Blog

Fractal Chain of Thought (FCoT) is a structured reasoning methodology that analyzes complex systems, problems, or texts by executing exactly three distinct iterations, with each iteration operating at a progressively deeper layer of granularity.

The methodology evaluates the target space across three specific layers:

• Macro Context: The global, birds-eye view. This layer establishes boundaries, identifies overarching paradigms, and evaluates macro-level objectives.
• Meso Context: The structural and architectural view. This layer focuses on interfaces, intermediate steps, operational logic, and programmatic components.
• Micro Context: The execution and mechanical view. This layer analyzes the exact syntax, precise edge cases, variables, and literal constraints.

The execution of these iterations is guided by dual objective functions designed to balance breadth and depth:

1. Maximize $f_{max}$: Maximize the coverage of global structural elements, systemic shifts, and completeness of the analysis.
2. Minimize $f_{min}$: Minimize noise, superficial fluff, unverified narrative claims, and redundant text.

Following a hill-climbing optimization approach, each iteration builds upon the last to refine the output, culminating in a highly structured, scannable response.