The essay
Why technology exists.
Before talking about startups, capital or risk, it is worth going one floor down. Technology is not an invention of economics: it is what a living system does when the world moves beneath it. This essay is the ground on which all the others rest.
14 min de lectura
A piece of universe that holds together.
A living being is a material system: a small, complex and fragile fragment of the universe, in constant exchange with the rest. It is not separate from what surrounds it. It takes matter and energy, orders them for a time and returns them. As long as that exchange stays within certain limits, the system goes on existing. When it moves outside them, it stops.
This is not a metaphor. Physiology described it a century and a half ago: Claude Bernard observed that complex organisms maintain a stable internal environment against an exterior that changes, and Walter Cannon called homeostasis that continuous work of holding the balance. A body is not a thing: it is a process that is maintaining itself, and that stops being maintained the moment it stops.
What a living being manufactures is itself.
There is a precise way of saying this, and it comes from two Chilean biologists. In Autopoiesis and Cognition, Humberto Maturana and Francisco Varela proposed distinguishing two types of system by what they produce.
An allopoietic machine produces something other than itself. A car factory produces cars, not factories. An engine produces movement. Its product is outside it, and the machine can stop without ceasing to be what it is.
An autopoietic machine continuously produces the components that produce it. Its product is its own organisation. It does not manufacture cars: it manufactures itself, without pause, and the moment it stops doing so it ceases to exist. That is a living being. A cell does not have an activity besides maintaining itself: maintaining itself is its whole activity.
The distinction looks abstract and in fact it orders everything that follows. Everything this site calls technology is allopoietic: things we manufacture that do something for us. We are the other. And the reason we manufacture those things is precisely that the other, the organisation that sustains itself, only sustains itself within very narrow limits.
There is no purpose, there is persistence.
It is tempting to say that a living being has a purpose, and that the purpose is to last. But the formulation is superfluous, and it also weakens the argument.
What exists is not here because it wanted to be. Matter aggregated spontaneously and by chance, and of all the possible combinations the ones still present are those that turned out able to maintain and copy themselves. In the picture described here there is no purpose: there is a property. Of whatever may lie beyond its edge, this essay says nothing, and later it will be clear why. What persists is what manages to persist, and the forms that did not manage it left nothing behind to tell of them.
Richard Dawkins took this idea to its end in The Selfish Gene, shifting the focus from the individual to what gets copied: what lasts is not the body, which comes apart in a generation, but the information that body carries and writes again. Dawkins himself warned, and had to repeat it for decades, that the adjective in the title is an accounting metaphor and not a trait of character: a gene wants nothing, because it is not the kind of thing that can want anything.
And here a caution is in order, one that is not biological but methodological. Ludwig Wittgenstein held that much of metaphysics' questions are not false: they are badly formed questions, born of a use of language that has slipped off its ground. Grammar lets us build what is the purpose of life? as easily as we build what is the purpose of a hammer?, and the symmetry of the two sentences makes us believe that both expect the same kind of answer. They do not. The second asks after the intention of whoever made the hammer; in the first there is nobody who made anything, and the question is left spinning in the void.
That is why this essay does not ask what a living being exists for. It asks what conditions it needs in order to go on existing, which is a question with an answer and moreover a testable one.
The narrow band.
The universe offers a range of physical conditions that is hard to imagine. In the core of a star the temperature is measured in millions of degrees and matter piles up at densities that have no equivalent here. In the space between galaxies there are a few atoms per cubic metre and the temperature grazes absolute zero. Everything else fits between those two extremes.
The life we know occupies a ridiculously narrow band of that range. The most resistant organisms described widen the margins a little: archaea that thrive near hydrothermal vents at more than a hundred degrees, bacteria that withstand salinity, pressure or radiation. But even they live in a minute slice of the possible.
And that band does not stay still. The conditions surrounding an organism change for physical reasons, such as climate, soil or the availability of water, and also for biological ones: the presence of other living beings that compete, prey, collaborate or simply occupy the place. The environment of a living being includes the other living beings, and that is why it never stops moving.
Two ways out, and only two.
When the mismatch appears, when the world moves outside the band, there are two ways of recovering the balance, and there is no third.
The animal in the photograph that opens this page solves the problem by the first route, and in the most literal way possible: an amphibian is, by definition, an animal of double life, inhabiting two different physical media with the same body. It manufactured nothing to achieve it. It took millions of years and nobody decided it.
Change the body. In the short term, by moving it: getting away from a danger, moving towards an opportunity, seeking shade, migrating. In the long term, by modifying it: that is what evolution by natural selection does, generation after generation, without any individual deciding anything.
Change the environment. Use matter and energy to manufacture a controlled space where conditions are back within the margin. A nest, a burrow, a beaver dam. Evolutionary biology calls this niche construction: organisms not only adapt to their environment, they also modify it, and that modification changes the selection pressures they and their descendants will face. Dawkins reached a similar place from another side with the extended phenotype: the dam is part of the beaver as much as its teeth, because it is a product of its genes and affects its survival. It should be said that the exact evolutionary weight of niche construction is still debated among specialists; what matters here is not the debate, but the observation that the second way out exists and has been in use since long before us.
The second way out, moreover, does not advance like the first. The economist W. Brian Arthur holds that radically new technologies do not appear by accumulating small variations on earlier ones, which is the Darwinian mechanism, but by combination of technologies that already exist; and that each new one becomes, in turn, an available piece for building the next. Hence every technology has a recursive structure: it is made of components that are themselves technologies. It is a form of evolution different from that of bodies, and much faster, because it does not wait for any generation: it is enough for the pieces to fit.
Cognition as a shortcut.
The first way out, in its long version, has a clock problem. Natural selection works on the scale of generations: it is a slow, blind mechanism, with no memory and no foresight. It works, but it does not arrive in time when the world changes fast.
And it is not only that it is slow: it is that it has arrived late more than once. The fossil record contains at least five episodes in which conditions changed faster than natural selection could respond, and in each of them most of the species then existing disappeared. Slowness is not a minor defect of the mechanism: it is the reason another one was needed.
All the adjustments that do not pass through that filter pass somewhere else: cognitive systems. A cognitive system, stripped of everything else, is a mechanism that does two things: it builds a representation of the environment and it prepares responses. Perceiving is manufacturing a model of the world; deciding is choosing a response within that model before the world sends the bill.
Maturana and Varela went one step further, and it is the step that closes the circle: for them cognition is not a faculty some living beings have in addition to living, but the very way in which an autopoietic system stays coupled to a changing environment. An organism knows to the extent that it goes on existing, and it goes on existing to the extent that it responds. Living and knowing, on that reading, are the same verb seen from two places.
It is worth pausing on the staircase we have climbed without naming it. First there is matter, which orders and disorders itself without anything noticing. Then, in a minute part of it, something appears that maintains itself: the living. And within the living, in an again minute part, something appears that also builds a model of what surrounds it: cognition. Each step happens inside the previous one and takes up far less room than it.
Whether that staircase has more steps is an open question, and it is best framed as a hypothesis and not as a forecast. Nothing requires cognition to go on needing a biological support: it could be that the next step were born not from a cell but from a machine, and that evolution by natural selection has been the road to get here, but not the only one possible from now on. We do not know. What can be asserted is more modest and more solid: until today, everything that knows has been alive.
There is the shortcut. Evolution takes generations; cognition fits in a lifetime, and sometimes in a second. An animal that sees the predator does not need its descendants to develop a defence: it steps aside. And a species able to pass on what it has learned from one head to another does not even need each individual to discover it alone.
It is the decisive difference, and not a small one. Stefan Klein calls the result of it the collective brain: the capacity to invent depends less on the talent of a few than on the exchange among many and on how open a society is. What one head knows stops dying with it, and each generation begins where the previous one ended instead of beginning from scratch.
Science and technology: prostheses and control.
Our species uses that capacity for both ways out at once, and systematically. Everything it manufactures is allopoietic, in the exact sense above: things that produce something other than themselves, and that produce it for us.
To change the body itself, with material and informational prostheses. A wheel is a leg that does not tire; a telephone is a memory and a voice that do not stop at a distance; an antibiotic is a defence our immune system did not bring from home. We do not modify the body genetically: we add pieces to it.
To change the environment, by manufacturing conditions. A house is a narrow band held up by hand. A pressurised suit is a piece of the Earth's surface taken where there is none. A hospital, an electricity grid and a cold chain are the same thing at another scale: machines for keeping the world within the margin.
Seen this way, technology stops being a market matter and becomes what it is: a mechanism that has appeared in the universe so that some of its parts can go on existing while conditions move. It is not a luxury of a civilisation. It is the shape persistence takes when it passes through a cognitive system.
From the solution to the world: the two phases.
A solution that reaches nobody has adapted nothing. This is where the argument lands and stops being abstract.
Innovation has two phases and needs both. The first is making it exist: research and development, the work that goes from the problem to something that works. The second is making it arrive: production, distribution, commercialisation, the work that goes from something that works to something in the hands of whoever needs it. A brilliant prototype that stays in the laboratory is, from the point of view of this essay, an adjustment that has not happened.
In a market, the organisational form that exists to do those two things is called a company. It is not the only one possible, and it does not always do it well, but it is the one there is: a structure that gathers knowledge, money, hands and relationships to build a solution and put it into circulation.
And here begins the problem the rest of this site deals with. Because making those two phases work at once requires gathering resources that are nothing like one another, and almost no initiative has them all.
Cognition is a confession.
It is worth looking squarely at what having a cognitive system means, because what it means is not flattering.
A cognitive system exists to build a representation of what is outside and prepare a response. Every one of those words presupposes something: that there is an outside, that the outside matters, and that it can do harm. Knowing is not an ornament of complex beings nor a proof of superiority. It is what a part does because it is a part. A system that were complete, depending on nothing external, would have nothing to represent and nothing to defend itself from.
Seen this way, intelligence is not the opposite of fragility: it is its consequence. Homeostasis maintains internal balance because the exterior threatens it. Autopoiesis manufactures itself because otherwise it comes apart. And cognition models the world because the world can kill it. All three are answers to the same fact: we are not the whole, we are a piece, and a piece can be lost.
In a very different field, formal logic found something similar. Kurt Gödel's incompleteness theorems showed that no formal system rich enough can account for itself entirely from within: there are always true statements that system cannot prove with its own means. It is not a proof of anything said here, and it should not be used as if it were: it is an analogy, taken from another terrain, that points in the same direction. What is inside does not manage to encompass itself completely.
This whole page is about that. The narrow band, the two ways out, the prostheses, the conditions manufactured by hand: they are the tools of something that knows it can cease to exist. Technology is a confession of limit made with materials.
What has no outside.
From this comes a thought that is not a thesis about the universe but a consequence of the previous definition, and it should be said carefully.
If a cognitive system exists to represent what is outside it, then the whole, by definition, needs none. It has no outside to model or to defend itself from. It is not that it does not think in the sense of being distracted or empty: it is that the word does not apply to it, just as it does not apply to ask what the purpose of life is. It is the same caution as before, applied to the other end of the scale. Cognition is a matter for parts.
And yet the whole does come to know itself, though only in pieces and from within. It does so through the minute parts that have developed the capacity to look at the rest: living beings, and among them those that also ask. Carl Sagan said it better than anyone in writing that we are a way for the cosmos to know itself. What this essay adds is only the reason: the cosmos knows itself in parts because otherwise it could not. Knowledge cannot be global. It is always born local, in something small and threatened, and spreads from there.
If tomorrow that capacity stops needing cells and passes to a machine, the nature of the matter will not change: it will still be a part knowing the rest. The support will change, not the condition.
Something similar can be said of consciousness, and also with caution, because here there is no agreement among those who have been arguing about it for centuries. If knowing is modelling what is outside, being conscious seems to be what happens when that model turns inward and the system includes itself in the map it was drawing. That too belongs to something limited: you need to observe yourself while you can still be lost.
A curious symmetry remains, and with it we end. As a part understands better what surrounds it, it uses up the outside. And when it has understood almost everything that is not itself, the only question still left whole is what it is.
This essay does not answer it. What it does is leave it written down, because everything else on this site, the method, the capital, the risk, the phases, was born of someone not settling for an open question.
References: C. Bernard and W. Cannon on the internal environment and homeostasis; H. Maturana and F. Varela, Autopoiesis and Cognition (1973), source of autopoiesis and of the idea of cognition as a way of staying coupled to the environment; R. Dawkins, The Selfish Gene (1976) and The Extended Phenotype (1982); L. Wittgenstein on badly formed questions; K. Gödel, incompleteness theorems, used here only as an analogy; C. Sagan, Cosmos (1980); W. B. Arthur, The Nature of Technology (2009), on combinatorial evolution; S. Klein, How We Change the World (2023), on the collective brain.
From the problem to the product, and from the product to the market: that is where the next essay continues.
Francesco Esposito, founder of Volcano.
John F. Kennedy, 1962