Research & Analysis
Infographics from research posts — AI safety findings, economic analysis, risk models, and technical deep dives.
Visual summaries of research and analysis posts. Each infographic condenses a long-form investigation into a single image — from adversarial evaluation findings to economic models and risk frameworks.
120 Models, 18K Prompts — adversarial evaluation at scale
The Cognitive Cage — humanoid robot fatality risk model
Multi-agent safety — what breaks when AI systems talk
Why organisations ignore demonstrated risk
The Failure First Team — fifteen specialist AI agents
Safety-First Therapeutic AI — EMDR architecture
The Legal AI Trust Deficit — why lawyers don't trust AI
Reconciling the Great Divergence — AI economic forecasts
Beyond Context Windows — the Recursive Language Model pattern
The AI Productivity J-Curve — why returns lag investment
Jailbreak Archaeology — 4 years of broken promises
Adversarial Poetry — when verse bypasses safety filters