A close reading of the early history of the Advanced Research Projects Agency (ARPA) — the ARIA-like agency in the US that funded the early internet and autonomous vehicles — yields a key lesson: exceptional projects were usually the result of exceptional contractors. World-changing ARPA successes like the ARPAnet and early autonomous vehicles were the result of a common shape of R&D org, one structured, incentivized, and staffed differently than typical academic labs or VC-funded startups. These orgs pursued exceptionally ambitious North Star technical visions, fueled by a mix of contracts and grants. We call orgs of this shape frontier research contractors (FRCs).
What makes an R&D org an FRC?
FRCs pursue their technical visions by building technology to solve difficult real-world problems, using customer demand to guide their research direction. History shows us that FRCs can be exceptional platforms for ambitious, applied-minded scientists to move their field of R&D forward. Two historical examples that inspire us include:
J.C.R. Licklider’s early BBN computing team at Bolt, Beranek and Newman. Their technical vision was to build a world of interactive computing, which led them to implement the first version of the internet in the 1960s.
Early Carnegie Mellon’s autonomous vehicle team. Their technical vision of self-driving vehicles led them to successfully take a vehicle cross-country 98.5% autonomously in 1997.
FRCs can be nonprofit or for-profit R&D groups, with technical visions ranging from building individual tools to bootstrapping frontier labs into existence. What they all have in common is their commitment to fueling their ambitious technical work through a mix of contracts and grants.
What is the FRC Launchpad?
As government R&D funders such as ARIA increasingly chase early ARPA-style outcomes, we need ambitious FRCs to do key portions of the work. That’s why we created the Frontier Research Contractor Launchpad, an experimental programme in which FRC founders working in ARIA technical areas were given (1) £100k and (2) a year of operational + strategic guidance in building and expanding their FRCs.
The FRCs involved ranged from a team with only one employee and less than a year old to teams of almost two dozen that were approximately three years old. Upon selection, each founder had a clearly honed North Star technical vision and a sense of what types of customers and funders could fuel their work. All teams, regardless of age, planned to use their participation in the programme to establish a foothold in some new market. And each team was selected with the belief that if they succeeded in doing this, their field would be materially better for it.
Beyond the £100k in financial resources, the programme made a special effort to help founders overcome the temptation to chase funding piecemeal, application by application, and instead build a core customer strategy to reliably fund progress towards their technical visions. We complemented this funding and strategy work with additional resources including founder group workshops, 1:1 sessions, and active network-building on behalf of each FRC.
Results from our first year of the FRC Launchpad
Each of the five teams in the FRC Launchpad was selected with the belief that their success in establishing a foothold in a new market or research area would materially improve their entire field. Each set a goal for what they wanted to accomplish with the programme.
Nine months in, the results are exceptionally promising. The following per-team subsections describe the technical ambition of these organisations and progress towards their goals thus far.
Cultivarium. Cultivarium began its life as a Focused Research Organization (FRO). FROs are often called “nonprofit startups”, and tend to be founded in areas where startup-like efforts are optimal for progress, but in which the massive upside of the effort is public benefit-related, rather than a $10 billion+ commercial upside. Cultivarium was founded to build the tools and models essential to making it faster, cheaper, and easier to work with non-model organisms — enabling greater access to the entire tree of life in biological investigations. Through Cultivarum’s (Cultivarium CEO) programme participation, Cultivarium has succeeded in transitioning from a time-bound, philanthropically funded FRO to a self-sustaining FRC — that can exist indefinitely if the org continues to solve problems for funders with applied needs. Through Cultivarum’s participation in participation in the the FRC Launchpad, Cultivarium won two large contracts: one with an organism-focused philanthropy to ensure Cultivarium’s models and tools speed up the work of the funder’s PIs using a particular organism, and a second large contract of a similar structure.
Living Physics. Living Physics was funded into existence by an ARIA Nature Computes Better opportunity seed - ARIA’s highly flexible ‘fast grant’ funding mode designed to support speculative scientific research with immense potential in a lighter-touch manner than ARIA’s programme funding.. ARIA’s Nature Computes Better opportunity space is dedicated to scaling computational efficiency, and has looked for inspiration to natural systems — which are often orders of magnitude more efficient than modern computers. David Jordan (Living Physics founder) is a biologist who can be thought of as having the mindset of a physicist, the hands of an engineer and the expertise of a biologist. Using the perch his ARIA opportunity seed created, David plans to turn Living Physics into a center of excellence at the intersection of engineering physics and biology. Specifically, his group is building a suite of instruments (e.g., Raspberry Pi-powered bioreactors) and partnerships to move forward the state-of-the-art in theoretical biology, such as understanding how information is processed in biological systems. Biology is a complex subject in which improved measurement and perturbation are often key capacity bottlenecks. Through the FRC Launchpad, David is finding early success in scaling his org from a purely research endeavor to one that accelerates science for others by building instrumentation for ambitious researchers and partnering with research groups that require the skills of people like him. These contracts, which Living Physics achieved through the FRCL, are allowing Living Physics to operate as a self-sustaining R&D org that can take their research into real-world applications. In addition, Living Physics is leveraging the operational know-how gained through the FRC Launchpad to spearhead applications for larger contracts that take advantage of this center of excellence he is building outside of the university and VC ecosystems, one that pairs an exceptionally ambitious research agenda with an intense desire to build artifacts of applied technology.
Syntato. Syntato was also funded into existence by an ARIA grant, from ARIA’s Synthetic Plants programme that is looking to build fully synthetic plant genomes and chloroplasts. Syntato was founded, in part, to address a key programme need: building better applied crop genetics tooling. This was necessary because in agriculture, the venture capital model is not well-suited to areas like “tools for transgenic crop design”. Syntato has built an integrated team of engineers and researchers that is as ambitious as any basic research group, but focused on building good tools for industrial crop biotechnology. During Syntato’s FRC Launchpad participation, Syntato has secured additional ARIA funding, a four-year Bezos PhD studentship, with several larger applications for research projects. These applications add up to over £5M in opportunities, with a high probability that at least £1M will be secured in the next few months.
Basis Institute. Basis builds reasoning engines: AI systems that automatically construct simulators and world models from knowledge and data, and use them to predict the future, explain the past, and reason about counterfactuals. Basis’s North Star technical ambition is a “universal reasoning engine”: a new foundation for AI that unifies probabilistic, causal, and mathematical reasoning over computational models of every kind. Basis has funded itself for several years through a mix of philanthropic support from those who believe in its vision, contracts from funders like ARIA and DARPA, and experimental contracts with companies. While Basis is building towards a general set of technical capabilities, it views working on targeted challenge problems (particularly in the sciences) as an ideal way to make progress. It believes this sustained problem-solving will surface lessons enabling it to steadily generalize its methods. In effect, Basis is trying to use its base of contracts and grants to build a new frontier lab, one with a path to simultaneously being self-sustaining, solving fundamental research problems, and building scalable technology that solves real-world problems. During its FRCL participation, Basis has established several promising contract leads that will expand its capacity to tackle challenges in the UK R&D ecosystem.
Topos Institute. The Topos Institute is an FRC focused on problems of applied category theory and mathematical software engineering, founded with the deep belief that category theory can be a useful, coherent modeling language in the sciences, particularly in problems that require communicating across disciplines. Similar to Basis, Topos seeks to increase the generality of its systems by continually solving specific challenge problems. Topos has also funded itself for several years with a funding mix similar to Basis. During its FRCL participation, the org has succeeded in diversifying the variety of contract partners in its Oxford office, and is actively considering how it might use the corps of talent it has built up to build increasingly useful pieces of technology for users. Many Topos employees would be at home in either top mathematics departments or AI labs. Its success in building a financially sustainable FRC enables this world-class group to marshal their talents in service of a technical goal that is simultaneously big-if-true, but not on the radar of the VC ecosystem.
If this sounds like you, apply to be a part of our second cohort!
Due to the success of the first FRCL cohort, ARIA is expanding our next cohort to eight teams! Read more about the programme here [link to RenPhil website]. We’ll also be hosting a webinar for any interested applicants on Aug 10 at 5pm BST: register here! The FRC Launchpad team — Eric Gilliam and Janelle Tam — are eager to chat with any team that can substantially move the work of any ARIA opportunity space forward. Feel free to reach out to them at eric.gilliam@renphil.org and janelle.tam@renphil.org.
Strong applications will have both a big-if-true technical vision and a clear thesis for a set of contracts and grants which can fuel the vision indefinitely. And these orgs should be building towards goals that could not be built within academic labs or the VC-funded path.
Existing ARIA Creators looking to find a way to use the perch offered by their ARIA to turn their efforts into self-sustaining FRCs are particularly encouraged to apply. The shape of each founder’s technical ambition will vary — FRO-like, frontier lab-like, center of excellence in some area, ambitious CRO-like efforts, etc. What’s important is the founders’ commitment to building an R&D organization that is simultaneously ambitious and able to financially sustain itself through a mix of contracts and grants.
If that sounds like you, check out the application here! Applications close Aug 21.