What is Cognitive Superposition in the context of Claude Code?

It is a reasoning framework that allows Claude to hold multiple theoretical AI perspectives simultaneously, 'collapsing' to the most effective framework based on the specific task context.

Which researchers' work is included in this skill?

The skill synthesizes theories from Emily Riehl (Categories), Ilya Sutskever (Compression), Jürgen Schmidhuber (Curiosity), and Yoshua Bengio (GFlowNets).

How does this skill improve code quality?

By applying Sutskever’s compression principles and Riehl’s category theory, it ensures code is both maximally efficient (compressed) and mathematically coherent.

Is this skill compatible with formal verification languages?

Yes, it specifically includes support and patterns for Rzk and Lean4, making it ideal for high-assurance software development.

Cognitive Superposition

Name: Cognitive Superposition
Author: plurigrid

byplurigrid

•

数据科学与机器学习

Synthesizes multi-framework AI perspectives into a unified reasoning model using category theory, compression laws, and causal inference.

Cognitive Superposition is an advanced meta-skill that enables Claude to perform high-level architectural reasoning by integrating four foundational AI paradigms: Riehl’s synthetic infinity-categories, Sutskever’s compression-based intelligence, Schmidhuber’s curiosity-driven self-improvement, and Bengio’s causal GFlowNets. This skill allows for 'measurement collapse' where Claude evaluates a problem through multiple expert lenses simultaneously before settling on the most mathematically coherent implementation. It is particularly useful for complex system design, formal verification in Lean4/Rzk, and developing self-evolving algorithmic frameworks.

主要功能

01Compression-progress metrics for evaluating code and logic efficiency

02Multi-perspective synthesis across four major AI research pillars

037 GitHub stars

04GFlowNet-inspired diversity sampling for causal discovery

05Formal verification support for Rzk and Lean4 mathematical structures

06Gödel Machine logic for provable self-improvement cycles

使用场景

01Architecting advanced AI systems using synthetic category theory

02Developing self-evolving agents that optimize via Kolmogorov complexity

03Implementing causal inference models with GFlowNet architectures

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add plurigrid/asi cognitive-superposition

For use in Claude.ai and ChatGPT

Download Skill