In the early 1920s, factories were mass-producing furniture that was cheap but ugly. People wanted beautiful furniture, and artisans knew how to craft it, but their techniques didn’t translate to the assembly line and so were impossible to scale. At the Bauhaus School under Walter Gropius, designers learned to translate between the artisanal and industrial ways of working. The Bauhaus developed a design language for creating objects that were both beautiful and manufacturable.
In the 1970s, computers could organize data for applications like accounting or inventory management. But navigating that data required specialized knowledge of terminal commands and memory structures. Researchers at Xerox PARC invented a graphical user interface (GUI) that allowed users to interact with their data though a system of windows, icons and menus. The GUI was a design language that made the complexity of the computer navigable to people who were not specialized computer engineers.
In the 2000s, advances in DNA sequencing technology were driving exponential growth in the size of public DNA databases like GenBank. The completion of the Human Genome project gave researchers access to (almost) the entire source code of a human being and many other organisms soon followed. New disciplines like systems biology, synthetic biology, and computational biology arose to deploy the awesome power of tech against the challenge of understanding and building with biology.
And then suddenly… not much changed. Biology in the 2000s did not undergo a revolution like industrial design did in the 1920s or personal computers in the 1970s. Biology labs in 2010 looked about like they did in 1990. The day-to-day work of a life scientist stayed about the same. Read a journal article and think up a hypothesis. Do an experiment with manual pipettes. Analyze the data with a t-test. Computational tools contributed, but mostly from the sidelines, a minor part of the overall workflow.
As 20-year veterans of the struggle to make biology look more like tech, we are as annoyed by this history as anyone. We thought we’d be vibe coding dinosaur genomes on VR headsets by now. We thought the market would be overflowing with cool new bioproducts that make people healthier, investors richer and the world greener. Instead we visit our former thesis lab and find the latest generation of PhD students doing Western blots on the exact same rigs that we struggled with in 2008. This rate of change is unacceptable.
What went wrong? If exponential increases in data and computing power didn’t transform biology in the genomics era, how can we be confident that they will in the AI era? What needs to happen for biology and tech to talk to each other more effectively?
We don’t think the missing piece is more interdisciplinary thinking. Many smart people recognized long ago that the application of tech to bio would require scientists educated in both fields. Our graduate training included both petri dishes and python scripts and we know plenty of other biotech professionals who can say the same. Interdisciplinary grad programs are necessary if we want to transform biology with tech, but they haven’t been sufficient.
We also don’t think the missing piece is more engineering frameworks. Synthetic biology is a 20-year effort to reshape biology into something resembling an engineering discipline. With tech-like formalism, synbio approaches DNA sequences as composable parts, transcription networks as genetic circuits and the scientific method as a design-build-test cycle. But the tools and techniques developed for synthetic biology have not been widely adopted. Whatever their merits, engineering frameworks have not changed the way most biologists work.
We believe the missing piece is a design language. To us, a design language means “a shared vocabulary and a set of principles that allow people to interact with a complex medium in an organized way.” Bauhaus transformed industrial manufacturing because it was a design language between the artisan and the factory. The GUI transformed personal computers because it was a design language between a user and a file structure. But a similarly transformative design language has never emerged for biology because we have no Bauhaus and no Xerox PARC, no product-focused design centers relentlessly driven to make things that people want.
All the great design languages accompany great products. Braun’s Ten Principles of Good Design, Apple’s Human Interface Guidelines, Google’s Material Design documentation. Design languages emerge from product development because of the unique constraints that product developers face. A product must work reliably in someone else’s hands, scale from prototype to thousands of instances and integrate with systems you don’t control. These pressures force systematic approaches: reusable patterns, clear interfaces, teachable principles.
Academic research doesn’t face these constraints. A successful experiment only needs to work once, in expert hands, under controlled conditions. The academic incentive structure rewards novelty and insight, not robustness and transferability. It simply isn’t the role of academia to ship products, so we should not expect a design language to emerge there.
Outside of academia, almost all biotech research happens in the context of pharmaceutical R&D. The biopharma industry is product-focused, but its unique structure makes it an unlikely incubator for a new design language. Drug development cycles span 10-15 years, far too long for the rapid iteration that crystallizes design principles. Biopharma’s culture of secrecy means that no general community of practice can form to share ideas across projects. The result is an industry that builds products on hard mode, managing complexity without a common language.
If we’re right about this, it means the challenge of integrating biology and tech is structural and deep. AI won’t be able to transform biology if AIxBio is developed only within the context of our existing institutions. If we want exponential gains in compute to drive exponential progress in biotech, we need a new way of working with biology. We need a new kind of entity that builds products, unlike academia, and builds them in public, unlike biopharma.
American Wetware is going to be that entity. Our mission is to develop the design language for biology. We’re going to do it by putting our hands to the hard work of making things that people want. We want to build our own products and we want to partner with teams of all sizes and industries to design bioproducts that people love, learn from, work with, and most of all, are happy to pay for.
Collectively, the three of us have been fortunate to work on dozens of bioproducts that have benefitted researchers, consumers, and patients. As we head back into the workshop, here is how we see the short list of things that we can make that the biotech world needs:
A Test Kitchen for AI Scientists
Today’s AI scientists still exist mainly to process data in the digital world. They will be changed completely once they are deployed against real world problems.
We want to be a sandbox lab for your science AI agents, bashing them against hard wet biology and sending them home best-in-class.
A Design Practice for Bioproducts
As technical barriers fall, strange new markets emerge. Peptides that taste like chicken. Wearable biosensors for THC. Leather made from your own skin cells. Do people want these things?
We want to be the MSCHF of biotech, building limited edition bioproducts to test hypotheses in consumer demand.
A New Model for Basic Research
If we’re right that design languages emerge from product-focused work, there might be a significant blind spot in the way we currently approach basic research.
We’re interested in working with public and private funding agencies to develop research institutes that want to shift focus from publishing articles to shipping tools that scientists really use.
A Content Channel for Building Biology in Public
Design languages were made to be spoken. Sharing stories and tech demos is the best way to test demand for the products that we are building.
We’re looking for partners looking to communicate science through short and long-form video.
A Marketing Firm for Communicating Biology to the World
Design languages enable marketing by making technically complex products legible to their intended users.
We want to be the marketing firm that can sell your frontier AI model, your lab reagent, your food ingredient or your platform biotechnology.
A Think Tank for AIxBio Futures
What does the future of biotech look like when AI allows users to co-create the bioproducts that they use?
We want to work with policy and biosecurity experts to deliver the best possible version of the biology-powered future.
If you’re building in this space, and this vision of product-focused biotechnology resonates with you, we want to talk.
We’re going to find the design language of biology by building things that people want to buy. Today we start by building a design studio that people want to work with.
I love this concept. I will say, though, that to a non-expert it may read as though you are arguing for the replacement of fundamental/hypothesis-driven exploratory research with bioengineering. Of course, in reality, we will always need to explore both. Might be good to acknowledge that, at least.
sounds intriguing, Jake. so is the idea to make some short-run, wild product concepts to see what gets traction? in my mind, it’s trivially easy to do lots of genetic engineering - but finding profitable use cases is much harder. i always use Z Biotics of an example of an idea that is (relatively) straightforward to execute on but was wild to propose.