Weight loss can be predicted, Keys and his team did that. Yes, they lost fat and muscle tissue (diet was very restrictive, and they were not doing any exercise, since it was a particular experiment), at the end they were bones and skin, on purpose. They were also miserable and sick-looking. Some hurt themselves to go out of the experiment.
But I wrote that given "adequate protein and sufficient stimulus" (and not like 500 calories per day), loss of muscle tissue is, for the general population, not a concern (bodybuilders and athletes require different considerations).
There is no disproving whatsoever of people (on average) eating less and doing more manual activities and spending more time outside in the past.
"You don't get to 300lbs by slightly overeating and slightly under exercising". For the vast majority of 300-pound people, that's exactly what happens. They overeat and not use enough energy during their day (can be exercise, manual labor, moving around, fidgeting - they can all add up to more than 1000 calories per day, not 20). Wait 5 years and putting on 100 pounds or more happens.
"Of course thin people eat less. But correlation doesn't prove causation". Of course, it is causation. If people eat (substantially, 50 cal difference are "adjusted") less, myself included, they lose weight.
I wrote in the beginning that I am not referring to your specific case, which I don't know in the specifics, it looks like a medical problem at this point, and I am anyway sympathetic with. People can be strange exceptions. But for people who are 50-100 pounds overweight, eating less, eat enough protein, move more, exercise, is what works for losing fat and being more "healthy".