

Cloudy with a chance of algorithms
23 snips Sep 26, 2025
Join Paris Perdikaris, an AI and deep learning expert from the University of Pennsylvania, as he dives into the evolving landscape of weather forecasting. He explains how AI enhances meteorology rather than replaces it and discusses the revolutionary Aurora system, which learns from historical patterns. Paris emphasizes the crucial role of extensive public weather data and raises concerns about funding cuts affecting forecasting accuracy. With optimism for future improvements, he highlights the untapped data sources still waiting to be explored.
AI Snips
Chapters
Transcript
Episode notes
Democratize Forecast Tools For Local Use
- Make AI weather tools accessible so people can run localized forecasts on personal computers.
- Paris Perdikaris says democratizing access helps users operate forecasts for their own regions.
Researcher Compares AI To A Sailor's Instinct
- Paris Perdikaris describes his decade of work applying deep learning to hurricanes, oceans, and weather systems.
- He compares a sailor's experience-based forecasting to how Aurora learns from historical examples.
AI Learns Weather By Pattern, Not Physics
- Aurora learns patterns from decades of global weather, ocean, and pollution data instead of solving physical equations.
- Paris Perdikaris says this lets AI forecast faster and sometimes more accurately than traditional models.