Forecasting & the drivers of AI progress

Inbound Marketing & Sales

This podcast episode focuses on the work of Danny Hernandez, a research scientist at OpenAI, who is primarily focused on measuring progress in artificial intelligence. He examines three major drivers of AI progress: compute, algorithmic innovation, and data, and the podcast explores the relationship between these drivers. Hernandez also discusses the challenges of forecasting in AI, particularly in the context of highly impactful and unlikely events. He emphasizes the importance of calibration training for improving forecasting skills and argues for a greater focus on understanding the economic impact of AI to better understand its progress. The conversation also touches upon the importance of AI safety, with Hernandez highlighting the work of various safety teams at OpenAI, such as the Foresight team, which focuses on understanding macro trends in AI, and the Reflection team, which concentrates on ensuring AI systems are aligned with human values.