Case Studies

This section contains concrete examples—both successes and failures—that illustrate how delegation risk principles apply in practice.

AI System Case Studies

Direct applications to AI systems, from failures to successes:

Case	Outcome	Key Lesson
Sydney (Failure)	Acute failure	What happens when constraints are missing
Code Review Bot (Success)	8 months in production	Sustained success through proper design
Support Bot (Near-Miss)	$50K error caught	Verification layers save the day
Content Moderator (Drift)	Gradual degradation	How small changes compound
Anti-Patterns	Common mistakes	What not to do

Human System Case Studies

Delegation risk principles aren’t new—human organizations have developed sophisticated approaches:

Case	Domain	Key Insight
Alliance Cascades	Geopolitics	Hidden alliances create systemic risk (WW1)
Criminal Organizations	Illegal enterprises	Trust without legal enforcement
Organizational Trust	Corporate hierarchies	Formal delegation structures
Jury Systems	Legal system	Adversarial trust filtering
Nuclear Launch Authority	Military	Extreme consequence management
Open Source Communities	Software	Distributed trust at scale
Cross-Civilization Trust	Historical	Trust across vast differences
The Oversight Dilemma	Governance	Who watches the watchers?

Power, Agency, and Authority Case Studies

Analysis of real systems through the lens of agency and power formalization:

Case	Power	Agency	Key Insight
Central Banks	Very High	Low	Institutional constraints limit agency despite high power
Recommendation Algorithms	High	Medium	Optimization creates emergent agency
AlphaFold	High	Very Low	Strong tools are achievable in narrow domains
Corporate Boards	High	Medium	Structural constraints substitute for alignment
Self-Driving Cars	Medium	Medium	Bounded agency through hard constraints

See Power Dynamics Case Studies for full analysis.

The Anomaly Chronicles

Fiction

The Anomaly Chronicles is speculative fiction, not a case study of real events. It explores how delegation principles might apply to hypothetical extreme scenarios.

A narrative exploration of delegation risk principles scaled to extreme scenarios:

Chapter	Focus
Introduction	Setting and premise
Containing Mr. X	Original containment engagement
Mr. X: A Perspective	The agent’s viewpoint
Power Struggles	Multi-agent coordination
Task Architecture	Structural approaches
Five Years Later	System evolution
Year Ten	Long-term dynamics

Reading Suggestions

New to the Framework

Start with the AI system cases—they’re most directly applicable:

1. Sydney (Failure) — See what goes wrong
2. Code Review Bot (Success) — See what goes right
3. Anti-Patterns — Learn what to avoid

Looking for Inspiration

The human systems cases show established approaches (with varying track records):

1. Nuclear Launch Authority — Extreme stakes
2. Criminal Organizations — Trust without enforcement
3. Open Source Communities — Distributed trust

Want the Full Picture

The Anomaly Chronicles explores where these principles lead at scale:

1. Start with The Insurer’s Dilemma for context
2. Then read the Chronicles in order

How Case Studies Connect to Principles

Principle	Best Illustrated By
Least Capability	Sydney (violation), Code Review Bot (success)
Verification Layers	Near-Miss (saved by verification)
Trust Decay	Drift (gradual failure)
Adversarial Oversight	Jury Systems, Oversight Dilemma
Bounded Exposure	Nuclear Launch, Anomaly Chronicles

Next Steps

• Apply these lessons → Design Patterns
• Understand the theory → Delegation Risk
• See other domains → Cross-Domain Methods