1. Phelps seems to be guessing: "Maybe eight to ten thousand calories per day," he writes... Also, the research shows there is extreme variation in people's resting BMR (perhaps due to large variation in height, body weight, stress levels, etc), so you eating 3500 calories isn't decisive against the thesis of this article.

2. That is exactly his hypothesis, and he has produced evidence for this hypothesis in a recent experiment. I don't know why you are presenting this as some kind of counter-argument.

3. That blog post isn't a coherent critique. The post's two points are that the drinking solution is also used for measuring energy intake, and that the Hadza actually burn far less calories than Western people until such numbers are adjusted. This first point isn't relevant without further explanation, and the second point only serves to further support the thesis.