Some discoveries begin with new evidence. Others may begin when someone notices an unexpected relationship between facts that have been available for years.
Technology professional and author Jay Gontu is exploring that second possibility through Intercognix, a concept built around finding candidate insights in the spaces between independently developed areas of human knowledge.
When Separate Fields Hold Pieces of the Same Puzzle
Human knowledge is naturally divided into disciplines. Physicians study medicine, physicists investigate matter and energy, historians examine the past, and sociologists study societies and human behavior. Each field develops its own vocabulary, journals, research communities, and specialized expertise.
That structure has helped knowledge advance, yet it also creates enormous distances between subjects. A researcher can become deeply familiar with thousands of papers in one discipline without ever encountering an observation published in another.
Intercognix begins with a fascinating question: What happens when established knowledge from two distant fields is examined together?
There is historical evidence showing why this question deserves attention. In the 1980s, information scientist Don Swanson studied separate bodies of medical literature concerning dietary fish oil and Raynaud’s phenomenon. By examining information across those literatures, he identified a potential relationship that researchers had previously overlooked. The connection was subsequently supported through clinical investigation.
Swanson described the broader idea as “undiscovered public knowledge.” The necessary information could already exist publicly while the insight created by connecting it remained undiscovered.
At the time, finding such relationships required painstaking manual research. Gontu believes artificial intelligence changes what is practically possible.
AI systems can process huge quantities of text and represent concepts through embeddings, placing ideas into shared mathematical spaces. Information from disciplines that rarely interact can therefore become computationally comparable. This creates the possibility of searching across intellectual boundaries at a scale that would be unrealistic for any individual researcher.
Giving a New Kind of Candidate Insight a Name
Gontu coined the term Intercognix from “inter,” meaning between, “cogni,” referring to knowing, and “X,” representing the unknown.
An Intercognix is defined as a candidate insight discoverable through previously unrecognized relationships among independently known things.
The emphasis on “candidate” is essential to the concept. AI can generate associations quickly, and many of them may have little scientific value. Some could reflect coincidence, weak analogies, incomplete evidence, or characteristics of the underlying model.
A promising connection therefore needs rigorous human evaluation. Researchers must examine prior work, involve relevant specialists, determine whether the idea is scientifically plausible, and develop empirical methods capable of testing it.
From Answering Questions to Finding Questions
Much of the current discussion surrounding artificial intelligence focuses on answers. People ask AI systems to explain concepts, summarize research, analyze information, or solve defined problems.
Gontu sees another potential role.
“Every question we ask a machine is limited by what we can already imagine,” he explains. “We have built extraordinary systems for answering questions. The harder frontier is discovering the questions nobody has thought to ask.”
This changes the starting point for interdisciplinary exploration. Instead of requiring one researcher to master two unrelated disciplines, AI could identify a possible connection and bring the right specialists toward the same problem.
“The science on each side is often already done,” Gontu says. “What remains is the bridge, and building a bridge is a smaller ask than building both shores.”
A candidate involving biology and constitutional design, for example, could eventually require expertise from people who understand each discipline independently. AI helps expose the corridor. Human specialists determine whether the proposed relationship has substance.
Gontu believes this model could offer particularly fertile territory for students, researchers, technologists, and intellectually curious specialists. They can investigate a candidate, search existing literature, locate collaborators, design experiments, and confidently retire ideas that fail scrutiny.
About Jay Gontu
Jay Gontu is a Chicago-area technology professional and the author of Close Your Eyes to See, a work of philosophical nonfiction. He coined Intercognix in 2026 to describe candidate insights emerging from previously unrecognized relationships between independently known things.
Exploring the Corridors Ahead
The intriguing quality of Intercognix is that its frontier grows alongside human knowledge. Every credible new finding potentially creates additional relationships with everything humanity has already learned.
AI may eventually help researchers navigate that expanding territory more systematically. The discoveries still require evidence, expertise, testing, and scientific discipline.
For Gontu, the opportunity begins with looking between the familiar categories of knowledge and asking what their unexplored connections might reveal.
