I've been research director on some communications projects focused on sustainability -- trying to get the public to understand the concept. It's easy to get fooled into believing that people understand what you're talking about, only to discover later that their perceptions were quite different from what you thought. There are certain parts of the sustainability story that are easy to tell -- resource depletion or economic viability, for example -- and then there are parts of the story that are nearly impossible to tell. One of these is the relationship between diversity and resilience (especially when it comes to sub-optimal, even maladaptive, traits and forms).  Across the long term a diversity of imperfections makes whatever system more resilient and ready for changes.   

Evolutionary biologists have the best grasp on it, since non-optimal forms are the stuff that evolutionary change is built of.  People who take the long view, ecologists like Wendell Berry or meta-farmers like Sharon Astyk, understand that this applies to our food system as well.  At the moment our tendency toward monoculture, genetically narrow breeds, and our one-size-fits-all approaches make us vulnerable to any kind of change (whether it’s a new pest, climate weirding or the declining availability of cheap fossil fuel).  Things always change eventually, and when change comes we'll want more rather than fewer varieties to choose from.

But Americans tend to believe there is always a "best" variety, and it is very, very hard for most people to really see the value in having a system full of waste and noise and inefficiencies (that is, diversity).  

When it comes to farming, there are two levels here. One is the farmer who has specific desires (like poplar that is straight for lumber, or fast-growing for landscaping) and she's certainly right to breed toward that end, even though it means reducing the genetic diversity of her tree stock. Across the short term, as long as conditions (whether the weather, the pests, or the economics) don't change too much it's fine. But I'm talking about non-optimal varieties being crucial for the long-term.

As Berry and Astyk point out, we can have the best of both worlds by breeding for (rather than just bulldozing over) a wide variety of local conditions. 

This can generate diversity without sacrificing much in the way of actual productivity.  A stringy, drought resistant sheep makes sense over in X-ville, where the heavy-wooled lummox that's favored over in Y-town wouldn't last a week. The end result is a much more diverse system that is better prepared for whatever changes the world throws at it.  It just takes a different kind of attention.

One analogy might be that it comes down to the same juxtaposition as between totalitarianism and pluralism.  Ideal totalitarianism assumed that humans where intelligent enough and disciplined to design and execute a Plan, and that Plan would encompass the future, realize the best possible outcome and exclude all others.  And that’s pretty much the model of agribusiness with its hybrids and its GMOs.  Ideal capitalism assumed that giving lots of people enough room to succeed or fail would mean that the next best thing emerges from that muddle, and so innovation and progress is ensured.

  

In nature, species, like farmers, respond to a great deal of pressure to optimize and specialize and limit their options -- and that is obviously a good and probably necessary short term survival strategy. But that's one reason why most species go extinct eventually. It's that wider, more diverse pool of options outside that species that makes the biosphere so durable and robust (in contrast to individual species which flicker through the fossil record like mayflies in comparison).   

When it comes to our more modest challenge of surviving the next couple of centuries, heading into global climate instability with a system built around extreme monoculture and cheap diesel is pretty much suicidal.