About This Report iii

Summary v

Figure and Tables xi

CHAPTER 1

Introduction 1

Report Objectives 1

Methodology 2

Chapter 1 References 6

CHAPTER 2

Defining the Target 9

Common Definitions 9

The Three-Stage Framework: Capability, Deployment, Societal Transformation 10

Why Definitions Matter for Strategy and Policy 11

Chapter 2 References 12

CHAPTER 3

Forecasting Methodologies 15

Expert Surveys and Elicitation 15

Crowd Forecasting and Prediction Markets 16

Compute-Centric Models 16

Empirical Trend Extrapolation 17

The Market Weight Signal 18

Interpreting the Forecasting Ecosystem: Three Archetypes 19

Chapter 3 References 20

CHAPTER 4

Understanding Disagreement: The Cruxes 23

The Capability Crux: Scaling, Architecture, and the Limits of Current Approaches 23

The Diffusion Crux: Speed of Transformation 24

The Takeoff Crux: Speed of Capability Gain 26

Summary: How Cruxes Generate Positions 27

Chapter 4 References 28

CHAPTER 5

The Validation Challenge: Limitations of Current Methods 31

The Benchmark Problem 31

Immaturity of the Forecasting Ecosystem 32

Institutional Independence Concerns 33

The Calibration Problem 33

Reflexivity: Forecasts Shape What They Predict 33

Foundational Objections 34

What Can We Rely On? 35

Chapter 5 References 36

CHAPTER 6

Strategic Implications: Decisionmaking Under Uncertainty 39

The Asymmetry of Costs 39

Policies, by Timeline Sensitivity 40

Distinguishing Strategy and Policy Triggers 40

Indicators for Ongoing Assessment 41

Summary: Principles for Strategy Under Uncertainty 42

Chapter 6 References 43

CHAPTER 7

Building Better Forecasting Infrastructure 45

Adversarial Validation of Forecasting Models 45

Intermediate Milestone Forecasting 46

Robust Capability Measurement 47

Economic Diffusion Modeling 48

Monitoring Infrastructure 48

Chapter 7 References 50

CHAPTER 8

Conclusion 53

APPENDIX

Evolution of AI Forecasting Approaches 55

Abbreviations 63

About the Authors 65