“The training compute of the largest AI systems has grown by approximately half an order of magnitude (0.5 ooms) per year for nearly a decade.”