The Living Compiler

Yesterday, UC Santa Barbara announced that molecular biologist Max Wilson had received a $1.7 million award to develop a »nucleic acid compiler«.¹

»Perhaps that is what happens when a metaphor becomes executable«

The idea is extraordinary. Instead of manufacturing molecules that carry biological instructions (DNA or RNA) in a lab and transporting them to a cell, the instructions for making them could be transmitted as light. The cell then manufactures the nucleic acid itself.²

Wilson’s team is engineering yeast with light-sensitive enzymes that respond to different wavelengths, allowing flashes of light to encode a genetic sequence. Imagine someone using a flashlight to send a Morse code message. Wilson is trying to do something similar inside a living cell. Instead of a person decoding the message, the cell receives it and builds it.

The ambition is to reduce some processes that currently take months to hours. But more importantly, it suggests a strange new relationship between information and matter.

We have spent centuries separating the two. A drawing represents a building. A blueprint represents a machine. A genetic sequence represents a biological possibility. A computer program represents a procedure. Then we build machines that execute those representations.

The compiler is a beautiful example. It translates instructions written for humans into instructions a machine can execute. In the early 1950s, computer scientist Grace Hopper helped develop some of the first compilers, creating an intermediary between what we want a machine to do and the machinery that does it.³

Wilson’s LUXCODE project proposes something like a compiler for living matter. But the living machine isn’t just executing instructions. It is also manufacturing what the instructions describe.

In What Is Life?, Erwin Schrödinger described chromosomes as containing a »code-script«.⁴ Molecular biology subsequently developed a vocabulary of information, reading, writing, copying, and editing. The genetic programme became an influential metaphor for understanding heredity and development.⁵ As philosopher Hans Blumenberg argued, metaphors can organise thought, structure questions, and open new domains of investigation.⁶ Here, the metaphor of biological information became a foundation for technologies that could eventually intervene in the thing being described.

Perhaps that is what happens when a metaphor becomes executable. The genetic code first gave us a way of talking about life. We learned to read it, then write it, then edit it. Now we are beginning to build systems in which a living cell might compile an external instruction into new genetic material. The metaphor has become an engineering problem.

And the implications extend far beyond genetic engineering. If information can be converted directly into biological matter, the same underlying principle could eventually change how we produce food, manufacture materials, generate energy, or make medicines. It could alter what is possible in places where transporting finished products is difficult or expensive, from remote environments on Earth to, eventually, other worlds. Imagine arriving on Mars with no building materials, only a machine capable of instructing the dust beneath your feet to become a building.

The same idea can be brought back down to Earth. Instead of transporting prefabricated material to a building site, what if we could bring the instructions and manufacture the building from what that site is made of? Prefab in reverse. The site would no longer simply be land on which a building is assembled. It could become a material substrate with a latent capacity for formation. We might seed materials with possibilities rather than finished forms, then activate those possibilities later.

This would change the economics of architecture, too. Land would no longer be valued only for what can be built on it, but for what it might be capable of becoming. A plot containing a dormant capacity for transformation would have something like stored architectural potential. Call it latent architecture.

The computer separated information from the machine. The compiler connected them. The living compiler might put the computation back into the material. Perhaps that’s where design eventually goes: away from specifying objects, towards specifying the conditions under which matter can form them.

Not a building, or a cell, waiting to be assembled. A substrate waiting to be instructed.

References