AI Forecasting: Two Years In

·Bounded Regret··

Two years ago, I commissioned forecasts for state-of-the-art performance on several popular ML benchmarks. Forecasters were asked to predict state-of-the-art performance on June 30th of 2022, 2023, 2024, and 2025. While there were four benchmarks total, the two most notable were MATH (a dataset of free-response math contest problems) and MMLU (a dataset of multiple-choice exams from the high school to post-graduate level). One year ago, I evaluated the first set of forecasts. Forecasters did poo...

Read full article →

Related Articles

Accelerating Gemma 4: faster inference with multi-token prediction drafters
amrrs · Hacker News · 3d ago
ProgramBench: Can language models rebuild programs from scratch?
jonbaer · Hacker News · 1d ago
ZAYA1-8B matches DeepSeek-R1 on math with less than 1B active parameters
steveharing1 · Hacker News · 1d ago
OpenAI’s o1 correctly diagnosed 67% of ER patients vs. 50-55% by triage doctors
donsupreme · Hacker News · 6d ago
A couple million lines of Haskell: Production engineering at Mercury
unignorant · Hacker News · 6d ago