I don't get how you have considered all these details yet didn't try to steelman the "hint" better, e.g. 30 minutes of relaxed meditation compared to 30 minutes of sauna usage, as opposed to some vague definition of "do nothing" and whether different social classes effectively have very different baselines of doing nothing, such as their stress levels, does playing golf count as free time, or sunning on the deck of a cruise ship is that "doing nothing", etc. at which point the discussion about confounders really gets in the weeds. Unlike CPUs human in/activity is not like a no-op instruction