Open Collider

Open Collider

Semantic collision engine for non-trivial ideas

Description

Ask an LLM for ideas and you get bland, familiar AI slop. Open Collider is a semantic collision engine that injects structurally distant knowledge domains into prompts and forces them to collide, surfacing non-trivial ideas.

It operationalizes Koestler's 1964 bisociation theory with empirical validation in a companion research repo.

Features



Domain collisions:Distant knowledge injected.

Bisociation:Grounded in theory.

Non-trivial ideas:Beyond the obvious.

Validated:Research data.