Would viewing your use of AI in terms of liters of water affect how you prompt?

There is one interesting article from the newsletter that turns your use of AI into gallons of water. I don't know how accurate the estimation might be in general, and I wouldn't rely on the numbers too much. But still, the thought stuck in my mind.

I always generate an answer five times if the first answer is a little bit off as I am not thinking about the cost of the extra attempts because the button is free to use, and I just click it. It would definitely change my behavior if there is a visible meter, as it does with the battery percentage on my smartphone.

My guess is it would work best for lazy prompts. Vague prompt, average result, retry, repeat.

Will such a meter change your behavior or would you ignore it in two days like everyone ignores other dashboards? It is really interesting if someone tried that in their own product.