Epilogue

Centaur Science—The Future of the Hypothesis

I began this book when I started thinking seriously about the scientific hypo­thesis in the context of graduate school science education; I end by thinking about how science education affects public perceptions of global climate change and how the hypothesis will fit into the science of the future.

For nearly 400 years, the Scientific Method, anchored by the hypothesis, has guided science. I've said that one of the most important concepts in the phi­losophy of science is the hypothesis, right up there with empiricism and fal- libilism, and I quoted the physicist, Herman Bondi, “There is nothing more to science than its method, and there is nothing more to its method than Karl Popper has said.” In retrospect, Bondi's remark seems a bit overstated; we do, after all, have to allow room for asking questions, making observations, Discovery Science, and Big Data. And yet, as I've tried to show, there are links among these other approaches to studying nature and to the hypothesis. They are all part of the method of science, and the purpose of that method is to ex­plain and understand.

Despite the almost incredible successes of human-powered, hypothesis­based science, everything may be changing. With the development of faster supercomputers (the current record is 200 x 10¹⁵ operations per second,¹ and quantum computing offers the prospect of significantly higher speeds); the avail­ability of limitless, cheap information storage in the “cloud”; and the growth of increasingly more “intelligent” computer programs, the advantages that the human mind has for conducting science are dwindling.

We seem to be approaching a crossroads, with the Big Data Mindset and the Robot Scientists heading off in different directions. The Big Data Mindset promises limitless ability to explore correlations beyond our imagination but forces us deny our nature and forgo the drive for understanding that has gotten us to where we are.

Garry Karparov's ego suffered a blow when he lost to the IBM Computer Deep Blue in the chess championship, but Kasparov did not become one of the greatest chess champions in history by lacking fortitude or imagination. He began working with computers, encouraging the formation of a man-machine hybrid team called “Centaur,” after the half-man, half horse beast of Greek mythology. He created “Centaur Chess” and organized a tournament open to all—human grand masters, supercomputers running chess artificial intelligence (AI) (more powerful than Deep Blue), and Centaurs.

And who won? If you guessed a Centaur you'd be right, but not the one you'd probably imagined. The Centaur that won, beating out all of the solo humans and solo supercomputers, was not the human grand master teamed with a powerful AI chess program, but a pair of amateur chess players teamed with three weak AI programs. In Nicky Case's account,² the dominant combination was not ge­nius or speed, alone or together: it was the symbiotic marriage of the best of the human mental traits with the strengths of AI. The guiding principle is that, in order to maximize intelligence, each intelligent entity must specialize in what it's best at.

As Case explains, “AIs are best at choosing answers; humans are best at choosing questions.” Their respective strengths are coming to the fore in the resurrected field of intelligence augmentation (IA), which is the flip side of AI.

The future ofbasic science research can be like that: instead of Robot Scientists, we'll have Centaur Scientists. And the human side of the Centaur Scientist will specialize in inventing hypotheses. Because that's what we do best.

Notes

1. World's fastest Supercomputer (as of June 2018) https://en.wikipedia.org/wiki/ Summit_(supercomputer)

2. Nicky Case, “How to Become a Centaur.” https://jods.mitpress.mit.edu/pub/ issue3-case?version=c847d892-97dc-40a7-a412-315d255b9b2d 1/