The Winning Essays for the Big Questions About AI
Abolishing pandemics/ Getting out of the way of AI automation/ Learning from Honk Kong MTR's business model
TL;DR
- Jassi Pannu suggests dedicating billions to AI-driven biosecurity to end airborne diseases, arguing this dual-benefit intervention (welfare and tail-risk reduction) is achievable through embedded physical infrastructure like far-UVC lamps.
- Ege Erdil recommends that countries not on the AI supply chain should focus on fundamental economic policies like strong property rights, low capital taxes, and open regulation to attract investment and benefit from AI-driven growth.
- Michael Li draws a parallel between AI labs and Hong Kong's MTR, arguing that AI labs, like the MTR's rail service, will not be profitable on their own and must instead capture value from complementary assets, such as property development for the MTR or deployment rights and integrated services for AI labs.
- Pannu outlines a 4-step plan totaling $40-$60B over 10 years to end airborne transmission, involving research, demand guarantees, large-scale deployment, and political infrastructure.
- Erdil identifies political and legal stability, increased capital formation, and fixing essential industries as key policies for countries to thrive in an AI-automated world.
- Li identifies four defensible strategies for AI labs to capture value: government-granted deployment rights, accumulated RL reward data, forward-deployed integration, and data trusteeship over national datasets.
- The essays were selected from over 600 submissions to a contest posing "big questions about AI."