More information does not guarantee better reasoning.
We are developing infrastructure aimed at the layer of reasoning quality. Modern AI can retrieve, synthesize, and generate enormous amounts of information, yet the quality of its reasoning depends on how evidence, assumptions, and uncertainty are managed. Our work focuses on the structural integrity of the reasoning process itself.
The Veritic Discernment Core
Truth before belief
We prioritize the objective reality of data over the subjective weight of human opinion.
Observation before interpretation
We separate raw sensory input from the cognitive biases that shape our initial understanding.
Discernment before conclusion
We evaluate the validity of assumptions before committing to a final, irreversible judgment.
What We Are Building
Our research is being translated into the next generation of reasoning infrastructure and human-AI systems.
In Development
Reasoning Infrastructure
Building the foundational layers of a structured reasoning architecture designed to manage uncertainty and expose hidden assumptions.
Coming Soon
AI Applications
Translating our research into practical tools for human-AI collaboration, focusing on the quality of the reasoning process itself.
Coming Soon
Human Discernment Tools
Empowering users with interfaces that facilitate structured observation and reduce the cognitive load of interpretation.
Coming Soon
Developer APIs
Engineering robust interfaces for developers to integrate our reasoning infrastructure into their own complex systems.
In Development
Research & Evaluation
Developing rigorous methodologies to measure the quality of reasoning, identifying degradation, and testing architectural improvements.
Coming Soon
The Veritic Corpus
A proprietary research corpus supporting development and avaluation.
Testing reasoning itself.
We are developing evaluation methods for comparing reasoning processes, identifying degradation, measuring improvements, and testing whether better reasoning architectures produce more reliable conclusions.
01
Process Evaluation
Our methodology focuses on identifying and testing the assumptions that accumulate during complex analysis, ensuring that conclusions are grounded in verifiable evidence rather than belief.
02
Assumption Testing
We are developing infrastructure aimed at developing proprietary methods for evaluating reasoning quality under uncertainty.
03
Uncertainty Metrics
Our research is documented through technical publications, experimental records, and datasets designed to support future development and rigorous benchmarking of reasoning systems.
04
Technical Publications
Kris Hollon
XKENOSIS was founded by Kris Hollon around the development of Veritic Discernment. We are building infrastructure aimed at the layer of reasoning quality, where evidence, assumptions, and uncertainty are managed with precision.
Built on XKenosis
A growing family of applications will bring the underlying research into practical use.
Coming Soon
01 / REASONING INFRASTRUCTURE
Veritic Core
Coming Soon
02 / AI APPLICATIONS
Discernment Tools
Coming Soon
03 / HUMAN DISCERNMENT
Human-AI Interface
Coming Soon
04 / DEVELOPER API
Core API
Coming Soon
05 / RESEARCH & EVALUATION
Research Evaluation
Coming Soon
06 / THE VERITIC CORPUS
Knowledge Corpus
Insights from XKenosis
Exploring reasoning, uncertainty, evidence, and the future of human-AI decision making. Read these articles.
Aug 15, 2026
Human Reasoning Benchmarks
Defining metrics for evaluating decision-making quality in collaborative human-AI reasoning environments.