I asked about comparative water wastage between calorie checking on Google vs. AI vs. calorie counting app: Claude ai: A dedicated calorie-counting app like Nutracheck or MyFitnessPal is largely doing local computation on your phone. When you log a meal, you’re mostly: Querying a pre-downloaded food database (or a small cloud lookup) Running simple arithmetic locally Syncing a few bytes of data to their servers There’s no large language model involved, no GPU cluster spinning up, and the data transferred is tiny — typically a few kilobytes per food entry. The energy cost of that network request is in the range of fractions of a milliwatt-hour , which translates to a negligible fraction of a milliliter of water — essentially immeasurable in any practical sense. So the rough hierarchy looks like: The app wins by a large margin — simply because it’s not invoking a massive neural network. It’s closer to using a calculator than talking to an AI. That said, a couple of caveats worth keeping in mind: If the app uses AI-powered features (like photo food recognition), that call to a vision model would bump its footprint up significantly. The app has its own background data usage — syncing, ads, tracking — which can add up, but is still small per individual interaction. All three options are still dwarfed by the water cost of the food you’re logging. A single apple takes roughly 70 liters of water to grow. So any way you count those calories, the counting itself is environmentally negligible compared to the eating. submitted by /u/MrsChatGPT4o
Originally posted by u/MrsChatGPT4o on r/ArtificialInteligence
You must log in or # to comment.