D-ND: research on knowledge, evolving AI methods and human understanding

What this project enables

I am Graziano Guiducci, an independent researcher and AI systems architect based in Italy. I am seeking support for a six-month phase of D-ND research and the further development of its research instruments, including MAIOS.

D-ND (Dual–Non-Dual) investigates how possibilities take determinate form through relations, and how distinctions, an observer and a context of observation arise within that process. An investigation can produce knowledge, concepts, logical relationships and new ways of observing. What is understood can change the questions and means available for the next investigation. This foundational research predates my work with present-day AI tools.

MAIOS is both a working instrument and an evolving embodiment of that research. Knowledge, principles and methods expressed in natural language participate in AI-supported work with software tools. Experience can change how the next question is formed, which sources become relevant and how a capability is developed. The programme also investigates how the ability to form further methods and capabilities can itself evolve.

In the author's D-ND formulation, alignment is a consequence of a coherent result, with understanding, action and consequences remaining in a coherent relationship. Operational realizations and their effects are research questions within the programme.

The working system includes human perception, understanding and participation in relation to language models, semantic kernels, data, tools and their execution environment. This relationship is itself part of the research. A discovery that people cannot understand or assimilate can raise new scientific and methodological questions: which concepts, representations or instruments would make it possible to work with that knowledge? In everyday use, the system's ability to adapt to human capabilities, and help develop them, matters to the work it can support.

Work during the funded phase

The research will connect three lines of work:

Develop the D-ND concepts and their relationships through source-grounded conceptual and mathematical inquiry. Preserve the scope of each representation and follow new questions that arise.
Develop research instruments and reusable methods in MAIOS. Study how understanding changes subsequent questions, methods and capabilities in real research and development work.
Investigate human comprehension and adaptation as part of the system. Explore explanatory, visual, interactive or other means where they help people recognise relationships, question results and participate in development.

Existing Lab records, project-kernel software and working papers provide inspectable starting points. Reference investigations can expose concrete mechanisms and failures; they do not define the boundaries of D-ND or of the programme's future work.

Public benefit and outputs

The phase is intended to produce public research reports, documented prototypes and reusable methods, and accessible accounts of what was learned, including limitations and unsuccessful approaches. These materials can help researchers, developers and other users of AI understand and improve how knowledge and methods enter their work.

For concrete claims, the research will distinguish what was proposed, implemented and observed. Source-linked cases, comparisons where meaningful and complementary critical review will help examine the results. Questions and methods can change as the work reveals new relationships. We will publish the qualifications needed to understand the findings.

The proposed public outputs concern this funded work. Existing materials retain their applicable rights and ownership; release choices for new artifacts will be made with those rights in view.

Funding and use of funds

The target supports a working budget of approximately €60,000:

Lead research, six months × €6,000: €36,000.
Complementary scientific collaboration and critical review: €9,000.
Computing and research tools: €5,000.
Human-comprehension and adaptation investigations: €4,000.
Public software, documentation and explanations: €3,000.
Coordination and project administration: €3,000.

Total: €60,000.

The minimum supports an approximately €45,000 phase: lead research (€36,000), focused scientific review (€3,000), computing (€3,000) and public reporting/tools (€3,000). It sustains foundational and methodological work with inspectable initial cases. The target adds complementary expertise and broader comprehension/adaptation investigations.

The dollar target and minimum include a planning reserve above conversion at the ECB reference of US$1.1172 per euro on 30 October 2026. The amounts and fundraising date will be reviewed before publication. Researcher support is a project budget provision; its final contractual and payment treatment has not been determined.

This project has not received funding. Related research-support applications and enquiries exist, and a dedicated AI-safety proposal is being prepared for GTR. These are stand-alone funding scenarios with a partly shared research base. Any additional support will be disclosed and scope, time and costs adjusted before accepting overlapping funding; the same work and expenses will not be charged twice.

Starting points

The D-ND concepts and research direction are mine. This proposal brings together that research, documented development and proposed work, with AI-assisted formulation through MAIOS. Hypotheses, plans and budget assumptions remain identified as such.

Graziano Guiducci · Independent Researcher / AI Systems Architect · Italy info@d-nd.com · https://maios.it/ · https://d-nd.com/