The Imitation Game
Aspiring to be the nest Silicon Valley is not going to work, but imitating DARPA might
Imitation in lieu of imagination
Rishi Sunak, in a speech to the CBI, laid out his vision for the future economy:
“I want to combine our technology and science brilliance with our formidable financial services to turn Britain into the world’s next Silicon Valley.”
This imitative aspiration is becoming a hallowed tradition in British politics. In 1999, Tony Blair urged British companies to ‘get on the net’ to rival the valley. In 2010, David Cameron announced he wanted East London to rival Palo Alto and Santa Clara. In 2018, Matt Hancock argued Britain could achieve similar tech status to the U.S. and China because, to my surprise, ‘we have the ecosystem’. The Johnson premiership also talked up such aspirations but dropped plans for the Oxford-Cambridge arc, which was also pitched as a potential rival to Silicon Valley.
On one level, it is a little sad that our leaders keep making this comparison, even as the gap between the real valley and Britain widens ever further. In 2018, California’s GDP became larger than the UK’s. This is not by government efficacy. California’s energy utilities are a disaster. The state cannot build a relatively basic high-speed rail network. California ranks 43rd in child literacy by state and 50th in adult literacy. Its success is attributable to Silicon Valley, which contributes about 20% of the state’s GDP, and the legacy industries that preceded it, such as aerospace. Having been there many times, it is amazing how much dysfunction and awful politics can be papered over by intellectual property licenses and advertising revenues. It is also alarming how the greatest density of talent on the planet cannot organize to stamp out open-air drug markets.
Sunak has spent some time in the valley, but based on his imitative comment I am not sure he understands what makes it successful. Silicon Valley’s success is defined by relatively small communities of very smart, ambitious, and secretive people building new traditions of knowledge, and with it, new markets. The ideology is built on eschewing competition and imitation and seeking monopoly status through platforms. Copying these novel developments is, by definition, near impossible. I believe the UK can effectively copy the negative aspects of Silicon Valley; the internal conformity, the gradual bureaucratization, the bloated workforces, the mad hype cycle, and diminishing levels of dynamism. But I think if a small country like Britain wanted to be a comparable center of innovation, the markets, technologies, and institutions would have to be novel. Just applying web 2.0 to banking and plastering ‘silicon’ onto every roundabout, alley, and cul-de-sac is an exercise in futility.
We also need to think about expectations. Ministers, MPs, and official documents talk endlessly of Britain being a ‘scientific superpower’. This is unrealistic language. Some might say it’s just feel-good marketing. But like the ‘green’ label, I think the ‘superpower’ label just gives too many stakeholders a way to dodge hard realities. It’s part of a pattern of behavior in our political class that fantasizes more than it implements. Compare to China. From direct subsidies to below-market credit, China spent $348 billion on industrial policy in 2019, or about two years of NHS spending. Yet looking at their documents, it is striking how modest they are in their goals. The Party’s goal for 2021 was to build a moderately well-off society, and its goal for 2049 is to build a modern socialist country. Semantics aside, these objectives are pretty grounded. China actually has built a legitimate counter to Silicon Valley in the form of Shenzhen. But this was an example of the local and national government leveraging Chinese strengths in manufacturing and a large surplus of engineers to capture key industries. China also built its own tech giants, had the population to scale them, and protected them by locking out American platforms. The lesson here might be that if you want your own Silicon Valley, you’ll have to practice protectionism to keep out the current one.
Some might argue that this imitation is another case of British policymakers being too America-brained. It is as if we cannot envisage changing our situation outside of becoming a variation of the center. This is true, but also understandable. We are so intertwined with the U.S. in foreign and economic policy that it would be hard for us to think otherwise. When Arm, the UK’s most prestigious company, is angling to go public on the NASDAQ rather than the London Stock Exchange, you might question the think pieces that argue Britain is distinct from the U.S.
A reserved cheer for ARIA
While imitation of the U.S. is by and large unproductive for our circumstances, the creation of the Advanced Research Invention Agency (ARIA), a British DARPA, is perhaps an exception. I won’t go into DARPA much. Benjamin Reinhardt has done a very thorough analysis and my fine employer Bismarck Analysis has analyzed its current travails.
Britain is far from unique in wanting a DARPA. Thanks to Mazzucato-mania, plenty of countries have looked to build their DARPA clones. Germany and Japan have developed their own DARPA-esque mission-led organizations. Russia alluded to developing one back in 2014. Even private companies like Google and charities like the Wellcome Trust have built their DARPAs with the help of former DARPA director Regina Dugan. The U.S. government has even built mini-DARPAs. IARPA is centered around intelligence and works pretty well due to its singular customer being the intelligence services. ARPA-Energy is focused primarily on funding renewable energy technologies.
DARPA and imitators
Most of these imitations, in our view, are not likely to deliver the results the original DARPA made. In the case of Russia, I think DARPA just means super secret R&D. In the case of Germany and Japan, the autonomy of the organizations is compromised by being tied to fashionable issues like aging societies, climate, and so forth. Germany’s Sprin-D is purely focused on non-military aspects, while a separate organization (Agency for Innovation in Cybersecurity GmbH ) does research on cybersecurity. Why is this a problem? Well, for the same reason DARPA has slowly lost some of its gloss from the glory days of 1958 to 1973 when its most startling innovations were made. The more definitive and politicized the organization’s mission becomes, the more restricted its research. In the mid-70s, the original ARPA got the ‘D’ for defense and had to focus more specifically on military technology. In turn, it became far more subordinate to what the U.S. defense establishment was interested in. So while DARPA retained a lot of the good things, like high autonomy for program managers and tolerance for risk, it became tethered to any pathologies or failures of thinking that permeated the U.S. top brass. Consider the Future Combat System debacle for more reading.
The original ARPA was quite a weird beast and owed its existence to glitches in the system. It was hated by all the armed forces when proposed in 1957. It got approved for three reasons. Sputnik created momentary panic in the system. The secretary of Defence at the time was not a military man but a former Proctor & Gamble executive. Lastly, President Eisenhower, the quintessential military bureaucrat, was fed up with the military establishment’s obstructive nature and liked the idea because they hated it. In ARPA’s first 15 years, the U.S. got the progenitor to the internet, the computer mouse, the M16 rifle, funding for the Saturn V rocket, and materials science. While DARPA has continued to produce a lot of impressive stuff since then, that particular period where the organization was at its most autonomous, informal, and unrestrained was quite unique.
It is from this period the people behind ARIA are drawing inspiration. It seems to have a very broad remit, is not tied to the military like DARPA, but is not blocked from military research like Germany’s Sprin-D. It does not seem to have the mission-centric baggage of the German and Japanese derivatives. The hiring of Matt Clifford as chairman is positive, as I know my boss has impeccable taste in company.
There are some concerns. It is located within BEIS and so is quite dependent on whichever individual is heading that department. BEIS itself is easily pressured by the overmighty Treasury. As with DARPA, the autonomy of program managers is quite dependent on the organization director’s determination to resist pressure from government departments. At an initial glance, ARIA looks like the most sincere non-U.S. attempt to replicate the original ARPA magic. But after a few decades, U.S. lawmakers and bureaucrats got quite tired of the informality and autonomy of ARPA managers. ARPA relied on high trust and was not performance-driven. Such an organization is always likely to attract scrutiny from larger bureaucracies. I think that the same tendency will manifest quickly in the UK, especially given the tightening budget constraints on government investment.
ARIA is clearly an imitation. The UK, much like Germany and Japan, is staring down the barrel of long-term stagnation and is throwing ‘hail marys’. to improve productivity. Countries in this predicament have unoriginally latched on to having a DARPA because it sounds dynamic and experimental. But more than the others, ARIA seems a sincere attempt to inculcate what made the original organization successful, rather than just build another organization with an innovation focus. The big question is whether 1950s American tolerance for mad scientists will be compatible with modern preoccupations with accountability and results-driven metrics. The best thing is to build enough credibility in technology circles to cause a stink if money is taken away, but innocuous enough not to elicit unwanted scrutiny.