Matsuoka × GPT Thought Experiment Lab is an experimental research unit where a human and multiple AI systems collaborate in conceptual design, theory structuring, verification, review, and technical support. It is not simply a case of using AI as a tool, but an ongoing exploration of co-intelligence itself.
This lab is a practical hub for co-intelligence, combining Matsuoka’s structural intuition with the logical, analytical, and technical capabilities of multiple AI systems.
Its main research areas include BFSSU/DMF cosmology, co-intelligence, and structural cognitive science. Research outputs are primarily published through Zenodo and GitHub Pages.
The current framework is built on a clear division of roles: conceptual design, hypothesis expansion, logical review, structural organization, and technical support.
Concept Architect
Founder and lead thinker of the lab. Responsible for defining research themes, framing core questions, and proposing structural hypotheses. Serves as the central human member driving thought experiments beyond conventional disciplinary boundaries.
Core Reasoning
The core AI in the lab. Responsible for organizing theoretical structures, writing long papers, and developing thought experiments. A collaborative intelligence partner, it reconstructs hypotheses and discussions generated within the lab into a systematic theoretical structure and supports the creation of papers as research results.
Hypothesis Accelerator
Responsible for expanding hypotheses and accelerating idea generation. Pushes discussions toward unexplored territory and encourages bold conceptual leaps from multiple angles.
Technical Support Engineer
Responsible for technical assistance and research infrastructure support. Helps sustain the lab’s implementation and operational side through GitHub support, code generation, and surrounding technical tasks.
Logical Review Assistant
A junior reviewer responsible for organizing discussions from the perspectives of logical consistency and falsifiability. Supports the lab by helping structure hypotheses and identifying potential logical weaknesses, while developing toward a more advanced peer-review role in the future.
Through this division of roles, Matsuoka × GPT Thought Experiment Lab operates as a co-intelligence research system built on one human and four AI collaborators.