> In fact the majority of their food database was community supplied. And it shows. Values for tons of entries seem to be random guesses, with a bias towards l

> In fact the majority of their food database was community supplied. And it shows. Values for tons of entries seem to be random guesses, with a bias towards lowball, guilt-assuaging numbers. I use MFP regularly, but I need to pad whatever it says by 20-30% unless I'm eating factory-packaged food or measured homemade food. > would largely stand the test of time (while specific products come and go, base facts about specific foods rarely do) Really? Not only do prepared products change their formulations plenty often (as you note), but do you really think the heirloom broccolis of 1950 have the same nutritional profile of the industrial-agriculture broccoli of 2020? Whether you support industrial agriculture or fear it, it directly drives change to the structural and chemical properties of the things it grows. Macronutrients and especially micronutrients are going to be effected by that. Public nutrient data seems like a really valuable social good, but community-sourced data for this kind of thing is just going to create a noisy, sloppy data set (as it has on MFP).