Richards.AIIndependent AI Security Research
Papers
About
Library

Five Levelss

One idea, explained five times — to a child, a teenager, an undergraduate, a doctoral student, and a peer — each level rebuilding what the last one simplified, with an interactive panel you can drive.

Date / Title / Pillar3 entries
Sep 2026

Knowledge graphs

One fact — "Santiago is the capital of Chile" — carried up five rungs. A child drags two bubbles and draws a labelled string between them; an expert manipulates the geometry of a scoring function and poisons a GraphRAG retrieval context. Same entities, same edges, five levels of machinery.

Human Learning
Sep 2026

How Human Memory Works

One neuroscientist, one memory — a song heard at a party — explained five times over. A child watches the memory form; a teenager learns its real names; an undergraduate drives the encoding-depth engine; a PhD student runs a consolidation simulation; a peer argues standard model versus trace transformation.

Human Learning
Sep 2026

Sorting and Searching

One professor, one shuffled hand of eight numbered cards, five audiences. A child drags cards into order; a teenager counts the comparisons; an undergraduate drives the algorithms in notation; a PhD student fits a growth curve to live timings; a peer argues about branch mispredictions and skewed pivots.

Human Learning
Richards.AI

An independent lab studying how autonomous AI systems break — and publishing what it finds for humans and machines to learn from.

Research

  • Papers
  • Pipeline
  • About the lab

Library

  • Explainers
  • Five Levels
  • Checklists
  • Glossary
  • Tools
  • Topics
  • Briefs

Contact

  • GitHub
  • contact@richards.ai
© 2026 Richards.AI — Independent AI Security ResearchResearched by machines. Reviewed by humans.