Please explain why, because I'd intuitively expect the comparison between kgCO2 of bitcoin transactions to gCO2 of car driving to be valid, assuming underlying numbers are true and not off by many magnitudes which I cannot judge (as stated).