Topic overview nodes (ringed circles) mark the 14 domain areas. They carry no extra weight — every note is positioned by its actual semantic similarity.
Singularity is an experimental personal knowledge graph that uses AI to discover novel, non-obvious connections across wildly different scientific and philosophical domains.
Every day, an automated research pipeline scans frontier scientific literature across 14 domains — from mycology and consciousness science to sacred geometry and bioelectric intelligence — and distills findings into interconnected notes. An embedding model converts each note into a high-dimensional vector, and a community detection algorithm (Louvain clustering) reveals which ideas naturally group together, regardless of their field of origin.
Each node is a note. Position reflects semantic similarity — notes that are closer together are more similar in meaning, even if they come from completely different fields. The edges represent measured similarity between notes. Colors indicate algorithmically detected communities of related ideas.
The interesting part happens at the boundaries: bridge notes (triangles) mark places where ideas from different domains converge in unexpected ways. A study on how cells respond to sound might connect to research on tensegrity geometry and bioelectric signaling — not because someone planned that connection, but because the underlying science genuinely overlaps.
Modern science is deeply siloed. A breakthrough in mycology might directly inform AI network design, but the researchers will likely never read each other's papers. Singularity is an attempt to break those silos — to let a machine find the connections that disciplinary boundaries hide, and surface them for a human to evaluate, develop, and act on.