The Banana Peel Project

Charles J. Murray’s The Supermen: The Story of Seymour Cray and the Technical Wizards behind the Supercomputer invokes all of the images and paradigmatic power struggles we’ve become familiar with in the world of high-tech business. There’s Seymour Cray, the reclusive genius whose tooth-and-nail battle against bureaucracy made him the personal hero of computer technicians and electrical engineers everywhere. There’s the endless struggle between creativity and control, and the sacrifices that are inevitably made when art is institutionalized. And then there are the questions that any good history of science raises: Why did some ideas fail, and others succeed? Are genius, art, and revolution concepts that are rooted in particular times and particular spaces, or are they universal? Despite its many flaws, The Supermen helps to frame those questions by painting a nuanced picture of the development of the modern supercomputer.

The idea that the military was single-handedly responsible for establishing the first computer research companies shouldn’t be foreign to you, especially if you grew up around Silicon Valley. In 1946, the United States Navy sponsored the creation of Engineering Research Associates (ERA) for the purpose of developing code-breaking technology. The new company was privately owned, but it subcontracted exclusively to the Navy and was initially staffed only by engineers who had previously worked on another secret Navy research project (CSAW) during the war. It was only later (and only with the permission of the Navy) that ERA and its later incarnations would open up its business to non-military customers.

Again, this story isn’t new. But it stands in stark contrast to the book’s portrayal of the supercomputer as something that sprung up from a combination of non-hierarchical, creative collaboration and spontaneous revelation. Over and over, we are shown the prodigious genius Seymour Cray toiling away in solitude over his workbench, sweating out Boolean logic on paper before having an epiphany that provided the missing link in his current project. The most valiant efforts at research and development in the story take place in project spin-offs taking place far from the centralized bureaucracies of the computer industry: in isolated cabins in the woods, in Cray’s souped-up home basement, in dusty warehouses staffed by sweaty engineers in their undershirts.

It’s a testament to Murray’s writing that he acknowledges the tension between bureaucracy and creativity. When writers about science, business, and technology do research, it can be tremendously tempting to glorify subjects, oversimplify narratives, and generally become historical determinists. Murray succumbs to the first two errors, as I’ll explain in a moment, but he is far from a determinist. The narrative tone of The Supermen shifts over time, from one that hails Cray’s secluded, free-for-all research lab as “an engineer’s paradise” to one that waxes poetic for the quaint, anachronistic spirit of independent creativity that inundated it. In the early nineties:

It was no longer like those do-it-yourself, pioneer days when engineers had built their own test benches, swept the floors, soldered the boards, and sent the state of the art soaring beyond everyone’s imagination…[It] wasn’t enough anymore…Gone were the days when an engineer could isolate his crew among the pine trees and build revolutionary machines without corporate intervention. It was a team game now. There would be no more engineer’s paradise.

Just as the market was dynamic, and just as the needs and resources of their customers changed after the Cold War, so too does the tone of The Supermen change as the narrative unfolds. This is good science writing.

Yet the story is not without its problems. The protagonist is described as nothing short of godlike. He was genius, pure and simple. Throughout the story, Cray is infallible. Yet from the beginning, as Murray is eager to point out, Cray compulsively avoided bureaucracy. He made it a habit to start new ventures and build them into profitable corporate machines and then bailing out as soon as responsibility beckoned. He ritually surrounded himself with brilliant salesmen and managers; on very few occasions does Murray acknowledge the role of people like John Rollwagen and Bill Norris in the establishment and consolidation of the supercomputer industry. Cray wasn’t a businessman, he was an inventor. He relied on businessmen to market and manage his creations, and without them he would have been a garage tinkerer. Murray makes a paltry effort to highlight this.

All of this is in stark contrast with Murray’s portrayal of Steve Chen, another “by most accounts” brilliant engineer hired by Cray Research (one of Cray’s abandoned corporate monsters). Chen, “a slim, soft-spoken Chinese immigrant,” “struggled with the language and culture” of American industry and “felt he understood parallel processing.” Murray portrays Chen as an attention hog and an overconfident scientist who stupidly took it upon himself to start his own creative lab in the woods, Supercomputer Systems, Inc. Little attention is paid to Chen’s own work and his eventual $150 million deal with IBM.

Ultimately, the real value of The Supermen lies not in geniuses or engineering utopias, but in the questions that it raises about the differences between evolutionary and revolutionary technologies. When computers were new, they had a “wow” factor. Governments, universities, and companies would pay top dollar for the fastest, the smallest, the first of anything. As the technology advanced and complexified, and as the computer manufacturing industry stratified and bureaucratized, customers developed more specific demands for more specific uses: software, maintenance, size, storage, etc. Murray’s narrative eloquently describes this process, but offers no suggestions as to why it happened. What is it about new technologies that forces them to follow such a curve?

The concept of a “maturing” technology is certainly an interesting one, but what does it mean? Are new technologies more exciting simply because they are new, or because they allow us to do something that we could never do before? What is the role of the market in the process by which a technology moves from revolutionary, to evolutionary, to obsolete? Does it have to do with market saturation and production costs, or is it something more intrinsic to the technology itself?

I don’t think that there are multiple answers to these questions. I think that there’s one, but I’m not yet sure what it is. In any case, cheers to Murray for producing a technologically sophisticated business book that is thoroughly engaging and overwhelmingly intelligent.

My next post will be a fun announcement.

[the banana peel project]