Decoding nature - computation and our physical world
Decoding nature means combining technology and nature to act on information we’ve never had before. Back in 2021's covid winter I was in a London park filling an old plastic bottle with pond water to catch billions of tiny DNA fragments, known as environmental DNA (eDNA). I then posted the sample to a little known company, NatureMetrics, who were taking advantage of rapidly falling costs in genetic sequencing and lab automation to decode eDNA. Combining technology with nature was giving their customers entirely new insights into their operations.
Of course, humanity has been decoding nature for millennia to find food, fuel, medicine and search for suitable partners! Back then, prehistoric programmers started by processing nature’s information using the only computer available – the human brain . Over time, humanity coded complex relationships between natural phenomena into cultural and religious practises (this is effectively indigenous knowledge!), such as the arrival of migratory birds and weather changes.
But this is where humans differ from all other known life: using technology to supercharge our ability to decode nature. Gradually, we invented tools that allowed us to interact more powerfully with nature – resulting in the expansion of natural resource production (agriculture, forestry, mining…) and underpinning the emergence of human civilisation.
Since the industrial revolution 250 years ago, natural resource production has dramatically increased. This has been underpinned by remarkable chemical and engineering breakthroughs. The legendary Vaclav Smil sums it up – growing and harvesting 1kg of wheat in 1801 took 10 minutes of human labour, today it takes <2 seconds. The incredible prosperity we live in today is a direct result. But so is a severely degraded nature.
As nature retreats, we will risk losing connection, delight, adventure, and the myriad ways in which nature provides for us practically. It is no wonder that we find nature rich with beauty and meaning - many heroes of mine have dedicated their lives to conserving nature as a result. And for many businesses, such as those relying on reliable weather patterns and functioning ecosystems, nature loss means lower and less reliable profits.
We also risk losing something fundamental – an irreplaceable repository of information evolved over 3.5bn years. Nature is an unimaginably complex system, and the source of outcomes as remarkable as human consciousness and climate conditions that support intelligent life. The loss of nature risks damaging its own information network – or nature’s intrinsic computational capabilities – before we truly understand it. We risk missing discoveries with profound implications for both people and planet, from novel medicines and materials to the inner workings of mysterious old growth forests whose fungal networks look curiously similar to a brain.
What has changed over the past 5-10 years is we now have the technology to engage with nature’s information in more detail than ever before. We’re poised to decode nature – that is, dramatically grow our understanding of nature’s processes from climate science through to genetics - with breadth and granularity we could have only dreamt of a decade ago. This means a better understanding of wild nature, that is the earth’s natural systems, and engineered nature, where nature’s processes are adapted by humanity, such as in agriculture and biomanufacturing. And importantly, the interaction between the two. This is being driven by costs falling dramatically across three interrelated areas:
Acquiring new data using low cost and precise sensors
Converting this data into information using the latest algorithms combined with abundant and powerful compute
Taking action based upon this information
A short and massively simplified history of how humans have decoded nature. Earth is already starting to look like a computer, with sensors everywhere, hyper-intelligent (across some but not all dimensions) AI and powerful robots to act. The result – precise and powerful application of technology into the physical world.
Cheap rockets equals a flood of data. Despite a few hiccups on the launchpad recently, Starship is set to drop launch costs to $100/kg, from $10 000/kg in 2015. New space-hardened servers, such as from Ramon Space, process more data on the satellite, enabling faster and more precise responses to incidents such as wildfires and other natural disasters.
And GPUs are gulping down this data. The exponential curve on advances in power and cost is insane - Nvidia's GPUs have increased performance of AI inference of 1000x in last 10 years. Despite Moore’s law slowing as transistor size butts up against the physical limits of silicon, Nvidia and other industry leaders are driving breakthroughs through new architectures, software optimizations, and innovations like lower-precision data formats and massive multi-GPU systems, enabling continued leaps in AI performance beyond what traditional scaling alone could achieve.
Robotics may seem an odd bedfellow for an investment thesis with ‘nature’ in it, but it’s essential. Information is particularly valuable if it can be acted upon. Robotics is poised to change the cost structure of ‘getting stuff done’ in the physical world by dramatically increasing the power and precision of acting. Like Kodama, who are supporting foresters manage very remote but fire prone forests with teleoperated machines – if you can fell and bunch timber without needing personnel onsite, you can do more cheaply, safely and consistently. It’s also the best way to get sensors into the field to collect even more data – epitomised by Aigen’s outworld inspired earth-rovers.
At the same time, our ability to work with the most fundamental building blocks of life – DNA and its relatives - is advancing rapidly. In the past decade, genomic sequencing costs have fallen by over ~90%, making powerful genetic insight and manipulation broadly accessible, powered by start-ups like Portal Biotech. This is opening up entirely new possibilities in industries as diverse as agriculture, mining and manufacturing.
This information will change industries responsible for trillions of global value annually, starting with more productive, resilient and adaptable natural resource production. One example is InnerPlant, a company combining genetic engineering, satellites and the latest computer vision algorithms to allow crops to communicate directly to farmers. This unlocks huge opportunities for improving farming efficiency by preventing unnecessary application of pesticides and even bumping up yields.
But we also believe the impact of this information will underpin fundamental breakthroughs in other sectors too. For example, decoding the relationship between a material’s structure on an atomic scale and its function in a battery or solar cell (such as Polaron), or leveraging the structure of the brain to design highly efficient computers (such as BBB-tech).
Ultimately, nature will be valued for its information, not solely its biomass. We want to invest into the companies at this frontier. These are companies that fuse computer processing power with nature’s own information, to create immense value for their customers and nature in the process.
Decoding nature is a core focus area for Systemiq Capital. With our climate lens, we expect this opportunity to play out across multiple domains:
More productive natural resource production – Greeneye’s computer vision spraying reducing herbicide use by 90%, InnerPlant’s soy biosensor communicating plant needs directly with farmers, Timbereye’s fully traceable soft commodity trading platform, Hoxton Farms leveraging machine learning and increased processing availability with proprietary bioreactor design and mammalian gene editing to grow fat without the animal,
More resilient infrastructure – Jupiter Intelligence’s climate risk models increasing resilient asset value vs non-resilient assets, Naturemetrics’ biodiversity intelligence reducing compliance risk and enhancing biodiversity, Helios Ai’s global pricing and supply chain mapping for climate exposed soft commodities
More efficient and adaptable manufacturing – Terrantic’s AI ERP for food manufacturing companies, Caladan Bio’s next-gen sensor dense bioreactor, SirenOpt’s tiny plasma tool to instantly identify materials, Bluestem Biosciences anaerobic superbugs for producing biochemicals, Tastewise’s better formulation software
More powerful and efficient computation – using nature’s own computation capabilities at different scales: from the neuron scale (such as BBB-tech) or sub-cellular scale, a focus area for the UK’s Advanced Research and Invention Agency (ARIA) Nature Computes Better focus area, to concepts such as the Internet of Animals.
New markets for nature’s information and services, rather than biomass – Basecamp Research is building the OpenAI of biology, leveraging handheld gene sequencing devices unavailable 5 years ago with access to enormous amounts of compute power to leverage biodiversity’s extraordinary genetic information for the development of new enzymes and medicines.
If you’re decoding nature to transform incumbent industries, please reach out.
Once the BFR or Starship as it’s now known gets going, expect to see Earth in a whole new way…