Questions and answers

Temporal Data Dynamics is a San Diego-based company developing Generative Temporal Referencing, the science of using time as a substrate for organizing, assessing, and investigating data, and the method for recovering the value held in records with AI. The work relies on a collaboration between a person who knows their field and AI models that can search, research, and process data at scale across a myriad of systems and formats.

GTR is a process of parsing complex corpora of varying types and sizes, across a temporal spectrum, and providing a reference related to its placement on this temporal spectrum. The process is similar to Geographical Information Systems that use location as a substrate to assign data and values to; GTR uses time as a substrate to reference various data elements to their proper place in time. How it is put to work is set out on The Method.

The Directed User Interface (DUI) establishes a collaborative structure that provides a hierarchy of roles and responsibilities and most importantly, an assignment of authority so the user remains in the driver's seat and in control of the session. In the 1990s, the Graphical User Interface (GUI) replaced the command line for data programming and put personal computing within everyone's reach. DUI is designed to do the same for Artificial Intelligence by allowing non-technical users to access and unlock the dynamic capabilities of AI by using normal conversation in the user's register.

No. DUI was built for non-technical users, who work with it using normal conversation in their own register. There is a learning curve, and the user's knowledge of their own field matters more than technical skill.

GTR was developed on Claude and runs within an AI service the user already has; it has also been used with other services. Results depend on the service and the plan: features such as file uploads, file creation, and longer sessions make a difference. Claude is a product of Anthropic; Temporal Data Dynamics is not affiliated with Anthropic.

With the paper closest to your question: GTR for the science, Latent Value Query for the value held in records, the rectification train for damaged text, the Directed User Interface for governing work with AI, and the data commodity exchange for where it could lead. All are free to read on Publications, and the book develops them in full.

All situations are unique and applications are designed with you for your specific data requirement. Please feel free to reach out for an initial consultation to see how this approach can fit your needs.

Inquiries

With a question. If your records raise one the method could answer, send us an inquiry.

Records a user already holds or can obtain: documents, filings, logs, correspondence, and data exports, across systems, disciplines, and document types. The case studies on this site worked from public records such as budgets, meeting records, maps, and filings.

No. Licensed professional work stays with licensed professionals. GTR is intended to support licensed professionals across a wide spectrum of functions and an array of capacities.

GTR works inside the user's own AI session. It manages the session, and the session manages the data, so records go only where they would go without it: between the user and the AI service they choose, under that service's terms. Temporal Data Dynamics does not receive them.

The part of the method that gives each record its place in time, so that records begin to explain one another and gaps and contradictions become visible. It is described in the book and the papers.

A conceptual model of how an exchange for recovered records might work: appraisal, de-identification, published standards, and certification. It is set out in a paper, not an operating business. See The Market.

Tests of the method on different kinds of records and different questions, each followed as far as the available data would allow. They are described on the Applications page.