Finally read the thing. It's pretty interesting. To be honest, I only read it because one of the authors is Nassim Taleb and I liked his books. Anyway, the gi

Finally read the thing. It's pretty interesting. To be honest, I only read it because one of the authors is Nassim Taleb and I liked his books. Anyway, the gist of it is that there are different kinds of risk. Some risks can be calculated with certain probabilities, and you can do cost-benefit analysis with them. For example, you should only spend money on a security system for your home or business if the cost-of-being-attacked times the probability of attack exceeds the cost of the system, right? However, there is another kind of risk for which you can't do that, because their cost is so high that nobody will live to tell the tale and collect the statistics (extinction events), or maybe they will but it will be too late (global financial crisis). The paper argues that monocultures present that kind of risk. You can't calculate "probability of massive blight that kills all our genetically homogeneous food" for example because we don't have much of a sample size. And by the time we do have the sample size, it will be too late. Then, the paper points out some other risks and explains why the risks of artificial selection are different from the risks of GMO. Long story short, it's because a bad mutation in a "natural" crop only kills a few people before they realize that they should plant something else.