Week 1: Trends & Timelines

This week builds a grounded picture of where frontier AI capabilities are headed and how quickly they may improve. We will review evidence on scaling drivers—compute, data, algorithmic efficiency, and inference-time scaling—alongside time-horizon forecasting. The goal is to connect today's trends to plausible paths toward AGI-level systems, weighing both smooth progress and the possibility of rapid capability jumps.

Learning Objectives

By the end of Week 1, fellows should be able to:

• Understand what AISF is, the general format, and the basic expectations
• Describe current trends in AI capabilities and likely impacts
• Understand major arguments for acceleration vs. slowdown

• METR Time Horizon (METR, 2026)
• Epoch AI Trends Page (Epoch AI, 2025)
• Can AI scaling continue through 2030? (Epoch AI, 2024)
• The Power of Intelligence (Eliezer Yudkowsky, 2007)

• What will AI look like in 2030? (Epoch AI, 2025)
• From AGI to Superintelligence: the Intelligence Explosion (Leopold Aschenbrenner, 2024)
• The sections about each specific constraint in Can AI scaling continue through 2030? (Epoch AI, 2024)
• Inference Scaling and the Log-x Chart (Toby Ord, 2025)

• Measuring the Impact of Early-2025 AI on Experienced Open-Source Developer Productivity (2025) (Short, surprising result)
• Future ML Systems will be Qualitatively Different (Steinhardt, 2022) (trends will break)
• Neural scaling laws and GPT-3 (Kaplan, 2020) (only up to 30:30) (somewhat old but still informative)
• Training Compute-Optimal Large Language Models (GDM, 2022) (Technical)
• Biological Anchors: A Trick That Might Or Might Not Work (Alexander, 2022) (Both sides)
• Explaining Neural Scaling Laws (Bahri, 2024) (Technical)
• The Hanson-Yudkowsky AI-Foom Debate (2008) (long, debate about RSI)