I'm all on board with what he's saying it and how he's saying it.

I don't think the evidence is quite as compelling as required to "reduce the sugar that crosses the point of sale by, say, 5 percent every year" because many individuals are not harmed by sugar. The influence/persuasive/manipulative lead-in is saying a crime has been committed and saying innocence has to be proven.

Moreover human beings are not simply a sum of statistics and I generally like to leave people to make their own decisions. Statistical reasoning is in fact designed to be discriminatory against changes in the system being studied and outliers. Statistical reasoning is designed to prevent innovation.