Vision

The future is open

Rebuilding the machinery of science for our abundant future

The coming years will bring an era of unprecedented scientific and technological advancement. Exponential progress in artificial intelligence is converging with central advances in other key fields to dramatically accelerate the pace of innovation across virtually all domains.

This inflection point provides an unparalleled opportunity – and an urgent mandate – to fundamentally rethink the institutions, systems, and tools that drive scientific progress. At Astera, we want to meet this challenge by seeding a vibrant ecosystem of mission-driven, open-source products and organizations. We aim to play a leading role in the transition to a shared and abundant future by experimenting with unexplored approaches, embracing scientific and technical risk, and leveraging our $2.5 billion endowment to chase impact at scale.

The Approaching Horizon

The next decade will be the fastest period of scientific and technological progress in history (so far). The case in brief:

Exponential Growth in Computing Power and AI Capabilities.

We are reaching a tipping point where advanced AI will increasingly augment and accelerate research and innovation across virtually all domains. As the world wakes up to the revolutionary potential of machine learning, an increasingly large fraction of global wealth, chip manufacturing capacity, and energy production will be plowed into further acceleration.

Growing automation of science and engineering.

Progress in high-throughput science, robotics, computer vision, materials simulation, computational biology, and other tools are allowing more steps of the research, development, and deployment process to be automated. In the near future, we will likely see large-scale efforts to replace the still-firmly-human planning components of science with autonomous systems.

Convergence of multiple maturing technologies.

Fields like genomics, synthetic biology, immunology, structural biology, materials science, renewable energy, neurotech, and quantum computing are now reaching key milestones. The next decade will likely see the combination and synthesis of these advancing technologies in powerful ways that will reshape many aspects of society.

Scientific knowledge and data growth.

The proliferation of sensors, storage, and data analytics have fueled rapid growth in the amount of scientific data available to fuel breakthroughs. This trend will accelerate dramatically as algorithmic and compute advances in ML mean progress is increasingly bottlenecked by the availability of high-quality data, and resources are increasingly allocated to its collection or synthesis.

As the horizon nears, we should not expect our old assumptions about how prosperity is converted into scientific progress to hold – or vice versa. By default, transformative change in the above areas is likely to be largely driven by private investment. Government and other institutions will take more time to adapt, but we can be partners for speeding this up by charting a path.

Our Vision

Given Astera’s risk tolerance, institutional agility, and substantial resources, we believe we’re in a unique position to keep pace with the leading edge of technological development.

In order to navigate this period of dramatic change and to pave the way for a shared, scientifically advanced future, we need to rebuild the machinery of science. Astera’s mission is to contribute to this project by supporting the creation of public goods to accelerate scientific and technological progress – goods like non-proprietary infrastructure, tools, datasets, playbooks, and standards.

We believe:

1
Abundance is Nigh

AI is on the verge of explosive growth. (Scientific) productive capacity is on the verge of explosive growth. At a civilizational scale, we are about to get much wealthier, very fast.

2
Science is Ripe for Transformation

There are huge opportunities to channel research in high-leverage directions, to coordinate at large scales, to improve the way scientific funding is deployed, and to ensure that scientific knowledge is widely shared. We must rethink these core aspects of science holistically, as they are interdependent.

3
We Are At the Crossroads

Our transition to a scientifically advanced future can go two ways: extension and entrenchment of existing systems, or replacement. The next 5-10 years are disproportionately likely to be the critical period for shifting the future of science in one direction or the other. This is our best chance to create new, more rational systems and approaches, and we must act with urgency to bring them about.

Given the novelty of emerging technologies and unknowns about the future, we believe more experimentation around these public goods is needed to develop conviction about what will work. We hope to provide a dedicated space for this exploration. We want to help make science better, faster, and cheaper, while ensuring that everyone participates in its benefits.

A few examples of the kind of programming we hope to support:

1
Data-Centric Publishing

To take advantage of new computational tools that can increase the pace of scientific progress, reduce bias, and root out fraud, we must shift from narrative-centric to data-centric approaches to sharing scientific results.

This requires establishing robust infrastructure, tools, platforms, and best practices around metadata and data formats to facilitate data-sharing and analysis. We can also leverage a shift toward data-centric publishing to transform meta-analysis: We can build infrastructure to make meta-reviews 2 OOM faster, 2 OOM cheaper, auditable, and automatically updatable. These changes could allow us to transition from a system where meta-reviews are tools for summarizing existing evidence on directed topics to one where they can systematically identify holes in our scientific understanding.

2
Open Datasets

Machine learning will be key to rapid advancement in many areas of science, and datasets are the key to realizing the potential of machine learning.

Trying to harvest the necessary datasets from publicly available sources is not currently possible. This is due in large part to heterogeneous data formats, uneven availability, incomplete metadata, and an absence of cohesive, shared infrastructure. It is also exacerbated by current predominant publishing practices and culture, which emerged from a vastly different technological landscape and throughput of information. There is a well-known bias in scientific publishing toward positive, novel results. This can paint a skewed picture of reality by overemphasizing certain findings and underrepresenting others.

In order to harness the power of machine learning and other modern technologies, it will be necessary to create large, high-quality, publicly available datasets that overcome these challenges. This task cannot be left to the devices of companies building proprietary repositories that will be shared with a select few. We plan to contribute to this project at Astera by supporting the generation, collection, or collation of open datasets across many scientific domains by a variety of means.

3
Open-Source Translational Science

Emerging technologies hold immense promise to transform industries and address global challenges, but their progress is often hampered by a lack of accessible, scalable, and interoperable infrastructure.

Many of the tools, datasets, and platforms necessary to unlock exponential advances in these fields are proprietary, siloed, or not yet developed. To help tackle this challenge, we plan to launch programming to accelerate the development and adoption of open-source infrastructure for emerging technologies. We will provide funding, resources, and support for entrepreneurs, engineers, and scientists to create and collaborate on projects aimed at creating critical public goods for these fields.

By bringing together the creativity of entrepreneurial minds with our resources and network, we hope to catalyze a new era of open innovation in emerging technologies. The resulting work will enable researchers around the world to more efficiently design, build, test, and scale cutting-edge solutions, unlocking the potential of these fields to revolutionize various sectors. Importantly, the open nature of these tools and platforms will ensure that the benefits of technological advances are accessible to a wide range of stakeholders, rather than being concentrated in the hands of a few.

At Astera, we believe an abundant future for all is possible – but not inevitable. To bring it about, we must rebuild the machinery of science to be more efficient, scalable, and open. We need visionary leaders who are compelled by the challenge of leveraging all the resources at our disposal to drive transformative change in the way science is funded, conducted, and communicated.

The future is open.