The end of insight?

A post was made on Gene Expression this past Tuesday called "The End of Insight-Monkeys lost in their own castles". It discusses the idea that many scientific explanations may soon be past the ability of the human mind to comprehend. One example is the so-called cellular automata studied by Stephen Wolfram. I happen to have read part of Wolfram’s book A New Kind of Science about 3 years ago, so I know the general idea being referred to here.

One important concept is that the simplified equations used to model phenomena in physics and chemistry do not adequately reproduce the complexity observed in nature, particularly in biology, such as the arrangement of spots observed on many animals, but the iterated application of a set of simple rules can. A far more general concept, however, and one far more limiting for human understanding, is the hypothesis that certain systems perform computations that are irreducible-their behavior cannot be predicted by a clever shortcut or intuition, but rather require building a replica of the system, either physically or within a computer, and observing it run.

Fly, who posted on Gene Expression, makes the claim that the limitations of insight “seems likely to afflict us in science.” I In some ways it seems that this shift is already taking place, to my great disappointment. My interest has always lied in mental pattern-finding and trying to predict the function of, or design, complex systems based on the properties of individual components and a mental sense of how things work. This is particularly true when the complexity in question is of a visual or spatial nature.

One of the fields I have been interested in for a while is how protein structure influences function, particularly the binding of small molecules such as metabolites or drugs and/or signaling. It seems more and more that the general consensus is that it is not practical to have humans try to predict these things using their own brains, but to use automated computer search algorithms to test hundreds of thousands of possibilities. This has chased me away from the field of protein structure as a career goal. I am seriously beginning to wonder if there is any place for me in biology, given that I like to predict and design things mostly using my “bare mind” and find using computer simulations to do most of the predicting and explaining boring. It seems in a way that chemistry is running into a similar situation—making any revolutionary or novel prediction tends to only succeed using quantitative number-crunching rather than qualitative mental models, and sometimes even these fail when judged against experiment. I have asked many scientists whether there is a niche for predictions based on mental models or design by human creativity in any field they know of, and very few can give me a straight answer. It may be that the usefulness of talented human pattern-matchers has declined in favor of systematic computer simulation of the type Wolfram describes. In any case, even if all other scientists think a phenomenon is far too complex to be understood by humans, that will not stop me from trying to explain it using my own insight.