The Bush Engine of Technology Innovation

Vannevar Bush argued that it was a proper role for the federal government to support scientific research. This proposition today is regarded as a truth that hardly needs justification. But in Science the Endless Frontier, Bush was not arguing for the value of scientific research in general, but rather a particular area of research and a means to accomplish that research. What Bush argued is so radical that it has largely been rejected, ignored, and converted into a general truism about the importance of research.

President Roosevelt asked Bush whether he could develop a technology innovation engine for civilian use comparable to the one he had created to develop new technology during World War 2 for military use. Bush’s answer was the radical argument in Science the Endless Frontier. In Bush’s approach, there are three elements. The first is an established order–the military (and the companies that supply it), professional medicine (and the companies that supply it), information management (and the companies that supply it), or government (and the companies that supply it). An established order has needs, a leadership mindful of those needs, and resources to address those needs.

The second element in Bush’s engine is what we might call a skunk works. It is a technology innovation engine that stands outside an established order to produce new things that the established order could not specify for itself or even imagine, much less shift resources from pressing needs in an attempt to make. In an established order, people with influence make claims on resources based on where their activities fall in a ranking of the established order’s needs. Outliers rank last, with the nutcases and the clueless. By contrast, a Bush Engine of Technology Innovation (a BETI) is purposely placed outside the reach of an established order, but makes a study of that established order to examine its purposes and practices, as well as its problems–but not just the problems that it sees for itself, but also its problems in seeing what it needs. 

A BETI is distinctive in multiple ways. A BETI stands outside the established order and its politics, so it does not have to follow the “party” line for what is “best.” A BETI draws its personnel from industry (an awareness of manufacturing), from gadgeteers (ability to make new things from scratch and get them to work), and from science (awareness of what is possible in the natural world). The idea behind a BETI is to supply an established order with fundamentally new things–things the order could not do for itself without challenging the premises on which it is led and allocates its resources.

The mix of personnel in a BETI is distinctive–know the infrastructure that’s possible; know how to make new infrastructure as needed; know the science that’s possible. A BETI must be led by someone with a nose for what an establish order needs, but also must be smart about how scientists, gadgeteers, and industrial engineers work together to propose development tasks. In each case, for a BETI to operate, participants must have a deep working knowledge of their area of expertise, but also be ready to adapt that knowledge to a new thing. Scientists have to be open to new science. Industrial engineers have to imagine new production techniques. Gadgeteers have to be prepared to make and fix things they haven’t made or fixed before.

It’s this third element, the science, that is the subject of radical argument in Science the Endless Frontier. Bush sees in new science the key driver of a BETI. Anyone can mine existing science for ideas and then apply those ideas to a problem, starting with the problems that an established order considers the most important ones. One can even start with those most important problems and in trying to solve them, produce new scientific knowledge. If that new scientific knowledge is specific to solving one’s problem, then it appears to be useful and the research is “productive” and “successful.” Otherwise, not so much. If someone demands that every time one finds unproductive new scientific knowledge, that knowledge must be patented and shopped to companies for development, you might see what a terrible distraction that would be to getting an important problem resolved through research and development.

But Bush was looking for something different. What he found was that the driver of the BETI he created for the war effort was new science, which then enabled new thinking about what was possible in the design of devices and methods. Development followed markedly different paths from the approaches adopted by the military–and the military was using what it took to be the best science and engineering available. A BETI, by contrast, does not use the best science and engineering available. It runs on new science–which by its nature is not mainstream, is not best, may not even be accepted; and it runs on new industrial practices, which by their nature are untested, inefficient, outside of best practices. No one in their right minds, as money managers or project managers in an established order, would risk career and resources on anything less than the apparent “best.” “Best practice” is the hedge of the institutional manager. As one told me, “Your ideas may be right, but I could never use them because they aren’t established best practices. If something went wrong, my judgment would be on the line. If I follow best practices and something goes wrong, it’s not my fault.” This is the kind of thinking that a BETI is designed to jump past.

The key idea in innovation is to introduce change into an established order. An established order does deal in change, but it carefully manages that change so that it does not upset or undo the established order. Other than adoption at some entry point into an established order–from the top or by the engineers at the bottom–a BETI couldn’t give a rip about “best practices” or what managers think of their reputations or status in the order. A BETI is designed to render best practices obsolete, to substitute new practices for old, to take advantage of the untested, the unimaginable, the “we’d never have thought of that, and if someone else had, we’d have laughed them out of the room.”

Vannevar Bush argues in Science the Endless Frontier that the key driver, then, for a BETI is new science, not just science in general, and not just the throw-away science that the established order produces for itself in trying to use scientific research to solve its own problems. Bush wants science at the frontier, science that changes what we observe, what we characterize, how we explain, what we can imagine. This is the “frontier” of science–where the unimagined becomes an imaginable unknown, where there is no possible limitation on sorting things into profitable or immediately useful (on the one hand) and goofball or distractingly off-topic (on the other).

How does one choose what to study in frontier science? How does one fund such work? In particular, how does one fund such work using the resources of an established order and yet prevent the established order from turning the exploration at the frontier into something that looks important or immediately useful to the order? These are the questions that Bush has to deal with in making his radical proposal. Here’s his idea:

  1. It’s a proper role for government to fund exploration research, even though there’s nothing to be procured, no pressing government problem to justify the allocation of money.This runs against everything that an established order considers proper: throwing money at projects without a plan, without merit, without a ranking for priority.
  2. The research should be led by the “free play of free intellects”–that is scientists choosing what to explore without being directed and managed by others.And especially not managed by representatives of the established order looking to make sure their money is properly used. “Free play of free intellects” does not mean random goofiness. Nor does it mean that scientists have a love of truth and therefore should be funded for their virtue. The phrase means that an established order is of no use in directing the curiosity of the explorer of the natural world and the mind that would recognize and account for what is there to be discovered.
  3. The research should be located at universities, where scientific talent can be assembled without institutional claims on the direction of the work but institutional resources can be made available to support the work.Bush saw in academic freedom and tenure the tools by which universities as institutions prevented themselves from taking administrative control of the science they supported. While for-profit companies and government laboratories also conducted scientific research, the results were often proprietary or classified, the results that were off-topic were ignored and not pursued, and scientists were recruited and managed and rewarded to stay on topic. That’s the idea behind management–that the managers set the agenda and the hired folks perform the required tasks, even scientific tasks.

The apparatus for a national research foundation in Science the Endless Frontier is then directed at finding a way to allow significant federal money find its way to “free intellects” for work that they propose to advance the frontiers of science. At the heart of this problem is how to choose the people to do the work. This is the problem of money as a scarce asset; this is the dilemma of having money and therefore attracting people eager to please; this is the dilemma of accountability for the use of public money. The proposal is radical: throw money away on low-priority, untested, nutcase “science” in the hope something useful will show up.

But actually, that’s not what Bush was proposing. His proposal was directed at creating the science that talent in a BETI could draw on, not science that the talent in an established order would draw on. It is easy to conflate Bush’s proposal with any generic support for scientific research or especially with research conducted at universities, as if merely being at a university makes a difference. And once one has made such a conflation, then it just makes perfect sense to attempt to sell back to established orders the work that has been done in university research, to become another supplier of science to feed the engines of commerce and government.

Each of these three areas of research practice then may produce “science”:

  • an established order
  • a BETI
  • exploration science

When an established order produces science–whether for commercial purposes or governmental purposes or for the good of the profession–it seeks to appropriate that science for its benefit. What’s useful is promoted; what isn’t is ignored.

When a BETI produces science, it does so while working on things that haven’t been done before, outside an established order. It, too, seeks to appropriate that science for its benefit–but that science is often substantially different science than the science accepted and used by an established order.

Exploration science operates outside the imagination of established orders. It digs where there is no reason to dig. It ranges into the low priority, the side effect, the defect, the odd predictive error, the anomaly. It is easily disregarded or mocked, even among scientists clustered in a consensus. By its nature, exploration science runs outside of scientific consensus, best practices, and even common sense. Bush envisions that the frontiers of science are “endless”–that in exploration of the natural world, we will reach no limit in our minds where there is nothing left to be discovered.

Perhaps that’s not true. Maybe there are limits, both in the natural world and in our minds to recognize, imagine, and explain. But it seems to be the case that we have not reached even “peak scientific discovery” and that there is plenty more to be found, even if not provably “endless.” Stuart Kauffman’s calculation on the number of possible chemical compounds available in the universe suggests that even if we merely create one per second (let alone test it for its properties), we will have way more to do than the number of seconds that have elapsed in the present university. That may not be endless, but it is effectively forever. And that’s just chemistry.

But the established orders that had their sights on research funding from the government were quick to rein in Bush’s proposal. They accepted the idea that there should be government money for research, but they rejected the idea that the money should go just to exploration science or should be managed in a special government agency that was almost but not quite entirely outside of government–just inside enough to be accountable to the President. University administrators rejected the idea. Government agencies rejected the idea. The National Science Foundation was the rejected love child of university administrators and federal agencies–a way to funnel money to university research, so long as more money was funneled through other federal agencies for research, too–and especially the military and medicine.

Even the NSF quickly became another institutional entity, adopting an approach to funding that created the competing project proposal reviewed by panels of scientists–and thus, the idea of a professional class of scientists deciding what constitutes the “best” scientific proposals, of course with an eye for their own standing. It would take a special person to advocate for funding proposals that ran counter to one’s own claims and reputation; it would take an even more special person to advocate for funding proposals outside one’s area of professional competence. But that’s the essence of the problem: how does one choose who to fund, or what to fund, if what’s to be funded is so new that no one has any competence to anticipate and judge it beforehand?

Bush proposed funding people–“free intellects”–not projects, and “free play”–not work pinned down by proposals reviewed for merit in advance and bound into contracts with deliverables. This has proven mostly impossible for established orders to tolerate. It is just this dichotomy that Carl Sagan highlights in Contact, between the goofball researcher Arroway and the cunning career scientist Drumlin–really, the movie isn’t about meeting aliens; it’s about exploration research running up against an established order, and where especially government money should “best” be used to do “good” science.

The established orders also rejected the idea of a BETI. The military figured it could draw on new science like anyone else and would take over. At its best, it created ARPA/DARPA, which in turn produced the internet–and that worked out. The Public Health Service skipped the idea of a BETI entirely and built itself to do just that with the NIH doing both the research and sponsoring the research, and then working to also do the development–and for leukemia and malaria, it did significant work. Elsewhere, though, it ran out of steam (or money) and just funded “early stage” research without any follow-on resources (and thereby creating the impulse for university patenting and “transfer” of results as monopolies to the pharmaceutical industry, which took on something of the BETI role, but in another institutionalized form.

Even professional science rejected the idea of a BETI and limited the role of exploration science. A BETI is predominantly an institution-free development environment, and does not focus on “fundamental” science. The government has organized various research consortia, such as the SEMATECH and the Electrical Power Research Institute, and the NSF has created many more–these do research and may even create “proof of concept” prototypes, but they are consensus organizations operated by their member companies. They do good science, and they even develop new things, but they are also managed by their respective established industrial leaders.

Bush envisioned that new exploration science would be consumed by BETI. The BETI would find that science as imagination struck, because the exploration science would be published, be available (not proprietary), and especially would be outlandishly beyond the prudent tastes of an established order’s priorities and methods. Once professional science settled in to control what constitutes “basic” or “fundamental” research, using merit-based proposals rather than a judgment of individual’s character, training, and potential to discover (all difficult things for a committee to put a number to), basic research itself was largely reined in to serve the pressing needs of professional “science.” Technology transfer was developed as a bridge between this professional sort of research and the companies and government agencies that might benefit from it. Patents were the property of choice, and thus “licensing by bureaucrat” came to appear not only unstrange but apparently virtuous.

As it happens, we have three distinct areas for research.

A. Research that an established order proposes to extend its activity

Almost all sponsored university research is directed at catering to established orders. Research is selected based on “merit” by professional organizations, companies, and government officials. While “basic” in the sense that the proposal for the work comes from the investigator rather than from the sponsor, the research is selected based on what the established order can justify based on priority (it selects the areas for grants) and merit (it selects projects based on consensus science).

B. Research undertaken in competition with an established order’s research

This is now the domain marked out by entrepreneurship. Often entrepreneurial research is outside institutional control. Where it gains power, entrepreneurship is backed generally by private investment capital, not the government. This is the space in which BETI would operate, had Bush’s proposal been accepted in its radical form. As private approaches, the startups that matter–the ones that might become animals (gazelles, unicorns)–are ones that beat up an established commercial order, changing the spreadsheet, changing the customer based, being generally destructive in a creative, capital-intensive way. The need for new science is not particularly strong.

C. Exploration research

This is research that does not arise because someone has an institutional interest or a commercial interest, but rather arises because someone has a remarkable curiosity. One explores because there is something there that attracts exploration. In geographical exploration, one might be motivated by any number of things–claiming land for a monarch, or opening up trade routes for business folk and their investors, or finding gold or other “resources,” or even to “be first” and become “famous.” Francis Bacon argued these were all “idol” motivations, and the reason to do science was “charity”–to contribute to the common good through knowledge. There are folks who think Bacon made this claim as a seductive lie to induce the government to allow research on otherwise taboo subjects, and even to fund it. For Vannevar Bush, the motivation for exploration research is to expand the frontiers of science, to expand the tools available to imagine the workings of the natural world. Perhaps that’s a weak motivation for individuals, who would rather find gold or be a hero or have an adventure. Maybe scientists such as Feynman just say that there’s a “pleasure in finding things out” to cover for these other pleasures, of good pay or recognition or cleaning up with a lucrative patent license.

Most university technology transfer is A research to an established order. Universities may license as well to startups, but often those startups are created to repackage university patents for acquisition by the leading companies in an industry–again, an established order. The cognitive dissonance in the effort is that the A research is sponsored and filtered by the government, while the intended beneficiary of patents on any discoveries made is a company in industry–typically either a leader or a startup that might be sold to a leader.

The research that backs entrepreneurial efforts backed by venture capital is often not institutionally sponsored, and where it has been institutionally sponsored, it is generally the research that has not been deemed to be “strategic.” Thus, Xerox could let go of such oddities as graphical user interfaces, the computer mouse, ethernet, and postscript. Vannevar Bush’s BETI is a form of entrepreneurship, but a form that does not cater to venture capital (itself a rather institutional activity these days). Instead of shopping ideas as the source of “wealth creation” for investors, a BETI aims to use a knowledge of science and industry to fundamentally alter the course of an established order. The results of BETI work might be adopted by an established order, but to do so means altering the direction (and research) that the established order otherwise had committed to.

I once hosted at my university an officer specializing in military innovation. What do you want to see? I asked. If I tell you what I want, then I won’t learn much from what you show me. So we toured five or six projects that I considered “out there,” cool and unexpected. Just what he wanted. Stuff beyond his imagination. We are three years behind in what we think are the areas for research, he concluded. Our projects are obsolete before we finish them. That’s the power of a BETI to grab new science, rapidly build a technology infrastructure, and have it ready for deployment before an established order can bring itself to specify and fund anything even close.

The technology transfer that Vannevar Bush wanted was the flow of new science to BETI. That was, he argued, how to adapt his engine to create new technology to civilian use. For that, Bush needed government funding for exploration science–the “free play of free intellects.” There was still plenty of reason to fund other sorts of research in government labs, and also to recognize that industry labs also produce good science. It’s just that most government lab science is narrowly scoped and may be classified and most commercial science is product focused and often a trade secret. Bush imagined science arising from exploration rather than as merely the offhand discovery of pragmatically tasked “applied” work. As Neal Stephenson has quipped in discussing innovation, “I saw the best minds of my generation writing spam filters.” We can task the best scientists to do the work that has the highest priority to the needs of society (please, rid us of spam by means of technology), and that makes perfect sense, in some weird way. But Bush was radical. He wanted some of the best scientists to do something other than write spam filters, or even try to create a new mouse model for disease therapy research. Yes, in those activities there may well be scientific discovery, and it may even be that the discovery is immediately useful–but most likely not–the discovery will be off topic, random, an anomaly, for later.

Bush wanted to focus on the off topic from the get-go, to expand the frontiers, and then build BETI that would seek out that new science, combine it with insights into the purposes of established orders–military, medicine, industry–and produce new products and methods that transform those established orders. Thus, for Bush, the transfer of “technology” that mattered was from exploration research to BETI, from exploration science to a skunk works with institutional scale funding, with insight into institutional purposes, with insight into industrial methods, but operating independent of institutional direction and controls. Instead, we have transfer of “technology” mostly directed at acquisition by established orders for whatever purposes those orders want–so long is there is a license and at least a mild attempt at using an invention, and payment, that’s sufficient. Where there is exploration science, the “inventive” results of that work, too, are packaged to go down the same channel to whomever is willing to pay–whether a leading company in industry or a startup backed by a venture fund.

Bush didn’t get much in the way of exploration research. Government money was diverted to research in general, research to meet the needs of established orders, research selected by professional science to advance the needs of professional science, leading university administrators to talk about “research as an industry”–the benefits of research are money spent doing research, rather than results of the research being useful–to the public or to industry or to entrepreneurs. But we don’t have government money funding individuals based on what they might do. MacArthur Foundation grants do something in this direction. Howard Hughes Foundation grants also. But not government funds, and not for exploration, and not directed at individuals rather than competed “projects” based on the King Lear premise of “I will divide my kingdom among those who love me best.”

Bush also didn’t get his BETI. There are things that have come close–the great industrial research labs, for instance, and the government-initiated industry consortia. The federal agencies figured they could run the BETI themselves and get to wonderful results more quickly and keep funding under their control. And it may be that we get by with what we have. It may even be that Bush was wrong about BETI, that skunk works don’t work, or maybe that the government cannot bring itself to fund either exploratory research or BETI as Bush envisions, that the apparatus that links an established order with either exploration research or a BETI becomes so complicated, so politicized, so amenable to institutionalization that it quickly takes over exploration and skunk works and institutionalizes these as well. As Matt Ridley has it, the “Moloch state” devours opportunity when it can, creating a “regime of paperwork and harassment, endless paperwork and endless harassment.”

During the development of the Apollo mission to the moon, I’m told that secretaries at NASA Ames had $50,000 in petty cash in their desks. When someone needed to buy a piece of equipment or supplies, they could go get what they wanted the same day. The Moloch state official would be horrified at the lack of controls. Bush, by contrast, argued (in Modern Arms and Free Men) that a government had to operate predominantly based on confidence rather than fear, that if the processes of research end up being adopted to defend the state from misuse of funds, then one heads down the path of fear, and the results will be limited. The way of confidence, argued Bush, is the better way. Perhaps Bush was wrong about that. Perhaps if there’s significant money from the government, there will be fraud and waste and diversion and institutional capture and not much will end up funding exploration science or BETIs or anything actually useful. But what if Bush was okay? What if research can be conducted on the basis of confidence, outside established orders, including professional “science”? In that case, the technology transfer we’d want to see is the movement of new science into BETIs and the production of new technology in BETIs directed at an established order and the adoption of the new technology by the established order, resulting in a change in direction for the better.

What’s missing in Bush’s ideas about BETI perhaps is that of competition to the death. Yes, there was urgency in finding new technology that the military could use but couldn’t imagine or fund. But the urgency had to do with the “free world” and sons and daughters dying in war, not just with urgency for the sake of getting something done efficiently and profitably. Thus, what’s the competitive pressure on a BETI, or more so, on the established order, so that it maintains its status and power, so that it wins out, as it were against forces that would overcome it? Folks tried to find that urgency by adopting the language of war–“an assault on disease.” The idea was that if an established order declared war on its problems, then it could in essence depict those problems as the things that might destroy it, and so must be conquered with urgency. But it’s not the same thing at all to work a metaphor in place of literal reality. The NIH depends on research to exist. Research is chronic. If the results of research consistently created cures, the NIH would have a lot less to do. Research has been institutionalized, selected, managed, controlled.

Similarly, national science was depicted as a race for international supremacy, as if science were an economic weapon that protected a country from the awfulness of having another country’s scientists discover and publish things first, or worse, that another country’s engineers might use our country’s science first, before our own engineers, and sell us a product that we want sooner and a lower price than our companies could do–and thus steal our jobs, our way of life, our international reputation. That, too, could be made to be an urgent matter. Bush thought that a civilian BETI might produce new industries, and jobs, and public welfare, and for that he needed government money, exploration science, a scientifically trained workforce, and access to results. But again, technology nationalism is not particularly compelling. The urgency is not so urgent as it is politically expedient. The Bayh-Dole Act was passed using such rhetoric, as if the 1% of federal patents involving government-funded research was the missing 1% that would make American technology competitive with those of Japan or Europe.

Perhaps if an asteroid was sure to strike the earth in fifteen years, we’d perk up the urgency. Or if everyone was dying from a new disease. Other than that, and war, it would appear there’s not much motivation in the established orders to create BETI or fuel them by exploration science. So BETI will have to remain skunk works, cleverly disguised so that established orders don’t recognize them and made to appear sufficiently unattractive that established orders don’t take them over. That’s one of the problems of “institution building” in an age already heavy on the institutions–one can get too much institution and not enough individual. Bush proposed two mechanisms that might change institutions–exploration research and BETI. These were to operate outside institutions, “free” but accountable. Consider that BETI might be an important element in innovation. Consider as well that you won’t find BETI in any accounts of university “technology transfer”–at least not as best practices, as institutionalized patent licensing. But consider that universities might be the organizations best suited to limiting their own institutional claims on research discovery, so that the results of exploration research might be broadly available to the operation of BETI, and from BETI introduce changes in established orders–that is, to innovate.


This entry was posted in History, Innovation, Policy, Social Science, Sponsored Research, Technology Transfer, Vannever Bush and tagged , , . Bookmark the permalink.