Breaking the suspended animation of defence innovation
SpaceX launch

SpaceX succeeds by learning fast. 'Failures' still produce progress.

Written by

Lesley Seebeck

As policy advisors and decision makers consider the problem of defence innovation, the video of the SpaceX Starship launch of 20 April 2023 should be required viewing.

When they watch it, they should note the audacity of the vision, the engineering, and the global audience. 

They should reflect on what is now almost commonplace, that a private company, little over 20 years old, is undertaking what used to be firmly within the realm of national governments. 

But most of all, they should pause and replay that moment when the rocket ship explodes—or in the words of SpaceX, experiences a ‘rapid unscheduled disassembly’—and cheers erupt. 

It is hard to imagine that same spirit—the celebration of an apparent failure in front of the world—enlivening the Australian Defence Department. Timidity is evident in Defence’s reluctance to release the Peever Report on defence innovation. That does not smack of an organisation—whether the department or its political masters—willing to embrace transparency, accountability and grasp unpleasant realities.  

Defence hesitancy to trust externals—especially new, unfamiliar players as opposed to existing defence primes—will be reinforced by the triumvirate (Head of the acquisition organisation (CASG), Vice Chief of the Defence Force and the Chief Defence Scientist) governing the Advanced Strategic Capabilities Accelerator (ASCA). Those arrangements are indicative of Defence’s determination to retain funding within its grasp and to balance competing, and typically short-term, interests internally. 

Then there’s the prospect of failure. And there’s the rub. True innovation is a paradoxical beast: as Elon Musk and SpaceX know, to succeed, one must be prepared to fail, time and again, and critically, be ready and able to learn from those failures. 

Given that, the 18-month planning period for the ASCA more likely reflects bureaucratic process, with what passes for public consultation in the modern APS, and careful risk management than open exploration, testing of ideas, and feedback and improvement loops. 

I’m hoping to be surprised but chances are that the outcome will be a standard prescription, some bundling together of programs in skills, co-investment in priority technologies, research grants for upstream emerging technologies, some prototyping and opportunities for testing/sandboxing, some infrastructure, opportunities for collaboration with AUKUS partners, outsourcing to investors, and support for export promotion. 

In short, nothing radical, nothing that hasn’t been tried before. Bureaucratically safe. A program that will meet the worthy prescripts of standard economic development. One that’s respectable enough to be tied into the 2024 National Defence Strategy, complete with a glossy ready to be placed on a shelf and forgotten.

As measured against the SpaceX standard of innovation, of US skunkworks, industrial labs and moonshot companies, of Silicon Valley lean startups, let alone the needs identified the government itself in a world of little warning time and of intensifying technological competition, the outlook is already one of, at best, stasis—and as such, almost certain failure.

The problem here is at heart is one of misdiagnosis. The government already thinks it knows the solution. That’s why it keeps enumerating lists of technologies, of kit and capabilities. 

But worthwhile military innovation is hard.  That’s recognised by the US Defense Advanced Research Projects Agency (DARPA), everyone’s favourite defence innovation agency.  DARPA, as reported in Richard Rumelt in Good Strategy, Bad Strategy, focusses on niche, ‘DARPA-hard’ military problems, precisely because they have no easy or obvious technical solutions or, alternatively, emerging technologies that may have unclear, far distant, and far-reaching military consequences. 

Writing in Modern Arms and Free Men in 1950, Vannevar Bush observed that while people, especially planners, can grasp the value of a device before them, they struggle to imagine intelligently devices for the future.  Novelty, as Clayton Christiansen’s research found, is rarely initially welcomed. 

Asking Defence to imagine, cultivate, nurture, and realise the sort of innovation that may, to be successful, necessitate disrupting the current basis for Defence culture, doctrine, and ways of working is expecting too much of Defence.    

The problem the government must itself grapple with the harder task of establishing the appropriate means of creation of new technologies and their rapid testing and production, and then their translation into military capability.  These are two very different functions, entailing different systems, business models and cultures. 

All of which suggests a new institution. That won’t be easy—building good institutions requires insight, goodwill, integrity, and tenacity. The Australian government has at best a mixed record.

It’s worth returning to Vannevar Bush. Bush wanted to overcome the divisions he’d experienced between civilian scientists and military during World War I. He was keenly aware that science and technology was going to be a key determinant of the outcome of the war in the 1940s and that a civilian-military partnership was essential.

Drawing from Bush’s experience, the ASCA must meet the same two fundamental criteria Bush identified as critical to the success of the National Defence Research Committee and later the Office of Scientific Research and Development (OSRD). 

First, it must have top-level, prime ministerial, authority, separate from the military.  And not simply for show: the prime minister must buy into the vision. After all, if science and technology is so critical to Australia’s strategic interests, it should be treated as such. Miring the ASCA in Defence red tape and subordinating it to military interests will not achieve the free thinking and competitive spirit needed.

Second, it must have an independent budget and sufficient, sustained, long-term funding. This, along with the authority given by accountability to the head of government is key to its autonomy. It would also allow the ASCA to fully fund research, including overheads, rather than nickel-and-diming academics or start-ups, helping build Australian capacity.

There are some further principles of operation. The ASCA should start small rather than allowing itself to be overwhelmed with bureaucratic overhead—though it will need an area able to manage contracts, IP, and budgets. But organisational size should not stint its budget: the ACSA needs an eye, and the capacity, to seize opportunities and enable scaling when needed. Starting small with scope for growth permits experimentation, learning and adaptation.

Speed should be prioritised, whether through bottom-up matching of deep civilian expertise with military officers who knew the problem in the field, or, as in the case of DARPA, enabling (through funding, infrastructure, and resources) the freedom to explore, prototype and test, including through competitions.

The ASCA should preference human capital and be able to offer the smartest and best people attractive packages, like DARPA, for block periods of time for specific programs. Creative, free-thinking people, and deep expertise, matter. That will almost certainly mean that it will not be a typical public service organisation, which will inherently challenge Canberran expectations—a further reason why unstinting top cover is critical. 

Capable management matters. The government cannot afford the ASCA to be average. Everyone who ever worked in a high-performing team yearns for that experience again: that, and the sort of professional exuberance evident at the SpaceX launch, should be the yardstick for the ASCA. And that in turn means that the government needs to choose the core team with care; it should not assume that current practice—the tick-and-flick 20-minute interview—will serve. 

And the ASCA needs translation mechanisms. First, it should be able to choose military staff able to imagine how technology could be used in practice. Second, it needs the equivalent of the OSRD’s field service office, to advise and learn from how the technology is used in place, in the field. Third, it needs a means of ensuring follow-on engineering and production.

All this can be realised. Many of the same precepts were evident in the early Office of National Assessments: responsibility to the PM, its primary customer; an independent budget; scope to choose the best staff; a recognition of the importance of smart, insightful, and objective intelligence analysis, with a care for the craft; and a strong, capable leader.

Institutions are not set and forget. Even if we manage to break the ASCA free from Defence, establish its own budget line, sufficiently engage the Prime Minister in its activities and, not least, find a modern Australian Vannevar Bush—all in a short timeframe— success is far from guaranteed.  But it would offer us a better prospect than we have following the current tedious pedestrian path.