the dataset is interesting, thanks for linking to it. the relationship is for now linear, which I believe doesn't invalidate the point that you and I both make that exponential is not sustainable. the GDP-energy consumption decoupling must also become exponential, which I lack vision for how to achieve.