Mostly agreed but with that much data couldn't we easily have an adjusted CO2 number for local sensors ?

I would imagine it's still relative unlike temperature on the wrist (which is too affected by body temp)