Unpredictable Atmosphere - lecture performance with Lucia Rebolino

Unpredictable Atmosphere reimagines the internet as an optical device to unfold climate models in digital space. It addresses the limitations of AI foundation models in climate forecasting, which often ignore outliers—extreme weather events such as hurricanes and heatwaves—that push the atmosphere to its breaking point. These events represent phenomena that cannot be fully modelled and remain hidden in AI’s blind spots. Unpredictable Atmosphere views the atmosphere as an archive of transformation, much like an architectural object, preserving technological residues. It crumbles conventions and explores how outliers in counter-modelling practices can reshape our understanding of climate, weather, and space.

Part of the LINA architecture programme.