It's not a very good example, because making food halal requires businesses to change the way they do things. One of the things required for food to be halal (AFAIK) is that the person doing the slaughtering must be Muslim. I can see how this requirement would prevent most businesses from "making all things halal", just by itself.
Besides, France's population is 5 to 10 percent Muslim, which negates the "three to four percent" 'dictating' claim of the article, especially because, last time I visited France, most meats I saw weren't halal.
My biggest problem with this Taleb piece is that it doesn't really take into account the probability of success from a particular intolerant minority. In practice there's a kind of signal-strength effect where the further up the chain you go the less effective any one person becomes. (At least for the extended GMO example.) Once you factor that into the model Taleb's argument falls apart.
eg. Let's say a given family has a 25% chance of their intolerant child swaying food preferences. (This is very generous, because I'm fairly sure in practice it's under 5%.)
And we'll assume that going up the chain you need at least a 25% threshold to accomplish anything. Then you're looking at multiplying probabilities that get increasingly small the further up the chain you go.
The cool part of this is that you could probably write a quick model in your language of choice to explore it.
If commenter apathy is the reason we see this article several times a month, allow me to chime in as a minority who is intolerant of repeat submissions.
