SOME of humanity’s finest acts of collaboration are carried out by people in special suits – the full-body kind, which protect them from radiation, lethal pathogens, extreme temperatures and aggressive substances. But then there’s the reverse, where the suit helps to protect delicate objects from human contamination.

The clean room is the par excellence example. Decked out with powerful fans and filters to remove airborne contaminants, clean rooms are some of the most controlled environments we have. For the past half-century, collaboration in these spaces has fuelled progress, facilitating the best of semiconductor research and development – and as a result, profoundly influencing US science policy and practice.

These labs were not only a locale of experimentation, but also the products of “institutional experiments”, as science and technology historian Cyrus Mody explores in The Long Arm of Moore’s Law. The institutions are “social technologies”, distinct from but co-evolving with “physical technologies”. Based on archives and interviews, as well as his time spent researching alongside nanotechnologists at Rice University in Texas, Mody charts the highs and lows of developing integrated circuits, or microchips.

In meticulous case studies, Mody traces the unprecedented interdisciplinary collaboration between universities, manufacturers and government funding agencies during and after the cold war. The space and arms race, which had spawned civilian microelectronics in the 1960s, was decelerating and, in its place, global competition to build better chips was picking up – fast.

“Interpretations of Moore’s law range from a law of nature to an intentional coordinating mechanism”

Research groups formerly financed by the military looked to civilian agencies such as the National Science Foundation or the Environmental Protection Agency. In turn, the agencies began developing their own applied research programmes. Giants like Intel and IBM increasingly shared the huge financial burden of research with universities and national laboratories.

As the development of innovative microchips became ever more complex and costly, other kinds of large-scale hybrid partnerships were forged. These brought together industrial, government and academic players in consortia, shared R&D centres and eventually meta-networks of organisations.

At the heart of all this was Moore’s law, the observation-cum-dictum that the density of microchips doubled at intervals: initially 12 months, then 18, now close to 24 months. The “law” took its name from a 1965 paper by Gordon E. Moore, a co-founder of the mighty Intel.

His law came to predict the exponential miniaturisation of integrated circuits, and hence the increase in computing power that could economically fit on a single chip. Since then, there have been many interpretations of Moore’s law. These range from seeing it as a sort of natural law (due to the physical behaviour of materials such as silicon) to a consciously coordinated mechanism – an engine for innovation that guides rather than predicts human activity.

Mody fully acknowledges the technological aspect, but approaches the law as a “social fact”. He thinks of it as a human construct continually “enacted” by the likes of grant officers and programme managers applying it in journals and conferences, or through labs, national working groups and industry roadmaps. Fuelled by the needs of big science and government for ever greater computing power, and the desire for ever smaller consumer goods, this rule of thumb becomes a self-fulfilling prophecy we maintain and are subject to.

Mody’s book concentrates on the mesoscale of organisations involved in developing microelectronics because he argues it helps us get a sense of how the semiconductor sector influenced scientific knowledge-making. He shifts fluidly between detail and contextual currents, capturing the coalescence of individuals and institutions into configurations. These groupings splinter, disband and regroup after massive or minute changes, like IBM’s foray into circuits involving superconducting materials, the drying up of funding sources or the exit of a key team leader.

From that vantage point, Mody zooms into the nanoscale of materials and zooms out to the macro-scale of global economics and national security. While the two scales were increasingly linked by policy-makers, in fact the civilianisation of computing gained traction. Chip technology always had dual-use potential: microchips could be embedded just as easily in a reconnaissance satellite as in a medical device.

Mody’s book offers a wide range of important issues providing food for thought on the R&D behind our modern systems. Amid the different models of innovation, and approaches to technology management, civilianisation was as pertinent to the advent of semiconductors as it is today.

To illustrate this, Mody evokes student and faculty disquiet at the relationship between the military, industry and academia at Stanford University, California, in the late 1960s. Administrators there had to communicate the dual potential of microchips, emphasising military or civilian advantages depending on who they talked to. This reassured security customers while diffusing criticism against the military’s role, and winning over researchers who would not otherwise have considered a career in defence systems.

It is a compelling thought, highlighting not only the application of a technology after its conception, or a debate that may occur while it is being conceived, but also its appeal to researchers and designers before they engage in its development.

Although the emphasis for microchip applications shifted significantly to civilian markets, defence and security players such as the Defense Advanced Research Projects Agency (DARPA) played a crucial part in addressing “network failures” in the growing semiconductor sector : the lack of trust, transparency and know-how exchange between different kinds of organisations.

The trick was to link stakeholders across time, distance and institution, to help them get to know each other and team up. The transparency and capability built by the networking efforts of government agencies (security, defence or civilian) was to prove vital in advancing microelectronics in the US.

Semiconductor technology today continues to depend heavily on R&D. If Moore’s law helps drive technological foresight, it may also help us to develop responsible innovation and to figure out what we should do, rather than just what we may soon be capable of doing. We need to know if we can cope with the implications of the potential technologies the law churns out – not just in our daily lives, but also with the impact this has on science and engineering sectors.

“Increasing computing power is one thing, deciding in which devices to embed it is another”

The Long Arm invites us to reflect on the technology we conceive, discover, develop, manufacture, consume and discard. We have to ask ourselves which technologies we regard as critical to national or global security. Who should develop these systems – corporate labs? Government? A partnership of both? How do we want to set up processes that devise our future technology? Are we looking at the right problems?

The journey doesn’t end at the next milestone of doubled chip capacity. Increasing computing power is one thing, the small matter of what to do with all the extra capability is another. We must decide in which devices to embed it or what systems to build around it.

Last year there was a lot of publicity about around Moore’s Law being “dead”. Mody sees such claims as not being strictly true or false, but as conditional prophecies that bring about changes to prevent the prophecy from coming true. Moore’s law may be pervasive, but it is also plastic. And since we enact it, the law is literally in human hands.

Ěýby Cyrus C. M. Mody, MIT Press

This article appeared in print under the headline “The power of Moore”

“Many of our systems were designed incrementally: they are insoluble puzzles that are always changing“

Will we become the masters or the servants of our technologies? (Image: Vincent Fournier/Gallerystock)

AUTONOMOUS weapons, self-driving vehicles, robotic manufacturing plants. If these technologies are transforming our future, wouldn’t it be wise to know what makes their designers tick, and what roles they envision for their human operators?

What these roles turn out to be, argues New York Times science writer John Markoff, involves not a technical but an ethical choice. In his new book Machines of Loving Grace, he X-rays the artificial intelligence and human-computer interaction communities to discover their respective value systems.

With remarkable access to the archives, test sites and conference rooms ranging from the Massachusetts Institute of Technology to Silicon Valley, Markoff probes the philosophical fault line that runs through university and commercial labs, and has divided organisations since the 1950s.

It turns out that the two camps roughly subscribe to either automating humans away, or assisting them by augmenting their capabilities. The deep rift is epitomised by the debate on crewed versus robotic space exploration: even if you decide in favour of an astronaut, should she or he be an active supervisor or a passive passenger?

Markoff also documents how robotics developers have surfed the wave of military spending, most notably by the US Defense Advanced Research Projects Agency (DARPA). Its dual-use inventions find civilian and military purposes – think autonomous vacuum cleaners versus their “milspec” cousins sniffing out roadside bombs.

And it is defence where Markoff sees the struggle as most evident: trading the cost-effectiveness and precision of autonomous systems with the “consequences of approaching the line where humans are no longer in control in decisions on life and death”. Will we become the servants of the technologies we create?

Markoff’s concluding question is the starting point for Wendell Wallach, an ethicist and scholar at Yale University. The title of his book, A Dangerous Master, warns what will happen if we fail to establish formal measures to anticipate and address the implications of technologies during the design process.

Broadening the scope, Wallach includes cybersecurity, 3D printing, nuclear energy, bio, nano and geo-engineering, surveillance, and robotic surgery.

Each of these technologies has a big impact on society, for good or ill. Wallach picks out warfare as one of the drivers of what he calls a “techstorm”, the torrent of innovation so fast, complex and untransparent that societies are in peril of losing – or rather ceding – control, of being swept away into a future that is unpleasant at best.

Take one technology being developed: lethal autonomous weapons systems, LAWS, that are “capable of initiating actions with little or no human involvement”, says Wallach. Here he is not worried about remotely operated drones, but rather about combat aircraft, such as Northrop Grumman’s X-47B prototype, that are capable of autonomous take-off and landing.

Such new weapons, he argues, should be debated by scholars, designers, policymakers and an informed public. Today’s arms control, however, is not an ideal blueprint for regulation. Instead, Wallach recommends setting up “governance coordination committees” to monitor technologies and liaise between stakeholders.

Such foresight, he stresses, must be undertaken at “inflection points”, windows of opportunity before a technology is deployed, to allow us to think through scenarios for its use (and abuse), and intervene if necessary. This need not endanger development of a technology since short-term drawbacks may be outweighed by benefits in the long run, and both must be anticipated.

Wallach and Markoff deliver sobering assessments of today’s engineering culture. Augment or automate? Seek-and-destroy robot or search-and-rescue rover? Far from being black and white, this choice is “a design decision that will be made by individual human designers,” Markoff says.

Thinking stuff through sounds like a no-brainer, but apparently it’s not, as Wallach underscores: “Many scientists and engineers do not believe that the ethical and policy challenges arising from their work is their problem.”

Neither alarmist nor affirmative, both books contain urgent, compelling and relevant calls to consciously embed our values in the systems we design, and to critically engage with our choices.

“Both books contain urgent calls to consciously embed our values in the systems we design”

Unless we design ourselves out, humans are part of any technical system that we commission, develop, use and hack. Don’t fancy a future of self-inflicted overcomplexity and unpredictability? Before welcoming our robotic overlords, read these books.

This article appeared in print under the headline “Gods of tomorrow?”