In fact she's discussing how a particular form of science--reductionist science--has turned out to be too primitive to cope adequately with complex organic systems such as nutrition and agricultural. The same is true in education, where factory-style schooling boosts test scores faster in the short run (just as Round-up ready crops yielded faster yields at first) but yield poorer overall results in the long run, and cause significant collateral damage (the same pattern seen with GMOs).
Whether its agriculture or nutrition or education, reductionist science tries to reduce a complex system to a few variables, and in doing so loses sight of the big picture that one must understand in order to grasp any complex endeavor. In nutrition, we tried to reduce eating to a list of vitamins and minerals, plus certain percentages of fat, protein, and carbohydrates, and then created artificial fortified foods that had those things--and we got very large and very sick. It turns out that the way you get those nutrients matters enormously, because natural, plant-based foods versus processed foods set in motion very different dynamics in the body, with one path leading to greater health, and the other leading to greater disease. In agricultural, we tried to reduce it all to crop yields, but learned that the way you grow the crops matters enormously--since local-organic vs. factory farming producing fairly similar crop yield but actually have dramatically different impacts on human health, ecosystem health, and even local communities and the strength of democracy. In education, we reduced it all to test scores, but here again, different paths to achieving similar test scores have dramatically different effects for student motivation, students' emotions and mental health, cross-curricular learning, and outcomes such as creativity, initiative, and respect for diversity. Healthy nutrition can't be reduced to counting calories or carbs or vitamins, healthy agriculture can't be reduced to greater short-term crop yields, and healthy education can't be reduced to greater short-term test score gains. The history of each field indicates that reductionist science has taken us down a counterproductive dead end, with seductive and narrow short-term benefits being followed by a broad pattern of collateral damage, and more troubling still, an increasing dependence on the unhealthy foods, unhealthy agriculture methods, and unhealthy teaching methods. Colin Campbell's book Whole captures this dynamic for nutrition, and I'm working on a book that examines this unhealthy dynamic (and the alternatives) for education. In each field, trusting reductionist science too much led to identifying as effective methods that achieve those short term gains in a way that undermines the capacity of the body, plant, or learner to achieve healthy gains in the future.
To be sure, a great deal of traditional, reductionist science backs up the points that Shiva is making--Indian farmers who only had to spray their cotton crops twice in year one were spraying them 20 times in year seven, and all for a crop that produced inferior cotton, produced in a way that destroyed local communities and the ecosystem, that promoted monocultures, that bred superweeds, and in which many Indian children and adults were poisoned and became sick and even died. The tipping point seems to be about year four--after that even the economics of the GMO approach goes into the red (not even factoring in health to humans or the planet. The same phenomena occurred with Round-up ready soy in South America, and other GMOs elsewhere.
Reductionist science can identify such specific problems, but because it is poorly suited to systems thinking, it tends to lead us to treating isolated symptoms of a problem rather than looking for systemic prevention and cure--a path that often requires an entirely different paradigm.