New attractor points for scientific utility
This is a variation of the talk I gave at Funding the Commons Conference, Paris, 2023
Modern science is witnessing unprecedented knowledge generation, with thousands of papers and patents published daily. Yet the return on R&D investment is diminishing, and the effort required to achieve technological milestones, such as doubling chip density, has grown exponentially since the 1970s. Ideas, it seems, are getting harder to find.
Recently, we've seen an explosion of new initiatives across all dimensions or how science is done and applied in the world. Below is a rough schematic of the emerging system (not comprehensive).
The task ahead is to realign our scientific endeavours towards maximising the utility of this knowledge, essentially catalysing the transformation from "what we know" to "what we can do". To increase the utility of science, I would argue that we need to create attractor points that coordinate talent, knowledge, and capital.
Funding landscapes
Funding drives both research and application. However, there’s a paradox: there's plentiful capital but inefficiently distributed. Despite the global availability of $300bn in venture capital, dwarfing the global R&D budget of $1.5tn, early-stage "deep-tech" ventures grapple for financial support while their later-stage counterparts witness a surge in capital influx. The existing academia-driven rewards system further exacerbates this disparity. This system is rooted in traditional structures and often leans towards consensus, inadvertently sidelining risk-taking innovations. Furthermore, the overarching pressure on researchers to frequently publish often overshadows the primacy of quality, leading to a volume-over-value conundrum.
New funding mechanisms
Rapid Grants: The need for swift, bureaucracy-free funding mechanisms has never been more evident. The inception of Fast Grants, especially in the backdrop of global crises like the COVID-19 pandemic, underscores this urgency. Launched in a mere 10 days, Fast Grants set a precedent by ensuring grant decisions within 48 hours, offering a beacon of hope and agility in a traditionally sluggish system. Longevity Impetus grants were also inspired by the Fast Grants model, but with a different thematic focus.
Crypto-native platforms: Embodying the spirit of decentralisation and collaboration, crypto-native platforms are trying to reshape the funding landscape. Molecule.to, for instance, exemplifies this transformation by fostering a collaborative environment for drug development, transcending traditional silos.
Incentivizing tool development: As science advances, the need for innovative tools and platforms becomes paramount. The current funding ecosystem, however, seems to lag in this regard. The challenge lies in ensuring that financial mechanisms not only support research but also incentivise the development of tools that catalyze cross-disciplinary collaboration and data sharing. Philanthropy has jumped in to support open research infrastructure efforts. For example, Mark Zuckerberg and Priscilla Chan announced that they were donating 99% of their Facebook shares to CZI to “cure, prevent, and manage all human diseases in our lifetime”, to be allocated over ten years. Other notable examples include the Critical digital infrastructure programme by the Ford Foundation and Data and computational research by More Sloan Foundation.
Talent: The driving force of scientific progress
Future science demands a diverse talent pool, including researchers, entrepreneurs, and interdisciplinary experts. However, there's a misalignment in the talent landscape, with many STEM PhD holders leaving the science and technology sector. This shift requires rethinking talent development beyond traditional academic routes.
Presently, the talent landscape reveals stark contrasts. A staggering 80% of STEM PhD holders find themselves outside the science and technology sector a few years post-graduation as there aren’t enough academic jobs.
This requires a paradigm shift in how we perceive and nurture talent, moving beyond traditional academic pathways.
Empowering early-career scientists: The modern scientific landscape often favours seasoned researchers, inadvertently sidelining young, dynamic minds. Organizations like New Science are combatting this bias by supporting early-career researchers, ensuring they have the requisite resources, mentorship, and platforms to bring their innovative ideas to fruition.
Training for scientific application: A significant challenge lies in the predominant focus on novelty in scientific training. The future demands individuals adept at translating knowledge into actionable solutions. To this end, initiatives aimed at bolstering the number of ambitious, skilled professionals dedicated to applying knowledge are crucial. An example is the Venture Science Doctorate by DSV, which I’ve had the privilege to help design and deliver. Recognising the gap between academic training and real-world application, programs like the Venture Science Doctorate aim to bridge this divide.
Promising potential scientists with little to no track record. Today, those academics are recruited from a small pool of people who have attended elite undergraduate or master ’s-level institutions, allowing their students access to research experience at an early stage. This narrows the talent pool massively – we are missing many, many potentially groundbreaking researchers who have not had precisely the life circumstances at an early age necessary to find themselves in an elite graduate program. Silver Beach is a new research organization working to reach 100x the number of top-tier scientists worldwide. They have developed Researcher Seed, a lightweight mechanism for funding new researchers with no previous published work, particularly those not part of a traditional graduate program in their field.
By adopting a multifaceted approach to talent cultivation, the scientific community can ensure a robust, dynamic, and diverse talent community. The key lies in recognising the evolving needs of the scientific ecosystem and crafting strategies that resonate with these shifts.
Institutions: Bridging Knowledge and Application
In the current scientific paradigm, many science ventures emerge from the confines of university labs, with the top 5 patent-producing universities in the US producing one-third of new biotech companies. However, while this tech-driven approach is commendable, it doesn't always guarantee success. Evidence suggests that nearly 47% of university-licensed biotech startups face the daunting reality of failure. This underscores the pressing need to shift from a technology-centric approach to a problem-centric one, ensuring that scientific endeavours are rooted in real-world challenges.
It is rarely, if ever, true that all the expertise or all the advances needed to achieve a breakthrough are resident in one laboratory or organisation. Currently, there is a lack of coordinated networks of diverse, multidisciplinary teams from multiple organisations, all working together to solve a problem they cannot solve alone that is testable and measurable. Wellcome leap, develops programs that aim to tackle the above challenge and deliver breakthroughs in human health over 5 – 10 years.
The traditional model of public-funded institutions often falls short when it comes to funding R&D projects that necessitate tight coordination and teamwork. In response, we're witnessing the rise of novel research entities like Focused Research Organizations, Arcadia, and ARC. These organizations champion centralized research programs, addressing challenges that demand scale and coordination but might not be immediately profitable. Expressly time-bound and outcome-driven to prevent mission creep and organisational ageing.
Outcome-oriented scientific venture creation: Trying to find a gap for technologies and inventions rather than starting with the problem/opportunity leads to misaligned incentives and biased, sub-optimal matching. Institutions need to transition from mere invention hubs to outcome-driven entities. For instance, with its outcome-oriented approach, Deep Science Ventures brings together scientific knowledge and entrepreneurial scientists to build ventures for human and planetary wellbeing.
Infrastructure for Enhanced Learning and Utility
The key to future scientific advancement, IMO, lies in developing a robust infrastructure that optimises knowledge production and maximises utility.
This evolution includes several critical aspects, which I will describe briefly below. The aspects chosen below are inspired by the work of Brian Nosek, who describes the layers required to change a research culture as a pyramid. At the pyramid's base is "Infrastructure" with the directive to "Make it possible." Above it lies "User Interface/Experience" to "Make it easy," then "Communities" to "Make it normative," followed by "Incentives" to "Make it rewarding," and at the top is "Policy" to "Make it required."
For the future of applied science, I’d like to see:
1) Easier outcome coordination via new digital infrastructure that deepen our understanding and application of knowledge.
The practical application of scientific knowledge requires sophisticated outcome coordination, achievable through (but not limited to) tools like discourse graphs and roadmaps and better ways to connect and coordinate talent. These provide detailed frameworks for organizing and understanding knowledge, enabling quick identification of problems and solutions. Moreover, the potential of LLMs in shaping the future is immense. These models can serve as knowledge coordinators at a higher order, offering insights, feasibility analyses, and risk assessments. They can also aid in bottleneck analysis, helping to identify and address factors that impede research and development. Moreover, by tapping into the combinatorial potential of innovation, LLMs can guide us in unearthing novel technoscientific combinations, thereby spurring groundbreaking discoveries. Please reach out if you are working on this; we are cooking something at DSV.
2) Establishing communities of practice to foster compound learning
The emphasis on compound learning underscores the importance of rapid, collective knowledge acquisition and hyper-productive research communities. New roles like field strategists and synthesis experts harnessing collective intelligence to predict outcomes are emerging. But, science extends its outputs beyond traditional papers to include elements like code, data, and infrastructure. This approach, coupled with a metascience ledger that tracks conception, implementation, outcomes, and lessons learned, can significantly amplify our understanding and application of science. The challenge lies in dynamically mapping these diverse outputs to identify gaps and interdependencies within the system. This mapping is crucial for the system to understand its own capabilities, revealing to funders where to invest and how various components are interdependent. For example, with Open Sustainable Technology, we provided the first analysis of the open-source software ecosystem in sustainability and climate technology that informed better funding towards the ecosystem. How can we have this constantly for all scientific artifacts everywhere?
3) Designing incentives that are specifically tailored to the desired outcomes.
The final puzzle piece is crafting incentives that genuinely resonate with the scientific community's goals. In the realm of mechanism design, there is a growing recognition of the need to strategically utilise various mechanisms that stimulate technoscientific markets, addressing supply and demand at different stages of technological development. This approach requires collaboration between the public and private sectors to create synergistic programs.
For instance, impact bonds are ideal for scaling up interventions that have proven effective. When funders are uncertain about the outcomes, retroactive funding emerges as a solution. Another innovative strategy is using advanced market commitments, which is particularly effective in ensuring a viable market for a product post-development. This can take the form of quantity or revenue AMCs in high-cost or price AMCs when demand uncertainty is a significant factor. Additionally, the choice between providing prizes and grants represents another facet of this diverse mechanism design landscape. Each of these tools plays a unique role in fostering a conducive environment for scientific and technological advancements. How could we pull the right mechanisms for applied RnD?
in lieu of outro
Our ability to solve society’s different problems at the scale required in light of systemic crises like climate change will require us to drastically increase the pace by which we apply science to problems, coordinate and share value equitably, and collectively learn at greater speeds.
As we contemplate the future of the scientific economy, numerous questions arise. What role will AI assistants play for every scientist and inventor? How will these models be implemented, and what economic and value exchange networks can be built upon autonomous agents? How can we create modular, multi-player, and multi-scalar roadmaps? Crucially, we must consider how to build effective conduits of knowledge exchange within the actors of this emerging economy so we can learn what works. Such an exciting time to be working on these!
Many thanks to Dom Falcao, Mark Hammond, Joel Chan and others for all the conversations and the work.