I've been building a library. It isn't quite finished, and I'd like to show it to you anyway.

Last week I said I'd share a fully functioning library, but it's taking a little longer than I planned. Reading matters and building infrastructures carefully takes the time it takes, and I'd rather give it that time than hand you something faster and thinner. So, one more week, altough you are more than welcome to take a peek here: https://www.biodesign.academy/reading-room 

That said, the Reading of the Marble Berry itself is finished. So here it is, ahead of everything else. The care is most of the work, and I hope you'll feel it in the pages.

Download it here:

Reading_MarbleBerry_v5_brand-v3.0.pdf

1.78 MB • PDF File

Download

It's a PDF, on purpose. Print it, mark it up, keep it beside you while you work.

The first Reading follows one blue all the way down

Two weeks ago I wrote here about the brightest colour in nature, and how it holds no pigment at all. That was the short version. This is the full reading which will be our first instalment in the new library.

It names the property precisely, the angle-shifting metallic blue of the fruit skin. Then it shows you how to find out what carries it, with two questions you can run on any matter you work with. Herbarium fruits collected in Ghana in 1974 are still bright, so the colour outlives the plant. Grind the berry and the blue is gone, because the colour lives in an arrangement. Which lands it at a structure you preserve.

Then it keeps going, which is the part I most wanted to build. Cellulose synthase, the enzyme laying the fibres, with three UniProt accessions you can paste in and open in about a minute.

The microtubules that rotate layer by layer and turn straight fibre-laying into a spiral. Why the fruit reflects both left and right handed polarised light, cell by cell, which no pigment could do.

And something you could actually make at the end of it: Cambridge's cellulose films, made roll to roll, coloured by geometry alone, ground into a biodegradable glitter with no dye in it.

That's the shape most Readings will take. One named matter, the actor that carries the property, sources you can open, and an honest account of what the property asks of the living thing it came from.

Here's what the rest of it will hold

Seven doors: mycelium, algae and cyanobacteria, bacteria, fibres, bio-cement, slime mould, and the uncommon in-between. Each one opens on a foundational read, the groundwork a matter needs before anything deeper makes sense, with Readings following every weeks.

Most of the shelves are still bare with only scheduled contents metadata in place, and the foundational reads are what I'm writing this week. You're seeing it early, which is the nicest stage to arrive at, because you'll watch the thing fill.

Come and look around: [THE LIBRARY READING ROOM]

Thank you for reading this letter these past weeks, and for giving me the extra one. The library exists because of you.

Yours,

Raphael

P.S. If you teach: the Reading is written to sharpen one person's judgement, and it isn't licensed for a cohort or a VLE. Something you can teach from is a different build, and I'd like to make it with you. Reply and tell me what you're teaching.

There's no right way to start a biodesign project. Some people begin with a material they've fallen for, some with a place or a community they want to work alongside, some with a question, a technique, or an ecological worry they can't put down. Those are all good doors, and none is more right than another.

So take this week's letter as one option among many, not a method you have to adopt. It's a way of thinking I reach for in two moments: 1) when a living material has been described to me with a promise I find slippery, and 2) when I'm stuck at the very start, holding something alive and not yet sure what I'm actually working with.

When it does help, it's usually because of a single move. Before you design with a living material, it's worth finding out what actually carries the property that drew you to it.

Name what you're holding and you can see where the real work sits, and how a biodesign promise (defined as “the sentence about what a living material does that a design project has come to rely on”) is most likely to fail: chasing a purer molecule when the property is really a geometry, say, or promising a shippable material when it only exists while the organism is alive.

Here are the four kinds of thing a property can live in. Try them on, and keep them only if they earn a place in how you work.

Every property rides on at least one of four things

Whatever a biological or living material is sold on, some colour, some strength, a way it seems to sense or heal, that property depends on one of four things.

I've started calling that thing its biodesign carrier: the level of organisation, or the relationship, the property primarily depends on. Take it away and the property goes with it.

These are not sealed compartments: A property can lean on more than one at once, so the useful question is which one is indispensable: which level has to stay intact for the property you care about to survive. That one is the primary carrier; the rest are supporting conditions.

Naming the property is yours to do: "the angle-shifting blue" rather than "colour", "re-seals a cut within a day" rather than "it heals". That takes looking closely, over days and in different light, and no AI tool can do it for you.

Working out which of the four carries it is mostly something the material itself will tell you, if you test it and give it time. I'll come to how, and to the one narrow place an AI tool helps. What the four add is the part nothing else gives you: knowing what each one then asks of you.

One example runs through the rest of this letter. The marble berry, a small forest fruit, holds an intense blue that comes from no blue pigment at all. The colour is carried by an arrangement of ordinary cellulose, stacked in fine helical layers in the outer cell walls of the fruit.

So it belongs to the second kind below, and the work is to preserve that geometry, not to chase a blue that was never there. I'm writing a full read of it now, taken all the way down, and I'll share it at the end of this week, when the members' library opens (Vignolini et al., 2012).

Each of the four asks something different of you 

Each one below says the same four things in plain terms: what actually carries the property, how you can tell you're looking at it, what it then asks of you to hold on to it, and the way it most often falls apart.

1. Substance (keep it). 

Here the property lives in a molecule: a dye, a pigment, a compound, a protein. You can tell because you can take the molecule out and it still works, even dissolved in a drop of liquid.

So you keep it, and the work is chemistry: making enough of it, pure enough, and stable once it's out of the organism. It falls apart when the molecule breaks down, when there's too little to matter, or when it stops working the moment it leaves the living thing.

Solid liquid pigment extraction. Image: Fer Angelone. CC0 1.0 via Wikimedia Commons.

2. Structure (preserve it). 

Here the property lives in an arrangement: a spacing, a geometry, the way something is packed. You can tell because it survives being taken out, but not being ground up or reshaped. Wreck the arrangement and the property goes, even though nothing was removed.

So you preserve it, and the work is fabrication: making that geometry again and holding it through drying, handling, and scaling up. It falls apart when the arrangement collapses as the material dries, is handled, or is made bigger. The marble berry's blue lives here.

Structural colour, remade by hand. Elissa Brunato's Bio Iridescent Sequin holds this shifting blue-to-rose with no pigment, dye or metal in it. The colour comes from ordered cellulose, the same move the marble berry makes in a fruit, here reproduced in a material you can sew onto cloth. Keep the arrangement and you keep the colour. Let it collapse as the film dries or scales up, and the colour goes with it. That making-and-holding is the whole of the preserve strategy. Image credit Elissa Brunato. CC BY-SA 4.0 via Wikimedia Commons.

3. Living process (tend it). 

Here the property lives in something the organism is actively doing while it's alive: sensing, repairing, growing, feeding. You can tell because it stops when that activity stops.

So you tend it, and the work is keeping the organism alive, fed, and in the right conditions, at the size and for the length of time your design needs. It falls apart when you can't keep it alive long enough or large enough, and the property dies with the organism.

This one is easy to oversell, because the activity is often shown briefly, at lab scale, and then described as if it will stay dependable in use.

A flask of dinoflagellates doesn't hold a colour, it performs one. The blue flash is a living reaction, triggered when the cells are jostled, and it ends when they perish, which is the whole of the "tend it" strategy, and the whole reason it gets oversold. Image: Mattfrantzdotcom. CC BY-SA 4.0 via Wikimedia Commons.

4. Relational system (tend the relationship). 

Here the property depends on an interaction between organisms, usually together with the conditions around them, and no single partner has it alone. You can tell because it weakens or goes when you separate the partners, or when the conditions that hold their interaction are disrupted.

A lichen greens bare rock through a fungus living with one or more photosynthetic partners, usually an alga or a cyanobacterium, inside a wider microbial community.

So you tend the relationship, caring for the partners together and the conditions that sustain them, rather than putting one of them to work. It falls apart when a partner drops out or the conditions shift, and it's easy to misread as a portable material when the property really depends on keeping a whole relationship alive (Karana, Barati and Giaccardi, 2020; Groutars, Kim and Karana, 2024).

Lichens greening a garden ball: not one organism but a partnership, a fungus with an alga or a cyanobacterium, held by the conditions around it. Tend the relationship and you keep it. Photo: Acabashi, "Easton Lodge Gardens, Little Easton, Essex, England ~ ball finial 2," via Wikimedia Commons, CC BY-SA 4.0 (not modified).

The idea is older than it looks 

This is a well-grounded instrument, not a new theory. It carries two established ideas into the moment a promise reaches you.

The first is the materials idea that a property lives in a substance or in a structure, and that a structural property is set by how a thing is made, not only what it is made of (National Research Council, 1989; Olson, 1997).

Many biological material properties arise from hierarchical arrangements across scales, not from chemical composition alone (Fratzl and Weinkamer, 2007; Meyers et al., 2008; Vincent, 2012), which is why the marble berry's blue is a stack of ordinary cellulose and not a dye.

The framework keeps that substance-versus-structure distinction and adds the two dimensions a materials paradigm built for inert matter leaves out: an ongoing living process, and a relationship between organisms. With them comes the separation test, which asks whether a property survives once it is taken out of the living thing, or the relationship, that carries it.

The second is the biological view that many organisms live and work through lasting associations with other organisms (Sagan, 1967; Zilber-Rosenberg and Rosenberg, 2008; Gilbert, Sapp and Tauber, 2012), so some properties we can see arise from the interaction between partners rather than from any one of them alone (Anderson, 1972).

Organisms also build, and then depend on, the places they live in (Odling-Smee, Laland and Feldman, 2003): they do not just occupy an environment, they change it and can come to rely on the conditions they helped make.

Together these ideas support the relational category, and the idea that a relationship between organisms, held in its conditions, can be the real carrier.

Most of this you find by hand, and by staying with it

I want to be careful here, because this is the part that matters most. Which of the four you're in is, at heart, a physical question, and the material will answer it more honestly than any description, if you let it.

Does the colour survive once a piece has dried? Compare an already-detached sample before and after, under the same light. Does it survive the structure being disturbed? Disturb a small sacrificial sample and set it beside an intact one. Does the repair depend on the organism still working? Change one condition at a time in a culture you can keep safely, and keep a control.

For a relationship, watch how the property shifts as conditions vary in the wild, or lean on published experiments, rather than pulling a lichen off its rock yourself.

These are diagnostic probes, not proofs: use a control, a safe or sacrificial sample, and the least destructive test that can answer the question. Run them on a bench or on a slow walk in a wood, and the watching is the whole education.

It is also how you come to care about the thing, which nothing can hand you and no shortcut can fake. Time spent with a living material, letting it stay a bit strange, is not a delay before the real work. It is the work.

So where does an AI tool come in? In one narrow place, and only after you've looked. When you want to know what is doing the work underneath, the specific molecule, the exact structure, the biology can sit behind a wall of papers you may not have time to climb. An AI tool (one of the chatbots you already use) can hand you a way in fast.

Use it to suggest search terms, name the candidate mechanisms, and point you at possible papers. Then open those papers and check them yourself: the title, the authors, that the paper is real, and that the passage really says what you were told. A fluent answer is not a proven one, and a fluent model will invent a convincing reference, so the wobble is worth catching.

The AI tool is a door into the literature, not the evidence itself. It's standing in for a library card, not for the bench, and not for your own judgement.

And keep the last question for yourself. An AI model might help you find evidence about the cost, but it cannot settle it for you: what does this property cost the organism or community from which it comes? That one you earn only by paying attention, over time, to the living thing in front of you.

That full read of the marble berry, how the colour comes to sit in the structure, the cell-wall growth that lays down those layers, and every source you can open yourself, arrives at the end of this week, with the launch of the members' library. If you are new here, the free Biodesign Promise Worksheet is the place to start, for taking a promise apart in the first place.

That is what From the Molecule Up is for. It builds the thin layer between a living-material word and the thing that actually does the work, one piece per issue, in the open, before it is a book.

Before you go, one favour. If a "living", "self-healing", or "grown" promise has stayed with you, from a product page, a degree show, or a studio brief, reply and tell me. I will read the sharpest of them through these four in a future issue.

Until next time,

Raphael

P.S. For those of you teaching: this runs better with a real sample than a screen. Give students a material they can handle, an afternoon, and one instruction: name the property, then test it.

Pick one safe test that fits the material, comparing it before and after drying, disturbing a sacrificial piece against an intact control, or changing a single growth condition while keeping one culture untouched, and watch what survives.

Only when they're stuck on the biology underneath should an AI tool come out, and even then as something to check, not to trust. They land on keep it, preserve it, tend it, or tend the relationship, with their own hands as the evidence.

References

Anderson, P. W. (1972). More is different. Science, 177(4047), 393–396.

Fratzl, P., & Weinkamer, R. (2007). Nature's hierarchical materials. Progress in Materials Science, 52(8), 1263–1334.

Gilbert, S. F., Sapp, J., & Tauber, A. I. (2012). A symbiotic view of life: We have never been individuals. The Quarterly Review of Biology, 87(4), 325–341.

Groutars, E. G., Kim, R., & Karana, E. (2024). Designing living artefacts for multispecies interactions: An ecological approach. International Journal of Design, 18(2), 59–78.

Karana, E., Barati, B., & Giaccardi, E. (2020). Living artefacts: Conceptualizing livingness as a material quality in everyday artefacts. International Journal of Design, 14(3), 37–53.

Meyers, M. A., Chen, P.-Y., Lin, A. Y.-M., & Seki, Y. (2008). Biological materials: Structure and mechanical properties. Progress in Materials Science, 53(1), 1–206.

National Research Council. (1989). Materials Science and Engineering for the 1990s. National Academy Press.

Odling-Smee, F. J., Laland, K. N., & Feldman, M. W. (2003). Niche Construction: The Neglected Process in Evolution. Princeton University Press.

Olson, G. B. (1997). Computational design of hierarchically structured materials. Science, 277(5330), 1237–1242.

Sagan, L. [Margulis, L.] (1967). On the origin of mitosing cells. Journal of Theoretical Biology, 14(3), 255–274.

Vignolini, S., Rudall, P. J., Rowland, A. V., Reed, A., Moyroud, E., Faden, R. B., Baumberg, J. J., Glover, B. J., & Steiner, U. (2012). Pointillist structural color in Pollia fruit. Proceedings of the National Academy of Sciences, 109(39), 15712–15715.

Vincent, J. F. V. (2012). Structural Biomaterials (3rd ed.). Princeton University Press.

Zilber-Rosenberg, I., & Rosenberg, E. (2008). Role of microorganisms in the evolution of animals and plants: The hologenome theory of evolution. FEMS Microbiology Reviews, 32(5), 723–735.

Structural colour, built not painted: the marble berry's blue comes from stacked fibres, not a dye. Illustration: Biodesign Academy.

Dear {{first_name | reader}},

Someone hands you a small hard berry from a plant called Pollia condensata, and it looks like a bead of polished blue metal. You turn it toward the light and the blue glints, flecked here and there with green and violet, like a tiny mosaic, the way an oil film shifts on water or a beetle's back shifts as it moves.

Pollia condensata, also known as marble berry. Image by Juliano Costa. CC BY-SA 3.0 via Wikimedia Commons.

Nobody painted it on. A plant grew it and the colour is the plant's own, made for its own reasons. Researchers think the plant spends on colour instead of food: the bright shine draws birds who take the berries, perhaps to display them, and the birds scatter the seeds, so the plant gets carried to new ground without having to grow a costly, edible fruit.

You could stand there turning the berry in the light for a while. Then you ask the plain question, What is this colour actually made of? And the answer is stranger than it looks.

Anatomy of Pollia condensata fruit. (A) SEM image of the fruit surface showing smooth cuticular layer. (D) TEM of the cellulose microfibrils that constitute the thick cell wall in layer 1. Image adapted from Vignolini et al., PNAS 2012.

There is no blue in that blue. Nothing in the fruit is a blue substance. The colour is made entirely by how the plant stacks tiny fibres as the fruit grows, and once you see that, the way you would work with this material changes with it. This week is about that shift, and the kind of attention it asks of you.

Fibres stack in a spiral, and the spiral is the colour

When this fruit grows, the cells in its skin lay down fibres of cellulose, the same ordinary stuff that stiffens a plant, in fine layers, each layer turned a little from the one beneath it, winding round like a spiral staircase.

The spacing of that spiral is close to the width of a wavelength of visible light, so light bouncing off the layers lines up at one colour and cancels the rest, and a single strong colour comes back to your eye.

The structure behind the colour. In the skin of the marble berry, cellulose is laid down in layer after layer, each turned a little from the one below, winding into a spiral. The spacing of that spiral is about the width of a wavelength of light, so the stack throws back a single brilliant colour, here the berry's iridescent blue, with no pigment in it at all. Illustration: Biodesign Academy, after the structure in Vignolini et al., PNAS 2012.

This is structural colour, rather than a colour that comes from a dye. A butterfly wing and a soap bubble are coloured the same way, with no coloured substance in them at all. Turn the berry, and the colour shifts, because the angle changes which wavelengths line up.

Each cell winds its own spiral, and cells wind them a little differently, so the fruit is not one flat blue but a scatter of separate flecks, blue next to green next to violet, like a pointillist painting made of single dots.

Hold the arrangement, work its conditions

The colour is something that comes with the way the plant lives, and a person happened to notice it. When you read the material down to what is doing the work, ordinary cellulose stacked with great precision, you are not finding a pigment to take. You are looking more closely at the plant's own life. That is the same care many of you already bring to living materials, followed one scale further in.

The extractive instinct, the one most of us are trained into, is to find the pigment and lift it out, or to hunt for a variety that makes a bluer one. It treats the plant as a store to draw from. And often there is something to draw.

Some colours really do live in a substance you can pull out and move: the purple-blue of a blueberry is a pigment, a molecule called anthocyanin, that can be pressed out and used as a food colour, and indigo is a single dye molecule that people have extracted from plants for centuries and carried onto cloth far from where it grew.

Photo by C Cai on Unsplash

Those are colours that live in a substance, and the substance can be moved. This colour is not one of them. Grind the marble berry to powder and the blue is gone, because there was never a blue substance in it. The colour is not a thing sitting inside the fruit; it is the spiral the cells wind as they grow. You could spend months looking for a pigment that was never there.

There is a clear sign that the colour lives in the arrangement and not in a pigment. Fruits of this plant collected more than a hundred years ago, kept dry in a cabinet, are as bright today as the day they were picked. A pigment would have faded long ago. This colour has not faded, because there is no pigment to fade. The structure holds, so the colour holds.

An illustration of a dried marble berry. The fruit is no longer alive, but the blue holds, because the colour is a structure and not a pigment that can fade.

What the colour asks of you instead is to stay with that arrangement, and with the conditions that make and hold it. This can be done, and it is worth being honest about how. A few years ago, scientists built the same kind of colour with no plant at all: they let cellulose, broken into tiny rods, settle and dry into the very same spiral.

It colours itself as it dries, no dye added, in sheets you can make by the metre and even grind into a structural-colour glitter. The colour is set by how tightly the spiral packs, and you steer that packing through the conditions around the film, never by adding anything:

• How the material is laid down

• How it is dried

• How much moisture it sits in

Leave it in wet air and the spiral swells and the colour slides toward red; let it dry and it slides back. That is closer to tending a relationship than to mixing a dye to a recipe.

And it asks for honesty about what you end up holding. The plant is alive as it winds the spiral, but the dried berry is not, and the cast film never was. The colour outlives the life that made it, which is real and useful, but it is a structure, not a living process. Calling a beautiful dead geometry "living" is a costume, not a description.

Keeping that line clear, not calling a structure alive, not promising a colour that behaves in ways it does not, is part of the respect this material is owed. It is also how you tell a real promise from a pretty one.

Meeting a living material on its own terms

This move, finding out what actually carries a property before you build on it, is not only about this colour. It is a way of meeting any living material honestly, as the thing it is, rather than as a store of properties to extract.

It has a simple shape: a single question with three possible answers, that you can run in a few minutes with the AI tools you already use, and that keeps the extractive reflex from sending you the wrong way.

There is a reason to do this now in particular. Ask one of the AI models you already use about a living material and it will hand you the mechanism in seconds, often correctly. What it will not do is the judgement that comes after: whether the promise is really finished, and what you would build differently once you know.

More and more, the mechanism is the free part. Telling a grounded promise from a fluent one is the part that stays yours, and reading the material down is how you get there.

I have written this up as a short framework, with a full readout of this colour beside it: the cellulose spiral, the spacing that sets the colour, and the references to check it against. You will see one material read all the way down, then run the framework yourself on the next one you are drawn to.

The framework in brief: one question with three answers, run on any living material before you build on it.

A page from the readout: the marble berry taken apart to the cellulose spiral that carries its colour, with the sources laid out to follow.

Members' library opens next week

That framework is the first piece of a paid members' library, and the library opens next week.

The library is where these worked readings live. Each one takes a single living material apart to the level that actually carries its property, with the sources laid out so you can follow every step and read the next material yourself. This first framework, and the full readout of the marble berry beside it, are ready now. I will share how to join, and what it costs, when it opens.

If you would like first access when it opens, reply to this issue and say so, and I will make sure you are in.

That is what From the Molecule Up is for. It builds the thin layer between a biological word and the thing that actually carries it, so that the care we bring to living materials can rest on what the material is really doing, and not only on what we hope it means. It asks no one to become a biologist. Each issue reads one living material this way. I am writing it in the open, and you are reading it before it is a book.

Before you go, one question. If a living material has caught your eye for a property you cannot quite explain, a colour, a strength, a way it seems to sense or repair itself, reply and tell me what it is. I will read the sharpest of them this same way in a future issue. The same reply is all it takes to claim your early access to the library.

Until next time,

Raphael

P.S. For those of you teaching: this runs well as a short studio exercise. Give students one living-material image or product page and a single question: is this colour a dye the thing contains, or something the structure is doing? Most will reach for a dye first. The exercise is the moment they stop treating the material as a container of properties, and start asking what it is actually doing, before they design with it.

References

1. Vignolini, S. et al. (2012). Pointillist structural color in Pollia fruit. Proceedings of the National Academy of Sciences 109(39), 15712–15715. https://doi.org/10.1073/pnas.1210105109

2. Droguet, B. E. et al. (2022). Large-scale fabrication of structurally coloured cellulose nanocrystal films and effect pigments. Nature Materials 21, 352–358. https://doi.org/10.1038/s41563-021-01135-8

The full citation set, including the helicoidal cell-wall mechanism and the humidity-responsive film work, sits in the members' asset.

Dear {{first_name | reader}},

Most of us did not come to living materials because we wanted a more efficient production system.

We came because biodesign seemed to offer another way of being with the living world. A partnership instead of extraction. A material you tend, feed, wait for, and care for. A thing with a life of its own, not a passive substance waiting to be used.

That instinct is one of the field’s strongest commitments, and I share it.

So when someone says that designers need to understand living materials at the molecular level, it is fair if something in you tightens.

Gif by therainbowbridge on Giphy

The phrase can sound like an old habit in new clothes: take the living thing apart, turn it into a mechanism, and make it easier to control.

Val Plumwood gave that habit a clear name. She called it the master model: a way of relating to the living world from above it. It depends on two quiet moves.

One is backgrounding, where you depend on the living world but treat it as scenery. The other is hyperseparation, where you make the living other seem so distant from you that you never have to meet it on its own terms.

Concept of backgrounding and hyperseparation in human-nature relationship

If molecular reading meant that, the suspicion would be right.

But I want to make a different claim.

Honouring requires knowing what it does

Close attention to what a living material actually does, including the molecular work when the material acts at that scale, is not a betrayal of the ethics many of us already hold.

It is one way those ethics stay honest.

This is not a correction of the projects that already put organisms, growth, care, and cohabitation at the centre. Those projects have made a rare and important turn. They face the organism when much of design still looks past it.

What I want to add is one step inside that same turn.

The material is alive? Good.

Now what kind of life is it living? What does it need? What does it do in return? Which part of its living process makes the material bind, sense, change colour, soften, stiffen, clean, decay, or heal?

A warm word left on its own can leave the meeting unfinished. The point is not to cool the relationship down. The point is to follow the care all the way into the behaviour of the material itself.

Noticing has a scale

This field already values attention. Anna Tsing calls this the “arts of noticing”: the practiced ability to pay attention to what is happening around you, especially in damaged, mixed, and more-than-human worlds.

Noticing, in this sense, is not a mood or an aesthetic preference. It is work. You get better at it by staying with a thing long enough to see how it lives, changes, responds, and depends on others.

That matters deeply to biodesign because a living material is never only a sample on a table. It feeds, grows, dries, resists, contaminates, adapts, and dies. It asks for conditions. It gives some behaviours and refuses others. To design with it, you have to notice those relations. You have to learn what the material is doing, not just what it appears to be.

The question is scale. At one scale, you might notice that a mycelium composite grows faster in one substrate than another. You might see that the edges colonize before the centre, that moisture changes the surface, that drying alters the feel, or that pressing changes the stiffness. This is already good noticing. It gives the student a real material, not an idea of one.

But there is another question inside it. What is doing the binding? What part of the fungus meets the surface? Which structures help it grip? At that scale, the answer is no longer simply that the material is alive. The hyphae grow into and around the substrate, while cell-wall polymers and surface proteins help create adhesion. That does not make the relationship colder. It makes the attention more exact.

Mycelial film formed by the fungus Xylaria polymorpha (Xylariaceae, Ascomycota) in axenic culture. Image was taken with a scanning electron microscope. Image: Alisa Atamanchuk. CC BY-SA 4.0 via Wikimedia Commons.

This is why words like alive, collaborative, responsive, or self-healing can be true and still not be enough. They are doorway words. They invite the meeting, but they do not complete it. Molecular attention is one way to complete the meeting.

The field has already mapped the larger scale carefully. Elvin Karana, Bahareh Barati, and Elisa Giaccardi describe livingness as a material quality in its own right: something an artefact can carry into everyday use as it senses, grows, adapts, and eventually dies. Their account follows the organism into the conditions that keep it alive.

What I am describing sits one scale inside that. Livingness tells you how the organism stays alive in the artefact. Molecular reading asks what that particular life is doing at the point where the material behaves. The two are not rivals. One holds the larger relation. The other looks inside that relation for the behaviour that matters to the design.

This is also how I read Adela Orcajada and colleagues’ idea of biodesign literacy. Biodesign literacy joins biological knowledge with design’s own ways of knowing. Molecular literacy is one scale inside that broader literacy. It does not compete with it. It helps biodesign literacy become more precise where the material demands precision.

Naming as attention, not domination 

The worry is real because there are different ways of knowing. Donna Haraway gave one of them a sharp name: the “god trick,” or the view from nowhere. This is the fantasy of seeing everything from no body, no place, and no responsibility. It is knowledge that pretends to be complete so it can speak as if it owns the world. That is the kind of knowing the ecofeminist critique was right to distrust.

Haraway offered another kind of knowing in its place: situated knowledge. In plain words, this means knowing from somewhere. You see from a body, a place, a method, a history, and a set of limits. You say what you can see from there, and you do not pretend that your view is the whole.

Molecular attention, done honestly, belongs to this second kind. It does not say, “Now we know the living thing completely.” It says something more modest and more useful: under these conditions, with this material, this behaviour seems to depend on this process. The account is partial. It can be revised. It has to answer to the material again.

That is not the god trick. It is not the view from nowhere. It is a located account of what can be seen, tested, read, and questioned from where you stand.

There is also restraint built into this. You do not go infinitely deep. You go as deep as the promise in front of you requires. If the claim is that a material changes colour, you need to understand enough to know what causes the change and what conditions it depends on. If the claim is that it binds, you need to understand what does the binding. If the claim is that it cleans, you need to understand what is removed, transformed, trapped, or degraded, and under what conditions.

Then you stop.

Mastery wants the total account. Situated attention takes responsibility for the specific account. The problem was never depth by itself. The problem is the wish to own what you find there.

The point is not only to catch bad claims

There is a defensive reason this matters. The language of ecological design is easy to borrow. Regenerative, circular, carbon-negative, living, self-healing: an honest project and a weak claim can use the same vocabulary. So yes, molecular attention helps you ask whether a physical promise has physical support. It helps you see whether the claim rests on a material process or only on a persuasive metaphor.

But that is the smaller reason. The better reason is that attention gives you design decisions.

Ectomycorrhizal mycelium (white) associated with Picea glauca roots (brown). Photo: André-Ph. D. Picard. CC BY-SA 3.0, via Wikimedia Commons.

A mycelium composite grows with the right moisture and enough time. Its final stiffness depends on the substrate, the extent of colonisation, and what happens after growth, including drying and pressing. Left vague, this sounds like a list of constraints. Read closely, each point becomes something the designer can act on.

You can choose the substrate. You can stage the growth. You can design the drying process. You can press toward a target stiffness. You can build growth time into the form instead of treating time as an inconvenience. The fact stops being a warning and becomes part of the material palette.

That is what attention buys. It gives the designer something to work with.

This matters even more now because students increasingly meet biological claims through fluent explanations, generated summaries, diagrams, and predicted structures. A confident account of what a protein does, or a clean picture of a folded structure, can be useful. It can also be a good guess dressed as a fact.

The student is not necessarily engineering a protein. More often, they are reading a claim and deciding whether it means anything for a material, a studio brief, or a public promise. Teaching students to ask what does the work, under what conditions, and with what evidence is becoming part of the craft.

A studio card for Monday

I have turned the practical part of this letter into a one-page PDF you can use directly with students. It is called The Molecular Attention Studio Card, and it gives you four ways to teach students to read a living material closely with no lab, no budget, and no biologist on staff.

The moves are simple.

1. Watch before anyone designs.
   Give students the material before you give them a brief. Ask them to change one thing at a time: moisture, light, temperature, time. Then ask them to record what the material does in response.

2. Build the brief around what the material wants to do.
   Start from the behaviour students actually saw. Ask what that behaviour makes possible before asking what concept the material should serve.

3. Add one line to every crit.
   When a student says their material heals, senses, cleans, grows, or responds, ask the same three questions: what does the work, what conditions does it need, and what is the evidence?

4. Borrow one scientist for one hour.
   You do not need a biologist on staff. One visitor can change a studio if students already know how to ask better questions.

The card is meant to be used, not just read.

Pin it up in a studio. Keep it beside a crit. Send it to students before they start working with a living material.

Molecular Attention Studio Card.pdf

337.90 KB • PDF File

Download

If you want to go deeper, pair it with the Biodesign Promise Worksheet and the Molecular Behaviour Reference.

The Studio Card gives the teaching moves. The Worksheet gives students a page to fill in. The Reference gives them a short palette of what biological materials actually do.

Where this is heading

Each of these newsletters takes one idea about living materials and leaves behind something you can use. Over time, the set will run from the molecule up to the materials, organisms, and relations we design with. That is what From the Molecule Up is for. I am writing it in the open, a piece at a time. You are reading it before it becomes a book.

Same invitation as always.

If you teach with living materials and you have a studio move that helps this kind of attention land for students, write back and tell me.

Until next time,

Raphael

P.S. If you are not teaching anyone right now, run the first two moves on yourself. Watch the material before you design with it. Then let its real behavior shape the brief. That works alone at a bench as well as it does in a room full of students.

Notes and further reading

The argument here stands on a tradition the reader already knows. Everything below was verified to source in the 2026-07-06 fact-check, in the order the ideas appear.

On the ideas

• Anna Lowenhaupt Tsing, The Mushroom at the End of the World: On the Possibility of Life in Capitalist Ruins (Princeton: Princeton University Press, 2015). The “arts of noticing”, chapter 1, pp. 17-25.

• Val Plumwood, Feminism and the Mastery of Nature (London: Routledge, 1993). The master model, “backgrounding”, and “radical exclusion (hyperseparation)”, chapter 2, pp. 49-50.

• Elvin Karana, Bahareh Barati., & Elisa Giaccardi. (2020). Living artefacts: Conceptualizing livingness as a material quality in everyday artefacts. International Journal of Design, 14(3).

• Donna J. Haraway, “Situated Knowledges: The Science Question in Feminism and the Privilege of Partial Perspective,” Feminist Studies 14, no. 3 (1988): 575-599, https://doi.org/10.2307/3178066. The “god trick” (pp. 581-583) and situated, partial knowledge as the more rigorous route to objectivity (pp. 583, 590).

• Donna J. Haraway, Staying with the Trouble: Making Kin in the Chthulucene (Durham: Duke University Press, 2016). On staying with a thing rather than looking away.

• María Puig de la Bellacasa, Matters of Care: Speculative Ethics in More Than Human Worlds (Minneapolis: University of Minnesota Press, 2017). Care as situated, implicated involvement attentive to the response of the touched, pp. 119-121 and 165-173.

• Oron Catts and Ionat Zurr, “Growing Semi-Living Sculptures: The Tissue Culture & Art Project,” Leonardo 35, no. 4 (2002): 365-370; and “Semi-Living Art,” in Tactical Biopolitics, ed. Beatriz da Costa and Kavita Philip (Cambridge, MA: MIT Press, 2007), 232-246. The Tissue Culture and Art Project from 1996 and SymbioticA from 2000.

• Anthony Dunne and Fiona Raby, Speculative Everything: Design, Fiction, and Social Dreaming (Cambridge, MA: MIT Press, 2013). Speculation as a public “what if”, chapter 1, pp. 2-6.

• Adela Orcajada, Isabel Ordoñez, and Valentina Rognoli, “In Search of the Definitions for Biodesign: Practice, Identity and Biodesign Literacy,” DRS2026, Edinburgh, 8-12 June 2026, https://doi.org/10.21606/drs.2026.1411. Defines biodesign literacy as a framework combining biological knowledge and design ways of knowing.

• On more-than-human design, for the “design with rather than over” line: Elisa Giaccardi and Johan Redström, “Technology and More-Than-Human Design,” Design Issues 36, no. 4 (2020): 33-44; and Laura Forlano, “Posthumanism and Design,” She Ji 3, no. 1 (2017): 16-29.

On the science

• Mycelium-composite properties and fabrication: Freek V. W. Appels et al., “Fabrication Factors Influencing Mechanical, Moisture- and Water-Related Properties of Mycelium-Based Composites,” Materials & Design 161 (2019): 64-71, https://doi.org/10.1016/j.matdes.2018.11.027.

• Sarah Schyck, Mark Ablonczy, Sourav Patranabish, and Kunal Masania, “Shaping of Biohybrid Functional Living Materials,” Advanced Functional Materials (2026), https://doi.org/10.1002/adfm.202530836. Living mycelium printed, then functionalised by its own growth, with particle size and carbon loading as tunable variables.

On the words we use

• The greenwash beat: European Commission, “Green Claims” and Directive (EU) 2024/825; and the U.S. Federal Trade Commission, “Green Guides” (16 C.F.R. Part 260), both aimed at environmental claims that mislead.

Dear {{first_name | reader}},

You have seen "five times stronger than steel" often enough to stop trusting it. Most of us have. When a material turns up with a number that good, some quiet part of the design mind files it under marketing and moves on.

That instinct is right, and for a designer who has to build with the material it is only the first step. "Stronger than steel" is not a lie. It is true, and it is still not the thing you need to know about the fibre. Distrust keeps you from believing the number. It does not tell you what to write in the brief instead.

So you go one step further than the doubt and ask the critical question: what is this number actually measuring, and is it the property my design needs? That one step is the whole of this week's letter.

The material is worth the trouble. A new kind of fibre has arrived in the studio, brewed rather than drilled. Companies grow a silk protein in steel tanks by fermentation, from plants and sugar rather than from oil, and spin it into yarn.

500-liter stainless-steel industrial bioreactor together with a tangential flow filtration (TFF) system. Photo: CC1984USA. CC BY-SA 4.0 via Wikimedia Commons.

Spiber, in Japan, is the best known, and more than forty fashion brands have put its Brewed Protein fibre into real clothing.

Spiber frames it not only as bio-based but as regenerative: the feedstock grown under regenerative farming, and the worn garment meant to return through a circular biosphere loop and feed the next batch. For a designer who wants materials that come from biology instead of petrochemistry, and that give back rather than just take less, this is the right ambition.

So it is worth being able to say precisely what the fibre does, rather than waving its promises away or holding them up.

Your suspicion knows something is off. It does not yet know which thing.

When you distrust "stronger than steel," you are sensing that the sentence is doing more work than it should. You are right. But "I do not buy it" and "here is the property I would specify instead" are a long way apart, and the distance between them is made of a few plain questions.

Start with the word. "Strong" is at least four different properties. There is strength, how hard you can pull before the fibre breaks. There is stiffness, how much it resists stretching at all. There is extensibility, how far it stretches before it gives. And there is toughness, the total energy it can soak up before it fails. Steel and silk trade places depending on which one you mean.

Biaxial mechanical testing machine. It is used to pull, push, bend, or fatigue-test a material sample from multiple directions at once. Photo: Kolossos. CC BY-SA 3.0, via Wikimedia Commons.

"Five times stronger than steel" is true only by weight. Gram for gram, dragline silk beats steel, because it is so much lighter. On a given cross section, in absolute force, good steel wins: dragline silk reaches roughly 1 to 1.6 gigapascals, and plenty of steel goes higher. The number is a strength-per-weight figure, which matters enormously if you are saving weight and not at all if you are not.

And strength is not even where silk is remarkable. Its real gift is toughness. It stretches by a third to a half of its length and absorbs around three times the energy of Kevlar before it snaps. A fibre you choose because it is "strong" and a fibre you choose because it is "tough" behave differently in a seam, a joint, or anything that takes a sudden load. The famous line names the least interesting of the fibre's strengths.

The strength lives in a structure, not in the word "silk"

This is the part your skepticism cannot reach on its own, and it is where the molecule earns its place. In real spider dragline, the protein has two kinds of stretch along its length. Alanine-rich segments fold into tiny stiff stacks, called beta-sheet nanocrystals, which are simply hard little blocks packed inside the fibre. Those blocks give the strength. Glycine-rich segments stay loose and springy, and those give the stretch and the toughness. The performance is the balance between the hard blocks and the springy regions, and it is set partly at the moment the spider draws the thread out of itself.

A brewed fibre does not arrive with that arrangement for free. Spiber's Brewed Protein is inspired by silk, not a copy of dragline, and in much of its clothing it behaves closer to soft wool than to a spider's safety line. There is a reason. The spider's full protein is enormous and highly repetitive, which is exactly the kind of protein that is hard to brew in a tank, so the made versions tend to be shorter and need careful spinning and drawing to come anywhere near the natural fibre. The strength in the headline lives in a specific sequence and a spinning step, not in the fact that the protein came from a vat.

Ask an AI model whether the fibre is stronger than steel and it will tell you yes, fluently, with the same five-times figure, because that is what the text it learned from says. It will sound far more certain than you do. That is the part worth naming: the confident answer is the marketing number with the hesitation removed, and your hesitation was the more accurate instrument.

The same move works on "biodegradable"

The second promise on the label behaves the same way, and you can feel the same doubt about it. "Biodegradable" is not finished until you ask under what conditions. Spiber's strongest figure is a seawater test, close to full breakdown in six months by a named international method. Breakdown in soil is a different number, and home composting is different again. The single word, with no condition attached, is not yet a fact you can design around. It is a promise waiting for its conditions, and the question that finishes it is "under what conditions, measured how."

"Regenerative" sits in the same place. It is a real and serious promise, but it lives in how the feedstock is farmed and whether the take-back loop actually closes, not in the thread you are holding. The fibre is the end of that story, not the proof of it. The question that finishes it is "regenerative where, and has the loop been closed yet, or is it a plan."

From a gut “no” to a written spec

Here is the move laid out, so you can run it on the next "stronger than," "tougher than," or "lasts longer than" line you meet. Six plain questions take you from distrust to a specification:

1. Which property is it really about: strength, stiffness, stretch, or toughness?

2. In what units, and measured how?

3. Per weight, or in absolute terms? (This is the steel question.)

4. Against which baseline, and in what state, wet or dry?

5. Under what conditions does it hold? (Silk weakens as it takes up water.)

6. Does the fibre actually in your hands match the one in the comparison, or is it a softer cousin that borrowed the line?

Run those on "five times stronger than steel" and your gut no turns into something you can write down. The number is real. It is a by-weight strength figure. The property you actually want is toughness. It was measured dry. And the brewed version in your fabric may or may not be the dragline the number came from. None of this makes the fibre worse. It makes it specifiable, which is what you needed all along.

The worked example is yours to keep

I have written this whole reading up as a worked example: the Spiber promise, taken from gut suspicion to written specification, one question at a time, with the verdict shown at each step. It is something you can keep beside you and teach from, and rerun on the next material that lands on your desk. You can download it below:

The_Strength_Promise_Teardown_v1_brand-v3.0.pdf

168.56 KB • PDF File

Download

It sits beside the earlier pieces. The Biodesign Promise Worksheet is for taking a promise apart. The Molecular Behaviour Reference is for naming what does the work. The Prediction Critique Map is for the moment a confident image tries to pass as proof. This one is for the moment a strong-sounding word stands in for a measured property. If you are new here, start with the Worksheet.

That is what From the Molecule Up is for. It builds the thin layer between the marketing word and the measured property, the layer that living-materials design still tends to skip, without asking anyone to become a materials scientist. Each issue takes one biodesign promise - defined as “the sentence about what a living material does that a design project has come to rely on”. It then takes that promise apart this way and adds one more piece to that layer. I am writing it in the open, and you are reading it before it is a book.

Before you go, one question. If a "stronger than," "tougher than," or "lasts longer than" line has stayed with you, from a press release, a product page, or a studio brief, reply and tell me what it is. I will run the sharpest of them through this same reading in a future issue.

Until next time,

Raphael

P.S. For those of you teaching: the six questions run well as a short studio exercise. Hand students one real material brand page, give them fifteen minutes to take its strongest "stronger than" line apart, and have them land on which property it is really about. They already distrust the line. The exercise gives them the words for why.

A biodesign promise is the sentence about what a living material does that a design project has come to rely on. "This mycelium heals its own cracks." "This algae cleans the air." It says what the biology does, and it usually sounds settled and obvious. Most of the time you did not sit down and choose it. It formed while you were writing, or you picked it up from a paper, an exhibition, or a supervisor, and it has sat inside the project as an assumption ever since.

The promise matters because everything downstream leans on it: the wall text, the decision to fabricate, the answer you give in the crit. So before you build on a promise, it is worth taking it apart to see whether it holds.

Why a promise can mislead you

A promise names an effect and hides the behaviour underneath it. And one word can hide several completely different behaviours.

Take "mycelium self-heals." Slow it down and it could mean any of these:

• the living fungus grows back across the damage

• two cut surfaces fuse together again

• the surface just closes over and looks mended

• the material swells when it gets wet, so the gap pinches shut

• a person feeds it or patches it by hand

That is five different things wearing one word. Each needs a different test, a different collaborator, and a different maintenance story. Each lets you say something completely different in public. A designer who says "self-heals" without knowing which one they mean has not yet decided what they are making.

How to test a biodesign promise

You take the loose promise and walk it down into something specific you can check. You do not need a lab or a biology degree. You need the promise your project has come to rest on, and the willingness to ask the next question.

First, find the promise the project is already carrying: the sentence about what the material does that the work has come to lean on. Then, if it hides more than one behaviour, split it and pick one. Then ask plain questions about that one thing:

• What does it do, exactly?

• What part of the living thing does the work? A whole organism, a cell, a protein, a molecule?

• What sets it off? What has to be present for the behaviour to start?

• What would you see if it worked?

• How would you know it is real? Does that evidence exist yet?

• When would it break?

By the end, the promise lands in one of three places. It is sharp and shown, so you can build on it. It is sharp but unproven, so you test it before you commit. Or it cannot be pinned down, which is not a failure: it is the exact list of what you would have to find out or invent to make the idea real.

One rule to keep you honest

Go as deep as the promise needs, and no deeper. "The crack looks closed" is a question about a surface. "It senses lead in the water" is a question about molecules. You stop at the level where the real answer lives.

This is also not hostile to speculative or provocative work. If a project means to provoke rather than function, an unproven promise can be a deliberate choice. Testing the promise just makes sure it is a choice you are making on purpose, not one you have mistaken for a working material.

The tools that do this

Biodesign Academy makes two free instruments for working a promise down.

The Biodesign Promise Worksheet is the fill-in tool. You bring a promise, split it, run it down the six questions, and see where it lands. The Molecular Behaviour Reference is the companion you keep beside it: a short reference for the second question, naming the small set of things a protein can actually do (bind, sense, signal, switch, cut, transform, assemble, stabilise, adhere, transport, move, produce colour or light) and helping you tell whether the protein is the thing you are making or just the machine that makes it.

Where the idea fits

The biodesign promise is part of the vocabulary behind From the Molecule Up, the design-education project and forthcoming book from Biodesign Academy. Taking a promise apart is molecular reasoning in practice: tracing a design intention down to the biology it depends on. Molecular design literacy is the underlying ability; better design judgement is what grows from doing it over time.

Dear {{first_name | reader}},

When someone hands you a grown material and tells you what it does, they are making a promise. "Strong, like leather." "It self-assembles." Those words are not yet facts about the material.

They become facts only when you can name three things: what inside the material does the work, what that thing needs in order to do it, and how you would know it is real. Until then, you are holding a word, not a property.

So here is the point of this issue, stated plainly. Before you build on a promise, find what actually does the work. This week I want to show you how, using a single sheet you can hold. The surprise is where the work turns out to live.

Photo by Tim-Oliver Metz on Unsplash

The sheet, and two promises

You are holding a sheet of bacterial cellulose: a thin, tough film that bacteria grow on the surface of their liquid, layer by layer, over several days. It sits behind a lot of "grown leather" and "living textile" work.

You have been told two things about it. That it is "strong, like leather." And that it "self-assembles." The sheet feels solid in the hand, so both promises seem settled already. Let us check them.

Nothing here assembles by itself

"Self-assembles" sounds like the sheet forms on its own. It does not. The bacteria build it, using a small group of proteins called the cellulose synthase complex (which just means "the cellulose-building machine").

Two of those proteins, BcsA and BcsB, take sugar units one at a time, join them into long chains, and push the chains out of the cell, where they line up and lock into fine threads. Stack enough threads and you have a sheet.

And the machine only runs on cue. It waits for a signal: a small molecule called c-di-GMP that works as the on-switch. No living, fed bacteria and no on-switch, no sheet. So "self-assembles" is true only in a loose way.

The word "self" is the part that misleads. What does the work is the cellulose-building machine, and it needs a living culture, the on-switch, a surface to grow on, and time.

“Self-assembled” cellulose nanocrystals. Image: Ahmad Hivechi. CC BY-SA 4.0 via Wikimedia Commons

The strength is not in any protein

Now the second promise, "strong, like leather." Look for the protein that makes the sheet strong and you will not find one. There is none left in the dried sheet. The strength comes from the structure the bacteria laid down: how straight the threads are, how tightly they pack, and how neatly they line up as the sheet dries. (Scientists call that tight, regular packing crystallinity.)

Pack them well and dry them well, and the sheet is strong. Leave it wet, or dry it carelessly, and the same material is weak and shrinks.

This is the part worth keeping. The strength was never in the word "grown." It was in the structure. The protein built that structure and then left. So the promise became real not inside a protein, but in the ordered material the protein left behind.

When you go looking for what does the work, it is not always something still in your hand.

Transformed E.coli expressing green fluorescent protein. Photo: DanceWithNyanko. CC BY-SA 4.0 via Wikimedia Commons.

The other way round: when the protein is the product

One more case, because it flips this one over. A recent project from a team at Northeastern, Living Textiles, embroiders living cells into fabric and asks whether a textile can be alive. Underneath the question is a clear mechanism.

The cells are built to act as sensors. Each one carries added instructions, so that when it meets a certain signal nearby, it notices the signal and shows it through a visible change in the fabric. The fabric is not alive. The cells are, and they sense and respond on cue.

Here the thing that does the work is the sensing machinery, and it is a protein sitting right there in the finished material. It is the product, not the maker. And it comes with conditions the makers put front and centre: the cells need water, food, and warmth, or they die. "Alive" is a promise that comes with care.

So, two materials and the same step. In the cellulose sheet, the protein is the maker, and the truth lives in the structure it leaves behind. In the living fabric, the protein is the product, and the truth lives in the protein itself. Either way, you find what does the work before you trust the word.

Why this matters now

You used to be able to leave this to someone else. The biology stayed in a lab, or inside a metaphor. Not any more. The grown sheet is in your own hands, and the same AI tools that help you design with it will describe it in smooth, confident language that sounds settled before anyone has named what does the work.

The words arrive polished. The thing behind them does not arrive at all, unless you go and find it.

A short reference to help

So here is the piece I am adding this week, a short reference called The Molecular Behaviour Reference. You do not fill it in. This reference sits under the tool from the first issue, for the moment when you ask which part of the living thing does the work.

It holds three things: a short list of the jobs a protein can do, one test for where to look for proof (is the protein the maker, or the product?), and a few worked examples, including this issue's sheet. Read it once and the next promise gets quicker to sort out.

The tool it sits under is the Biodesign Promise Worksheet, from the first issue. That is the part you fill in. You take one promise, separate out the smaller promises hiding inside it, and for each one name what does the work, what it needs, and how you would know it is real. If you have not used it yet, start there: the worksheet is where you do the work, and the reference tells you where to look.

You can download the worksheet here:

And you can read the new reference here:

Where this is going

This is the second issue, and the reference is the second piece of something I am building in the open. Each issue takes one promise about a living material apart and adds a piece, until the whole thing runs from the molecule up to the materials we make from it. That is what From the Molecule Up is for, and I am writing it as we go. You are reading it before it becomes a book.

Same invitation as last week. If a promise has stuck with you, one that sounds solid until you slow down and ask what carries it out, reply and tell me. I will follow the sharpest ones down in a future issue.

Until next time,

Raphael

P.S. For anyone teaching: this works as a ten-minute studio exercise. Give students one promise from a brief, have them name the job the protein does and decide whether it is the maker or the product, and the discussion about where the promise really lives tends to run itself.

Biodesign literacy is the mix of biological knowledge and design ways of knowing a person needs to work well with living systems. The term was set out by Adele Orcajada, Isabel Ordoñez and Valentina Rognoli in a 2026 paper at the Design Research Society conference (Orcajada et al., 2026, doi.org/10.21606/drs.2026.1411).

Their argument is plain. Biodesign has grown for two decades without a shared language. Designers and scientists use the same words to mean different things, and there's no agreed definition of who counts as a biodesigner. So the field needs a shared literacy, a way to move between biological and design ways of knowing.

Why it matters

The paper backs this with a small survey of 38 practitioners. About 81% hold formal qualifications in the arts and humanities, and most pick up biology on their own, through their own experiments (89%) and books (68%). Few are formally trained in biology.

Almost all of them report trouble collaborating across disciplines. The barriers they name most often are about language: 61% point to differences in disciplinary language, and 65% see a significant gap in the terminology used across fields. The diagnosis is hard to argue with.

Where it gets thinnest

Here's what I'd add. A literacy matters most where it's hardest to build, and in biodesign that's the molecular scale. By molecular I mean the level of proteins and the machinery inside cells, the layer AI tools now let designers work with directly.

The molecular end is also where claims get hardest to check. Trusting a molecular result is procedural work: it depends on how the result is documented and read over time, and on who is accountable for it. That's the subject of a recent framework I built on molecular trust in AI-era bio-digital systems (Kim, Heinis & Pschetz, 2026).

The survey hints at this without naming it. Territorial knowledge was the main motivation for under 3% of respondents, and 80% of them were based in Europe. The field's attention sits at the visible end: mycelium grown into a vessel, bacteria that colour a textile. The molecular end, where many of the real design decisions in AI-era biology are now made, stays thinly mapped.

How it relates to molecular design literacy

This is the gap I work in, and I have a name for the literacy it needs: molecular design literacy. It's the same literacy taken down to the scale where a protein is the unit of design. It belongs inside biodesign literacy, at the depth the broad view tends to skip.

Biodesign literacy names the gap. Molecular design literacy is what that literacy looks like once you take the molecule seriously.

Common questions

What is biodesign literacy? The mix of biological knowledge and design ways of knowing needed to work well with living systems, and to move between biological and design vocabularies and methods. The term was proposed by Orcajada, Ordoñez and Rognoli at DRS 2026.

Who proposed the term biodesign literacy? Adele Orcajada, Isabel Ordoñez and Valentina Rognoli, in a 2026 Design Research Society conference paper (doi.org/10.21606/drs.2026.1411), building on earlier work on biological literacy.

How is biodesign literacy different from molecular design literacy? Biodesign literacy is the broad capacity across the whole field. Molecular design literacy is that same literacy at the molecular scale, where a protein is the unit of design.

References

• Orcajada, A., Ordoñez, I., & Rognoli, V. (2026). In search of the definitions for biodesign: Practice, identity and biodesign literacy. DRS2026. doi.org/10.21606/drs.2026.1411

• Kim, R., Heinis, T., & Pschetz, L. (2026). Trusted Molecular Memory: Investigating Molecular Trust in Hybrid Bio-Digital Systems. doi.org/10.5281/zenodo.19002474

Molecular design literacy is a designer's ability to understand, question, and shape a biological idea at the level of the molecules that actually do the work. It means being able to ask, of any biological promise, what is doing this, under what conditions, with what evidence, and where does it break.

You do not need to become a protein engineer to have it. You need enough biology to know where a promise lives, so you can tell the difference between an idea that is finished and one that only looks finished.

Why it matters now

AI changed the stakes. Tools can now generate a protein structure, an image, or a confident explanation in seconds. They look like answers. Often they are guesses dressed as answers. A designer who cannot read the molecular layer has no way to tell which is which, and ends up trusting a picture because it is convincing.

At the same time, living materials like mycelium, algae, and bacterial cellulose are now in the hands of students and studios. So designers are making real biological claims, in public, more than ever. Molecular design literacy is what keeps those claims honest.

A short example

Take a common claim: "this mycelium material self-heals."

Self-heals how? Does it regrow? Re-bind where it was torn? Recover when you add water? Repair its structure under load? These are four different biological behaviours, with four different conditions and four different kinds of evidence. The phrase "self-heals" hides all of them.

Molecular design literacy is what lets a designer stop at that sentence and ask the next question, instead of putting it on a wall label and moving on.

What it is not

It is not a demand that every designer learn biochemistry. The rule is simple: go as deep as the claim requires, and no shallower than the promise demands. A small claim needs a small amount of biology. A bold claim needs more.

It is also not hostile to speculative, ecological, or more-than-human design. Those traditions are valuable. Molecular design literacy gives them a floor, so that a beautiful idea is also an accountable one.

Where the term comes from

Molecular design literacy is the core idea behind From the Molecule Up, a design-education project and forthcoming book from Biodesign Academy. It sits at the centre of a wider vocabulary: molecular reasoning is the practice of doing it, and better design judgement is what a designer gains from it over time.

If you found this page useful, the weekly Biodesign Academy newsletter is where this work develops, piece by piece. You can subscribe here.

The concept is not the design

Dear {{first_name | reader}},

Biodesign is full of promises. In particular, a biodesign promise which is defined as “the sentence about what a living material does that a design project has come to rely on” . Here is where one of them meets the bench.

You have grown a slab of mycelium over a week or two, and it is finally in your hands. The whole project was built around a single promise: this material "self-heals." That promise carried you through the sketch, the precedents, the search for a lab that had grown something close, and the crit.

Now the material is in front of you, and you are testing it.

You score a crack into the surface, mist it, feed it, keep it warm, and wait to see whether the gap closes on its own. This is the moment the idea has to become a real material doing a real thing, and something is missing.

Ectomycorrhizal mycelium (white) associated with Picea glauca roots (brown). Photo: André-Ph. D. Picard. CC BY-SA 3.0, via Wikimedia Commons.

The word that first caught you named an effect. It never named the behaviour underneath that effect. And here at the bench, with the material in your hands, that behaviour is exactly what you now need.

Does the living fungus actually regrow across the damage? Or does the surface only close over and look mended? The word will not tell you. The material might, if you know what to ask it.

So let me show you what I mean, with the word that started all this.

“Self-healing concrete crack healing”. Images credit: Niranjan Prabhu K. CC BY-SA 4.0 via Wikimedia Commons

One word, several different promises

Take "self-heals." It sounds like a single property. It is really several different promises sharing one word.

It might mean the living fungus regrows across the damage. It might mean two living surfaces fuse back into one. It might mean the surface closes over and looks mended, or that the material swells when it gets wet so the gap pinches shut on its own. It might mean that a person has to feed it nutrients or pack in fresh mycelium by hand.

These are not shades of a single idea. Each one calls for a different test, a different collaborator and a different maintenance story. Each one also licenses a very different sentence in public.

The underlying science can be genuine. In one study of a pure mycelium leather, fungal cells that had gone dormant during processing revived when researchers fed them a nutrient broth. They grew back over punched holes and recovered most of the original strength at the wound.

That is real self-healing. It is worth being precise about what made it possible: living cells survived the making of the material, and were then given the moisture, food and time to resume what fungus does.

It is worth setting that beside a building-material story from last year that travelled a good deal further. The headline described a living fungal material that "repairs itself for over a month." The press release underneath was clear that self-healing had not actually been tested. What lasted a month was viability, which means the cells stayed alive longer than usual.

That is a real and useful result, but it is not repair. Somewhere between the laboratory and the coverage, a month of staying alive quietly became a month of healing.

Same word. Two completely different promises. Only one of them would let you honestly promise repair.

Why the gap suddenly matters

For most of biodesign's life this kind of slippage was survivable, because the biology stayed at arm's length: a metaphor in a manifesto, or a result in someone else's lab, and the designer rarely had to put a number on what a material actually did.

That distance has closed.

Biofabrication has put the living material directly in your hands, so the person making the promise is now also the person who has to grow the thing and stand behind what it does.

At the same time, AI tools answer biological questions on demand, in a register that sounds more grounded than it is. The promise now leaps straight from the organism to the public version, with nothing in between to say which behaviour is doing the work, or under what conditions.

Aspergillus structures in LPCB tease mount of culture showing conidia, sterigma, vesicle, conidiophores, foot cell, hyphae, and mycelium. Photo: Ajay Kumar Chaurasiya. CC BY-SA 4.0 via Wikimedia Commons.

How far down do you actually go?

None of this means you have to become a biochemist. It does not mean every promise has to be dragged down to chemistry. You go only as deep as the promise requires, and no shallower than it demands.

"The crack looks closed" is a question about a surface. "The material is alive" is a question about cells. Neither needs to be taken to the molecular level for its own sake.

This image shows a group of elongated cells (hyphae) from the filamentous fungus Podospora anserina. They are labeled with a fluorescent stain, JC-1, that labels areas of high metabolic activity (orange-red staining). It was taken by a high-resolution camera attached to a fluorescence microscope. The magnification is 630x. Photo: Christian Scheckhuber. CC BY-SA 4.0 via Wikimedia Commons.

But some promises reach that level faster than designers tend to expect. Take a material that "senses pollution and changes colour." That promise does not stay poetic for long.

Which molecule is being detected, and what binds it: a pigment, a protein, a receptor? What changes once it has bound, and how does that change become a colour the eye can actually see? How specific is it, and what might set it off by accident?

Protein Structure of NCAM1 (Neural Cell Adhesion Molecule 1). Image: Emw. CC BY-SA 3.0, via Wikimedia Commons.

Four questions in, and you are already working at the level of molecules, because that is where this particular promise turns out to live. The skill is not depth for its own sake. It is knowing where a promise lives, and going exactly that far.

With AI in the loop, that skill only sharpens. A predicted structure is not a proven function. A generated sequence is not a working protein. The literacy to tell a real result from a convincing depiction of one is quietly becoming the thing that matters most, and it is the one thing these tools cannot hand you along with their output.

The tool I am building

The practical instrument for this is the Biodesign Promise Worksheet. It begins where this email began, by taking the promise your project has come to lean on and laying out the different things it might actually mean, the way "self-heals" opened into several versions.

You choose one of those versions and put only that one to a short series of plain questions: what it does, exactly; what part of the living thing does the work; what sets it off; what you would see if it worked; how you would know it is real; and when it would break.

By the end, the version you started with has settled into one of three places. It is sharp and already shown, and you can build on it. It is sharp but not yet shown, and you now know what to test before you commit to it. Or it cannot be pinned down at all, which is not a failure. It becomes a brief, the list of what you would have to find out or invent to make it real.

The worksheet is free for subscribers. If you're already on the list, it's in the issue waiting for you. If you're not, subscribe here, and it's yours: we’ll include it in our welcome email.

That is what From the Molecule Up is for. It is a literacy project for designers and educators working with living materials, biofabrication, and AI biology. Its aim is to build the layer between metaphor and mechanism that living-materials biodesign still lacks, without asking anyone to become a biochemist. Enough, that is, to know where a biological promise actually lives, and what can honestly be promised once it has been found.

This email is where that gets built, in the open. This is the first issue. Each one will take a single biological promise apart the way we took "self-heals" apart today, and add one more piece to that missing layer, until the whole runs from the molecule up to the things we make from it.

I am writing the book as we go, and you are reading it before it is a book.

Before you go, I would like to ask you one thing. If a biological design promise has stayed with you, one that sounds wonderful but goes vague the moment you slow it down, reply and tell me what it is. A project, a press release, a tool, a studio brief, anything at all. I will work through the sharpest of them in a future issue.

Until next time,

Raphael

Dear {{first_name | reader}},

This month closes out an AI-driven protein robustness study focused on what happens to experimentally validated, AI-designed protein binders under structural stress.

What started as a technical investigation into mutation responses became something more fundamental: a design question about the difference between working once and actually being reliable.

Between experimental validation and reliability

The current standard of success in AI protein design is essentially: can we produce a protein that performs the intended function? Increasingly, yes. AI-designed proteins now bind to targets with high precision, behave as expected in lab experiments, and pass experimental validation. At that point, the design is usually considered done.

But this project began from a different assumption: that working once doesn't necessarily mean something is reliable. Most AI-driven protein design systems are optimised to answer that first question and stop there.

Structural predictions, confidence scores, and lab validation tell you that something works in a specific setup. They don't tell you how stable the structure is under change, where the weak points are, or whether the design can be trusted beyond that initial success.

The study took AI-designed binders that had already passed experimental validation and asked what happens when you push them, introducing small controlled changes rather than designing new proteins from scratch.

Three patterns came up consistently.

1. Working can hide fragility: some proteins look structurally clean, score well, and pass validation, but destabilise under small changes.

2. Reliability is uneven: within the same design system, some proteins remain stable under perturbation while others sit close to failure.

3. Failure is often invisible: in several cases the overall structure stays intact, but the functional region where binding happens fails. The artifact looks stable but has lost its intended behaviour.

One further finding worth noting: attempts to benchmark these proteins against natural equivalents often failed, not due to poor matching but because there simply isn't a clear natural equivalent.

AI-designed proteins may occupy regions of design space biology never explored, which means reliability has to be actively constructed through evaluation rather than inferred from nature.

What this looks like in practice: bacterial cellulose and living materials

Bacterial cellulose is one of the more compelling materials in biodesign right now: a nanofibrous scaffold produced by bacteria, mechanically strong, highly pure, and finding applications in wearable textiles, wound dressings, and flexible electronics.

Controlling its properties precisely remains a significant challenge, and this is where AI-designed proteins are starting to enter the picture. Researchers are exploring how engineered proteins can direct cellulose assembly, modify fibril surface chemistry, or interface the material with biological and synthetic components.

But bacterial cellulose fabrication is not a controlled laboratory environment. The bacteria are subject to temperature variation, nutrient fluctuation, and mechanical agitation across days or weeks of production.

A protein that functions reliably in an optimised lab setup may behave very differently inside a living production system. If that protein sits close to a structural failure threshold, as our study suggests many AI-designed proteins do, those fluctuations may be enough to push it over quietly, in ways that don't surface until much later in the pipeline.

The same logic extends to engineered silk, mycelium composites, and biofilm-based structures. Anywhere the designed biology has to perform reliably across a production process rather than a single validated experiment, structural robustness under variation is not a secondary concern. It is central to whether the design actually works.

Toward trust-aware biodesign

The paper formalises this into a post-success evaluation framework: a structured method for stress-testing protein designs and distinguishing hidden fragility from genuine robustness. In traditional design disciplines, we don't stop at "it works."

We ask how something behaves under stress, where it fails, whether it holds across conditions. In AI-driven biodesign, we're mostly still at "it works, move on." The argument here is for a different posture: it works, now evaluate whether it holds.

That shift matters as AI-designed proteins move closer to deployment in therapeutics, biosensors, molecular control systems, and living material fabrication. Treating experimental validation as the finish line made sense when getting proteins to work at all was the hard part. That's no longer the hard part.

The full paper is coming soon. If any of this connects with work you're doing in protein design, living materials, or the broader question of how we build trustworthy biological systems, I'd be glad to hear from it.

Until next time,

Raphael

What Happens After AI-Designed Proteins “Work”? A Framework for Reliability, Robustness, and Trust in Biodesign

Key Insight

AI-designed proteins can pass experimental validation and still fail under real-world conditions. True success in protein design is not whether a system works once—but whether it remains stable, functional, and reliable under variation.

Why “It Works” Is No Longer Enough in AI Protein Design

The current benchmark in AI-driven protein design is functional success:

• Does the protein bind its target?

• Does it behave as expected in a controlled experiment?

Increasingly, the answer is yes.

However, experimental validation only confirms performance under a specific set of conditions. It does not evaluate:

• Structural stability under perturbation

• Sensitivity to mutation or environmental change

• Reliability across time, scale, or production environments

This creates a critical gap between initial success and real-world reliability.

From Experimental Validation to Structural Reliability

This study began with a different assumption:
A protein that works once is not necessarily a protein that can be trusted.

Methodology Overview

• Selected AI-designed protein binders already experimentally validated

• Introduced small, controlled perturbations (e.g., mutations)

• Observed structural and functional responses

Rather than designing new proteins, the focus was on stress-testing existing designs.

Three Core Findings from Protein Stress Testing

1. Functional Success Can Mask Structural Fragility

• Proteins with high confidence scores and clean structures may still destabilize under minor changes

• Standard evaluation metrics often fail to detect hidden weaknesses

2. Reliability Is Uneven Within the Same Design System

• Some proteins remain stable across perturbations

• Others operate close to structural failure thresholds

This suggests that robustness is not guaranteed—even within high-performing AI pipelines.

3. Failure Is Often Invisible

• Structural integrity may appear intact

• Functional regions (e.g., binding interfaces) can fail silently

Result:
A protein may look correct but lose its intended behavior.

Why Natural Benchmarks Don’t Always Apply

Attempts to compare AI-designed proteins with natural equivalents revealed a limitation:

• Many designs do not map to known biological structures

• These proteins may occupy novel regions of design space

Implication

Reliability cannot be inferred from nature.
It must be explicitly engineered and evaluated.

Real-World Application: Bacterial Cellulose and Living Materials

What Is Bacterial Cellulose?

• A nanofibrous material produced by bacteria

• Known for strength, purity, and flexibility

Current Applications

• Wearable textiles

• Wound dressings

• Flexible electronics

The Role of AI-Designed Proteins

Engineered proteins are being explored to:

• Direct cellulose assembly

• Modify fibril surface chemistry

• Enable integration with biological and synthetic systems

The Reliability Problem in Living Systems

Unlike controlled lab environments, biological production systems introduce variability:

• Temperature fluctuations

• Nutrient variability

• Mechanical stress over time

Critical Risk

Proteins designed near failure thresholds may:

• Degrade under production conditions

• Lose function without visible structural failure

• Fail late in the development pipeline

Broader Relevance

This challenge extends to:

• Engineered silk

• Mycelium composites

• Biofilm-based materials

In all cases, robustness under variation determines success.

Toward a Post-Success Evaluation Framework

What Is a Trust-Aware Biodesign Approach?

A shift from:

• “It works → move on”

To:

• “It works → now test whether it holds”

Core Components of the Framework

EEAT in Practice: Experience, Expertise, and Evidence

First-Hand Research Experience

This framework is grounded in direct experimental analysis of AI-designed proteins under controlled perturbations, rather than theoretical modeling alone.

Alignment with Emerging Research

Recent advances in protein design (e.g., deep learning-based structure prediction and generative models) have significantly improved functional success rates. However:

• Studies show mutation sensitivity remains a major limitation

• Protein stability remains a key bottleneck in therapeutic and industrial deployment

Multiple Perspectives

• Engineering view: Success = function achieved

• Biological systems view: Success = function sustained under variability

• Design perspective: Success = reliability across contexts

Limitations and Open Questions

• Stress-testing increases evaluation complexity and cost

• No universal benchmark for robustness currently exists

• Trade-offs between performance and stability remain unresolved

Why This Matters for the Future of Biodesign

As AI-designed proteins move toward deployment in:

• Therapeutics

• Biosensors

• Molecular control systems

• Living materials

The definition of success must evolve.

Key Shift

• Past: Difficulty was making proteins work

• Present: Difficulty is ensuring they remain reliable

Summary

• Experimental validation confirms function—but not reliability

• AI-designed proteins often exhibit hidden fragility

• Robustness must be actively tested, not assumed

• Living systems amplify the consequences of instability

• A post-success evaluation framework is essential for trust in biodesign

FAQs About Protein Reliability in AI-Driven Design

What is the difference between validation and reliability?

Validation confirms a protein works under specific conditions. Reliability ensures it continues to work under variation, stress, and real-world environments.

Why do AI-designed proteins fail after initial success?

They may be structurally fragile or sensitive to small perturbations not captured during initial testing.

Can natural proteins be used as benchmarks?

Not always. Many AI-designed proteins exist in novel design spaces without natural equivalents.

What industries are most affected by protein reliability?

• Biotechnology and therapeutics

• Materials science (e.g., living materials)

• Synthetic biology and biofabrication

How can reliability be improved?

Through systematic stress testing, mutation analysis, and evaluation across environmental conditions.

Final Takeaway

AI has solved the problem of making proteins work.
The next challenge is ensuring they keep working.

Reliability is not a byproduct of design—it is a requirement that must be engineered, tested, and validated explicitly.