If you apply the same logic to hunger, for example, even more people would be obese if they ate when they felt hungry, or if they followed their ancestors' instincts to gorge until full whenever the opportunity arose.
It's possible that we've historically survived in a local maxima of resource scarcity, and that "optimal" metabolic performance exists in a place beyond that, where resources are less scarce.
How does infant mortality help us to reason that we should drink beyond thirst?
Why not? Humans generally avoided arid climates until fairly recently, so I wouldn't consider the idea of drinking 2 liters per day as absurd on its face.
Also, one of our killer features is the ability to cool our bodies down by sweating. That makes us incredible endurance hunters, who can eventually exhaust even much larger and faster prey by simply following it around until it collapses from excess heat. But that requires drinking water.
Depending on where you live, you will sweat out a lot more water on a daylong hunt.
But in semi-arid climates, where streams are rare, animals tend to gather around watering holes, so that might actually be the best place to start your hunt.
Thirst is an emergency signal not a useful metric to depend on planning water intake.