What if AI weather models are solving physical equations after all, just not the ones in conventional NWP? Different models turn out to represent the atmosphere in strikingly similar ways, and the authors propose why: each mesh point is a particle in latent space, sliding down a learned free energy landscape - with GraphCast and Aurora working from large scales to small as the layers deepen.
13 Sept 2026
How AI predicts weather without physics
A long, unhurried walk through an open-source tool for poking at GraphCast's latent space: organise the processor's internal representations, run cosine similarity and PCA, and watch linear directions emerge that line up with mid-latitude waves and specific humidity.
12 Sept 2026
How AI weather models actually think
A simple, easy-to-use visualisation tool for exploring the latent spaces of AI weather models, and what it finds when pointed at GraphCast: directions in latent space that seem to track synoptic-scale waves and humidity.
Reach and audience
Public platform figures. Ratings count people who left a rating, not total listeners.
Apple Podcasts (US)
5.0 / 5
1 ratings
Contact The AIGS Pod
Guest appearances
Does not typically book guests
Based on episode analysis; this does not confirm that the show is currently accepting guests.
Host of The AIGS Pod?
Claim your podcast to manage its listing and keep your show details accurate.
Pod Engine is an independent podcast discovery and analytics service and is not affiliated with or endorsed by this podcast. Artwork and show content belong to their owners. Full legal notice.
Explore this show Podcast research with Pod Engine
