YANDI Core

An experimental architecture exploring the boundaries of adaptive AI.

YANDI Core explores whether a relatively small AI core can form and modify external structures of memory, knowledge and specialized intellectual modules.

Research directions

External memory and concepts

Forming, storing and revising representations outside the model's main weights.

Experience and verification

Learning from experience, choosing tools and checking the system's own results.

Specialized experts

Connecting separate modules instead of putting every capability into one core.

External storage

Exploring SSD and other storage as a long-term part of the computational environment.

Status

This is a research direction, not a finished AGI and not a promise of autonomous self-development. Implemented experiments must be distinguished from hypotheses and future ideas.

Status: research / prototype.