No, that's totally not what he claims, although that might be what he thinks he argues.
There are enough data points of (thin person, does not exercise or eat properly) and (fat person, exercises and eats "properly") to make that point basically wrong.
Perhaps the problem is with the definition of "properly". But the causation is not total (that is, there are reasons other than food and exercise that cause people to become fat, some known and some unknown), and the population correlation is time reversed to what you would expect (people stop exercising AFTER they gain weight, not BEFORE), to make this not a trivial matter at all.
