Authoring Life by Means of Artificial Biological Intelligence: On the Future of Species by Adrian Woolfson

About ten thousand years ago, in a narrow river valley in what is now the Mexican state of Michoacan, a group of Neolithic farmers began writing a genome.

Presented By Our Partners

They were not calling it that. They did not know the word “DNA.” They did not know that the tall, wild grass they were coaxing, season after season, toward larger and more exposed seeds carried chromosomes or nucleotides or a regulatory code. What they knew was that some plants produced better food than others, that those plants tended to produce children that were more like their parents than not, and that if you planted only the children of the best plants, you would, over many generations, end up with something the earlier generations of your tribe would not have recognized.

The grass they were working with was called teosinte. The thing they eventually produced was maize. By the time European colonizers arrived in the Americas in the fifteenth century, maize had become the dominant food production system of two continents. Today, it accounts for roughly one-fifth of all the calories human beings consume on the planet. The genetic changes that turned teosinte into maize, as modern research has shown, were concentrated in a small number of regulatory genes, the master switches that control how and when other genes are expressed. Small adjustments at those switches produced outsized changes in architecture. Stony-sheathed seeds became exposed kernels. A scraggly grass became a cob. A wild plant became a crop.

That is genome writing. It is slower, less precise, and less ambitious than what synthetic biologists are doing in laboratories in San Diego and Cambridge and Palo Alto today, but it is the same project. The work those Neolithic farmers were doing, whether they could have articulated it or not, was the deliberate reshaping of a living organism’s genetic code to serve human purposes.

That continuity is the quiet argument at the heart of Adrian Woolfson’s new book, On the Future of Species: Authoring Life by Means of Artificial Biological Intelligence, published in February 2026 by Bloomsbury and MIT Press. It is also the argument that makes the book worth reading carefully for anyone thinking about where medicine, agriculture, environmental science, and human health are heading over the next quarter-century. The subtitle is the part that gets the attention. The continuity with the last ten thousand years is the part that does the work.

For readers new to Woolfson, Part I of HealthcareDiscovery.ai’s Featured Voices profile, How Adrian Woolfson Is Helping Medicine Move From Reading DNA to Writing It, traces his career arc from Oxford and Cambridge under the Nobel laureate Cesar Milstein, through two decades in oncology research and development at Bristol-Myers Squibb, Pfizer, and Nouscom, through his tenure as executive vice president of research and development at Sangamo Therapeutics during the clinical readout of the first in vivo human genome-editing trial, to the synthetic genomics company he now leads in San Diego. Part II, Adrian Woolfson and the Battle Over the Future of Genome Writing, examines the critiques, the biosafety concerns, and the open questions raised by his work. This essay is about the book itself, and about why its central argument happens to be one of the clearest explanations we have encountered for why HealthcareDiscovery.ai’s editorial thesis, the meeting of ancestral intelligence and artificial intelligence, is not a marketing line but a serious claim about how biology is heading.

Fred’s Library and the books we have yet to read

The metaphor Woolfson returns to throughout the book is a variation on one of the most famous short stories of the twentieth century. In 1941, the Argentinian writer Jorge Luis Borges published The Library of Babel, which describes a library of indefinite, and perhaps infinite, size containing every possible book that could be written using twenty-five alphabetic characters. Most of the books are nonsense. The meaningful ones, the novels and histories and biographies of all possible human lives, are scattered across the library’s endless expanses like rare islands in a sea of gibberish. Finding them, Borges wrote, would require a lifetime of wandering through aisles that most librarians would never attempt.

The philosopher Daniel Dennett, writing decades later, observed that biology has a similar library. The alphabet of DNA is four nucleotide bases rather than twenty-five letters, but the combinatorial logic is the same. Every possible genome sequence has a defined address somewhere in that mathematical space. A relatively small number of those sequences code for viable organisms. An astronomically larger number are gibberish, incoherent instructions for forms of life that would never assemble. Dennett called this the Library of Mendel. Woolfson prefers to call it Fred’s Library, after Frederick Sanger, the British biochemist who gave us the ability to read any entry in it. A recent essay adapted from the book for The MIT Press Reader sets out the metaphor in its clearest form.

What makes Fred’s Library useful is that it imposes a particular shape on a question that is otherwise difficult to hold in the mind. Nature, through four billion years of Darwinian evolution, has explored a vanishingly small corner of this library. The genomes of every organism that has ever lived, every ant, every dandelion, every pterosaur, every human being, occupy a small island in an ocean of unexplored possibility. Most of what could be alive, never has been. The organisms that did exist, and the ones that exist now, arrived through a blind and slow process of mutation and selection. The library’s remaining territories, the much larger fraction of the map, have never been visited.

Woolfson’s argument is that humans are on the verge of being able to visit them. Not by waiting for evolution. Not by relying on the random biology that happens to arrive. By writing.

A short history of writing biology

The history of how humans acquired this capability is one of the pleasures of the book, and one of the reasons the book reads as well as it does. Woolfson is careful to resist the temptation to present synthetic biology as a sudden arrival. The capability is old. What is new is the scale.

He starts at the beginning, or at least the modern beginning. In 1927, the American geneticist Hermann Joseph Muller reported to a Berlin audience that he had successfully increased the mutation rate of fruit flies one hundredfold by exposing them to X-rays. This was the first time a human being had deliberately altered the genome of a species, using a tool built in a laboratory, in a way that was heritable. Muller received the 1946 Nobel Prize for the discovery. It was a quiet but permanent shift in the relationship between humans and biology. Before Muller, genetic change was something that happened. After him, it was something that could be done.

Featured Partner

Invest in the Infrastructure Behind Modern Medicine

As healthcare expands beyond hospital walls, the buildings and campuses supporting that shift are generating compelling returns for investors who move early. The Healthcare Real Estate Fund offers qualified investors direct access to a curated portfolio of medical office, outpatient, and specialty care facilities.

Learn More →

The next move came in 1955, when the British biochemist Alexander Todd and his colleague A. M. Michelson chemically synthesized, from scratch, a fragment of DNA two nucleotides long. It was a tiny achievement by modern standards and a revolutionary one by the standards of its time. For the first time in the history of biology, a piece of the genetic material that underwrote all living things had been constructed by human hands rather than inherited from an ancestor. The synthetic fragment was indistinguishable, chemically, from the natural variety. Todd had established, in principle, that the entire landscape of possible DNA sequences was accessible to synthesis. All that remained was a matter of scale.

The scaling took decades. In 1977, Frederick Sanger gave the world the chain-termination method for reading DNA, which eventually made genome sequencing routine. In 2012, Jennifer Doudna and Emmanuelle Charpentier described the CRISPR-Cas9 system, which made genome editing precise enough to be clinical. In 2019, a team at the MRC Laboratory of Molecular Biology in Cambridge, including Kaihang Wang, who is now one of Woolfson’s cofounders at Genyro, published the first synthetic bacterial genome to use a recoded genetic code. Syn61, as the organism was called, used 61 of the genetic code’s 64 codons rather than the full set. It was the largest synthetic genome ever built at the time and the first to demonstrate that the fundamental grammar of DNA itself was negotiable. In 2025, Wang and his colleagues published the Sidewinder DNA construction technology in Nature, which dramatically lowered the cost and time required to write long DNA sequences to order. That same year, the Stanford biologist Brian Hie, another Genyro cofounder, published Evo 2, a genome foundation model trained on more than nine trillion nucleotides of biological sequence, capable of generating plausible genome sequences from prompts.

Woolfson’s book is the first accessible, synthetic account we have read of how these capabilities fit together into a single emerging discipline. He calls the discipline Artificial Biological Intelligence, or ABI. The term is deliberate. It is meant to stand alongside Artificial Intelligence as a named technology, rather than being subsumed by it. Its outputs are not sentences or images. They are living organisms and the therapies built from them.

The ancestral case hiding in the book

Here is where the book’s argument starts to intersect, unexpectedly, with the editorial thesis HealthcareDiscovery.ai has been building around for the last year.

The phrase “Ancestral Intelligence meets Artificial Intelligence” was, when we first coined it, meant to describe a tension. On one side, the foundational wisdom about human health that has accumulated across cultures over tens of thousands of years, about nutrition, sleep, movement, breathwork, and mindset, about fermented foods and walking and fasting and community. On the other side, the exponential rise of computational biology, AI-designed therapeutics, genome foundation models, wearable diagnostics, and precision medicine. The two sides were supposed to be doing different work. The foundations kept you alive long enough to benefit from the breakthroughs. The breakthroughs extended what the foundations could achieve.

Woolfson’s book reframes the relationship. The foundations and the breakthroughs, on his account, are not two different things. They are two stages of the same project. The Neolithic farmers domesticating teosinte in the Balsas River Basin were doing, at a vastly slower pace, what Brian Hie’s Evo 2 is doing today. They were exploring the genome library. They were moving, trial by trial, through adjacent sequence space in search of organisms that served human purposes better than the ones they had started with. The fact that they were doing it through selective breeding rather than direct synthesis is a difference of method, not a difference in kind. The work is the same work. The tools have changed.

That reframing has implications. It means, for example, that the fermented foods traditions that HealthcareDiscovery.ai has written about in the context of gut microbiome research are not a throwback to a less scientific era. They are one of humanity’s earliest and most sustained engagements with microbial genome writing, carried out through the slow selection of bacterial populations that produced desirable flavors, preserved food, and, as we now know, conferred measurable health benefits on their consumers. The yogurt cultures that have survived millennia are, in this framing, ancestral synthetic biology. So are the domesticated dogs, the cultivated fruit varieties, the beer yeasts, the bread cultures, the selected chickens. Every one of them is an entry in Fred’s Library that a long line of our ancestors helped identify.

The continuity runs in the other direction as well. It means that the most advanced work being done at Genyro, or at the synthetic yeast genome project, or at the Arc Institute, is a continuation rather than a departure. The humans doing this work are not severing themselves from ancestral practice. They are accelerating it. The tools are different, and the pace is unrecognizable, but the underlying project, the deliberate shaping of biological organisms to serve human purposes, is continuous.

What Neolithic farmers knew

One of the book’s more useful observations is that selective breeding, over enough generations, can produce changes that look, from the outside, like magic. The teosinte-to-maize transformation is the canonical example. Wild teosinte has seeds the size of a grain of sand, encased in hard shells, arranged in spikes of about a dozen per plant. Modern maize has kernels the size of a fingernail, exposed to the air, arranged in dense cobs of several hundred per plant. The transformation took roughly six thousand years, which is a long time by human standards and an instant by evolutionary ones. Recent studies have shown that most of the work was done by changes to a small number of regulatory genes, and that the process accelerated dramatically when second-wave farmers in the Mexican highlands crossed their domesticated lowland teosinte with highland wild relatives, infusing new genetic material into the population.

Woolfson’s point is not just that this was an impressive feat. It is that the technique, which he calls, in the book, a “biological time machine,” has continued to work throughout human history and is still working now. The Victorians took selective breeding to elaborate extremes with dogs, producing breeds like the Dalmatian and the King Charles spaniel whose distinctive traits, he notes, sometimes came with real genetic costs. Modern agriculture does the same thing with corn, wheat, cattle, and salmon. The difference in the synthetic biology era is not that the project has changed. It is that the time machine has gotten faster. What took Neolithic farmers six thousand years can now be done in a laboratory in weeks. What took Darwin’s nineteenth-century breeders decades can be done in hours.

This is where the ancestral intelligence framing does the most useful work. It suggests that the right frame for understanding AI-augmented synthetic biology is not “a radical new capability,” though it is that. It is “a tool our species has been developing, in increasingly sophisticated forms, for ten thousand years, now arriving at a phase of exponential compression.”

The acceleration problem

That compression is not free of consequence. The book does not pretend that it is.

The strongest chapters are the ones where Woolfson describes what changes when a biological capability that used to require thousands of years can be accomplished in months. Some of those changes are clearly beneficial. Vaccines for diseases that have historically defied them become easier to design. Therapies for rare genetic disorders that affect too few patients to justify traditional pharmaceutical development become economically feasible. Crops adapted to climate-changed growing conditions can be developed at timescales that actually match the timescales of climate change itself. Whole ecosystems, Woolfson argues, might be repaired or restored using engineered organisms built for specific ecological purposes.

Other changes are harder to evaluate. The book engages seriously with the biosafety and biosecurity questions that HealthcareDiscovery.ai’s Part II profile of Woolfson covered in detail. The concluding chapters, as the former Wellcome Trust CEO John-Arne Rottingen observed in his endorsement of the book, sketch “a scaffold for a manifesto” for how humanity should navigate the emerging capability. Whether that scaffold is adequate is a matter of serious disagreement among scientists including Eric Topol, Kate Adamala, and Kevin Esvelt. The scaffold being offered at all, in a book by a sitting chief executive of a synthetic genomics company, is itself a notable move.

What the book does not do, and cannot do, is guarantee that the oversight architecture will be built in time. That is not a failure of the book. It is a failure that can only be addressed by regulators, funders, and the scientists themselves. Woolfson’s argument is that the architecture is necessary. The construction of the architecture is someone else’s job.

Where the book leaves the reader

The last third of On the Future of Species is the hardest to summarize and, in some ways, the most important. It is also the part most likely to divide readers along lines that have more to do with temperament than with evidence.

Woolfson argues that the convergence of AI-guided design and at-scale DNA synthesis will, within a generation, make it possible to author organisms that have never existed. Not just to edit existing genomes, which has been the project of the last decade, but to design new ones from scratch, for purposes ranging from therapeutic medicine to biomanufacturing to environmental restoration to, eventually, the reconstruction of species long extinct. His optimism about what this will mean for human health is considerable. His caution about the risks is also considerable. Whether the book’s balance between the two feels appropriate to any given reader will depend, in part, on that reader’s prior relationship with ambitious predictions in biology.

What is harder to disagree with, across temperaments, is the book’s underlying empirical claim. The capability is real. The tools are maturing. The infrastructure, laboratory, computational, commercial, is being built now, in real time, in specific buildings in specific cities, by specific people with specific funding. Whether the field moves faster than Woolfson predicts or slower, it is moving. The question of what kind of oversight, what kind of public conversation, what kind of ancestral wisdom the field carries with it as it moves is up to the rest of us.

Why this matters for HealthcareDiscovery.ai

The tagline we keep coming back to, “Ancestral Intelligence meets Artificial Intelligence,” has always been an editorial bet. It is a bet that the best way to think about modern medicine is neither “the ancients had it figured out and we should go back to them” nor “technology will save us if we just let it.” It is a bet that the two streams of human knowledge, the accumulated wisdom of human culture about how to live in a body, and the accelerating capability of modern computational biology, are not in competition. They are in dialogue. Each one needs the other to produce outcomes that are both humane and durable.

Woolfson’s book is the cleanest articulation we have encountered of why that dialogue is not a marketing construction but a description of the actual shape of the field. The same species that domesticated maize in the Balsas River Basin is the one that built Evo 2 at Stanford. The same instincts that led humans to select for fermented foods, hardier crops, and domesticated animals are the instincts now being applied, at vastly faster tempo and with vastly more precise tools, to the design of therapeutic organisms, regenerative tissues, and entirely new categories of medicine. The Four Shadows that dominate modern chronic disease, cardiovascular disease, cancer, neurodegenerative disease, and metabolic dysfunction, are exactly the conditions that the next phase of genome writing will be pointed at, because they are the conditions that kill the most people and that have resisted the most attempts at conventional pharmaceutical intervention.

The book does not say, in so many words, that ancestral intelligence and artificial biological intelligence are continuous with one another. It is more useful than that. It shows, across 350 pages of argument and example, why they have to be. A species that has been writing genomes, slowly and through proxies, for ten thousand years, is now acquiring the ability to do it directly, quickly, and at scale. The question is not whether that capability will change medicine. It will. The question is whether the wisdom that accompanies the capability, the ancestral sense of what humans are, what we owe each other, and what we should not attempt, scales along with the tools. That is the question the book leaves the reader with. It is a good question to sit with.

Adrian Woolfson’s On the Future of Species: Authoring Life by Means of Artificial Biological Intelligence was published by Bloomsbury and MIT Press in February 2026. It is available from Amazon, the MIT Press, and major booksellers.

References and further reading

• On the Future of Species: Authoring Life by Means of Artificial Biological Intelligence (MIT Press, 2026)
• Biological Babel: Why DNA Is Humanity’s Infinite Atlas, The MIT Press Reader, 2026 (adapted from the book)
• Featured Voices Part I: How Adrian Woolfson Is Helping Medicine Move From Reading DNA to Writing It
• Featured Voices Part II: Adrian Woolfson and the Battle Over the Future of Genome Writing
• Syn61 synthetic E. coli genome, Nature, 2019
• Evo 2 genome foundation model, Nature, 2025
• Sidewinder DNA construction technology, Nature, 2025
• ABI and generative biology: a new paradigm for gene therapy, genome engineering, and engineered cell therapy, Molecular Therapy, 2025
• On the Future of Species, in conversation with Eric Topol, Ground Truths, 2026
• The HealthcareDiscovery.ai library

Free Daily Briefing

The Latest Longevity Science.
Delivered Every Morning.

Join researchers, physicians, and health professionals getting daily breakthroughs in AI-driven medicine, epigenetics, and longevity research.

Support the research that powers this editorial

Subscribe Free

No spam. Unsubscribe anytime. We respect your inbox.