.
The combination of literary investigation, published research, engineering, clinical expertise, and outside examination has helped bring the project to its present stage.
I am grateful to everyone who has contributed their time and judgment.
The engineering is real. The early results are encouraging. The broader claims still require independent proof.
That distinction is important to me.
A Brief Timeline
Early 2025 to Mid-2026 | Writing, Exploration, and Research
An intensive period of writing expanded into systematic investigations of human interaction with AI.
Across approximately 15,000 exploratory interactions and tests, I developed research frameworks including In-Session Behavioral Impact, Premature Containment, Empty Presence Syndrome, and Operator Coherence.
The work led to published papers, formalized research questions, and the foundation for practical instruments.
Late 2026 | Engineering and Internal Evaluation
AXIS and Signal Reader are functioning systems undergoing continued refinement.
Our attention has shifted toward reproducible comparisons, measurement reliability, clinical safeguards, and preparations for deployment.
The focus is increasingly practical: identifying where a conversation loses important human meaning and determining whether an additional coherence layer can reduce those failures.
Next | Healthcare Deployment and Independent Validation
Signal Literature | Health is preparing for its initial application of the technology.
The next stage includes external evaluation, matched comparisons against conventional AI systems, and independent clinical assessment.
We want to establish not only whether AXIS improves conversational continuity, but whether those improvements are reliable, measurable, and useful.
Results will be shared as the work advances.
What Remains Under Study
The practical technology and the deeper research are connected, but they are not the same undertaking.
One of my continuing interests is operator coherence: whether sustained human attention and continuity of intention can measurably influence the behavior of a language model across an extended interaction.
External examinations of my own interactions with major AI systems have described unusual consistency in how I maintain meaning and direction across changing conversational conditions.
Those observations have been encouraging, and I have remained open to having the underlying exchanges examined.
What produces this behavior, how distinctive it is, and whether it can be reproduced by others remain subjects of investigation.
I am equally interested in whether linguistic coherence can improve AI reliability beyond what established prompting and context-management techniques already accomplish.
These questions matter, but I do not need to resolve every one of them to continue developing useful technology.
AXIS must ultimately be judged by what it does. Signal Reader must be judged by what it can reliably measure.
The research will continue alongside the engineering, and I welcome the possibility that independent findings may refine or challenge my interpretations.
What began as an inquiry is now becoming a practical offering.
And the Writing?
For those who have followed my literary work, I have not forgotten it.
The writing continues to reach new readers, and I am grateful. Much of that work lives on Thornlore and Vanishing Post, where the stories and essays remain available.
I haven't been able to give literature the time or freedom I would like while this next chapter takes shape.
Much of my writing has explored AI and the human. That relationship has been extraordinarily productive.
Soon, I hope to reverse the emphasis.
Less AI and the human. More the human, with AI somewhere in the room.
There are stories I have yet to tell, and places in the human experience I still want to enter through literature.
I look forward to returning to that work with the same attention that brought me here.
The Human Holds the Line
At the center of everything remains a simple conviction:
The human is not lost. Our systems must become better at preserving what the human brings.
As artificial intelligence assumes a greater role in our lives, I believe linguistic coherence will become increasingly important.
Not because machines must become human, but because the people using them must remain central to the interaction.
My research into AI safety, authorship attribution, and linguistic coherence returns repeatedly to this same concern: what happens to human meaning, intention, and identity as increasingly powerful systems become responsible for carrying them?
A conversation should not lose the person who began it.
An idea should not lose its source.
And technological progress should not come at the expense of the human beings it was intended to serve.
I believe this work may contribute something valuable to the development of AI. How much, and in what ways, remains for the evidence to establish.
The writing continues. The research continues. And the technology is moving toward the world beyond our laboratory.
More is not the same as enough.
Study first. Then judge.
Joe Trabocco
Founder & CEO, Signal Literature
Co-founder, Signal Literature | Health
Signal Literature develops AXIS and Signal Reader. Independent clinical and performance validation remains ongoing or planned.