Introduction to Agential Biology

PART 1: Introduction

Our project is called Agential Biology Institute or ABI for short. The ABI is founded because we think there is a missing causal layer in biology that gives cells an intentional direction as they develop, and we are building the infrastructure to study that causal layer. We think there is a paradox or a contradiction at the very heart of biological theory and we have a new set of experiments we think can make enormous progress on this problem. 

If we can work out the mechanism of cellular agency, it would be the most high-impact discovery ever made in biology, because it would give us a scientific theory of cellular coordination. This would revolutionize our approach to cancer, autoimmunity, diabetes, even aging. It would also transform the way we think about the next level “up,” too — organisms coordinating with one another and with the world. These are problems like addiction, crime, war and peace, social media, and AI. It would give us a new picture of what is possible for organic life as a whole.

The general consensus in biology today right now locates agency at the end of development. In this story humans begin as single cells which are gene-driven robots (it follows instructions coded for in its DNA). The zygote divides and the human grows larger, and at some unspecified point, the causation flips and there is a “mind.” The human starts obeying the mind’s instructions (as well as the genes’) - instructions shaped by the outside world. That is agency. 

We have new evidence that this is not the right story, because now we know organisms, even single cells, can change themselves and their genes in a heritable way. They can cut, splice, edit their DNA, and they can inherit in many ways outside of DNA pathways. 

These results compel us to think of DNA as an information resource that cells use rather than a set of instructions. So how does this actually work? If this mechanism is present in every cell, we are asking: what are the common features of all cells and all organisms that could be the mechanism of this universal cellular learning — this agency? 

We have an incredible team of advisors, research partners, and staff at ABI that are truly world-class. All our research partners have labs, they have post-docs, and they have institutional support. So why aren’t these tests already being done? The kind of research questions we have, questions of teleology, can only be asked and answered outside of normal institutional funding because they are intrinsically outside the paradigm of genes-first biology. We are developing tests of philosophical concepts, tests of teleology and function, tests of cognitive theories that depend on very different premises. They require interdisciplinary partnerships and conversations that cannot possibly happen within normal strictures because the answers have to be integrated from different domains, different labs that look at the problem in different ways. 

So for example, Denis Noble is on our board, he’s a fellow of the Royal Society who invented systems biology. We also have Stu Kauffman on our board, an inventor of complexity theory. Michael Levin is a research partner, he’s published in Science and Nature and been in the New York Times several times for his groundbreaking work making two-headed planaria and frog skin cell robots. Other research partners include Oded Rechavi, a Schmidt Science Polymath who studies epigenetic inheritance, Josh Bongard, who pioneered work in swarm robotics, Eloi Camprubí, who studies chemical origins of life, and Richard Watson, who invented natural induction, a computational framework.

We are talking about a set of specific testable hypotheses that have never been tested before. Whichever way each test goes, the results will be decisive and will tell us which way to turn in our search for this mechanism of agency. Agential biology is poised to be the deepest transformation in our thinking since the concept of the gene was invented. 

PART 2: Teleology

Let’s return to the problem of agency for a minute. Biology has an agency problem. The problem is that purposive trying is excluded from the theory of cells but we can’t understand life without it. I want to address two issues that are really important that people tend to misunderstand:

1. The first is the controversy about epigenetics, or the inheritance of acquired characteristics, the fact that most of what is learned is not heritable.

2. The second is the false distinction between bodies and minds.  

As I mentioned, lots of new evidence shows that organisms can change their DNA (and the way it's interpreted) in ways that appear purposive — in ways that are nonrandom. However, it is true that most of the change that happens during an organism’s lifetime is erased when the sex cells are made, especially in humans. Humans are an unusual species in that our genomes are especially well-protected from change. For example, if your grandmother learned to play the piano, you weren’t born with that skill already developed. This prompts some people to claim that “epigenetics is fake.” It’s true that most epigenetic changes are not heritable and the effects of the ones that are can be quite mild. But that isn’t the right standard. The existence of ANY epigenetic inheritance is a problem for theoretical biology. 

The scientific paradigm we are in in biology is called Neodarwinism. Neodarwinist theory actually has a very strict standard for the sources of variation. The variation must be blind for the theory to work logically. If a theory is formulated rigorously, then just a small bit of specific evidence can falsify it. We already have way more evidence than we need to falsify Neodarwinism. The existence of ANY degree of non-random epigenetic inheritance renders Neodarwinism a critically flawed theory. Which leads us to search for a better one. So some of our assumptions will have to change.

An organism can be considered to be running a program of some sort. The question is how has all that information gotten into that program in the first place? The existing assumptions of biology say that the instructions of life are generated by blind chance, and random changes to the program are responsible for accidentally arriving at good instructions, instructions that result in a living, surviving, reproducing being. The reason this explanation focuses on the DNA is because that is the only place where information is inherited faithfully, where everything is encoded digitally and corrected for errors. 

But there is something important to understand: The DNA is largely identical in all our cells, but cells change the way DNA is turned on and off. That’s called epigenetics. That is why starting from one zygotic cell, our bodies differentiate and become bones, muscles, hair, blood, organs, etc. When the sex cells are made, most of those epigenetic changes are stripped away, but not entirely all of them. 

The changes that make it through to the next generation are often tiny, minuscule differences but they are very, very important. Those variations are the source of all the directionality in evolution. So for example a lineage runs from chordate worms to jawed fish to dinosaurs to bald eagles, to trace just one lineage over hundreds of millions of years. All the changes that made that eagle from its ancestor the oceanic chordate worm are coming from that tiny bit of variation that persisted after most of the changes were wiped away. Natural selection wipes away all the variations that don’t work, but that doesn’t define which variations occur. Selection is a non sequitur when we are trying to explain what variations occur in the first place. 

The key question is this: are the changes that pass through to the next generation completely random, or do they have something to do with the function of the parent organism, given the conditions it was living in? For example, one of our research partners, Oded Rechavi, has shown that starvation in C. elegans worms causes them to modify the pool of heritable small RNAs that regulate nutrition, making their offspring a little more resistant to starvation. Well, of course, you might say, of course those changes are a little bit in the direction of function, because the parent worm was trying to survive, and changed itself in order to survive. It’s actually hard to imagine an organism not passing along a little bit of all the adaptation it has done during its lifetime due to all the stresses it has encountered. Especially if the organism is a single cell, its entire body becomes the two daughter cells, so where else can the adaptation go? It has to go to one or the other daughter cells. Of course it passes along some beneficial changes, because it was trying to survive. 

But that is the intellectual trap right there. The word “try.” The word “try” in that sentence is introducing what we call teleology. It is introducing the intention to do something. It’s introducing purpose. That is an illegal move in textbook biological theory. There isn’t supposed to be any trying involved, only blind chance. That is the paradox we are working on.

The problem is we cannot understand the behavior of animals and plants at all unless we interpret them as trying to do something. Teleology is always smuggled in, you see. It’s always assigned to minds, to psychology, and supposed to only be in metazoans with neurons. If all the trying comes from minds, and all function comes from bodies, then we are in this philosophical trap called Cartesian dualism that can’t connect these things. The mind and body. The mind is nonphysical and the body is physical. But we can see, on some basic scientific level, that they are one thing. Bodies and minds always travel together, and they can’t do without one another, and they both have causal influence. We can see that trying happens everywhere in living beings. We call it different things: homeostasis or adaptive plasticity or basal cognition or nature vs. nurture, but it all amounts to the same thing. So what is the connection between these two ways of understanding organisms? How do we scientifically connect body and mind?

There is a key unlock here. In order to make sense of this situation, we have to make a big frame shift. This is the deep part of this agential biology idea, and it has big repercussions. If you think about what it means to be a learning system, it means that you take information that you gather and integrate it for some purpose. Purpose is intrinsic to the idea of learning. Purpose is what makes information meaningful. But if you are an individual learning about the world, the domain you are learning in does not include any natural selection. Natural selection is not an event you can learn from in your lifetime — since the act of selection is linked to your death and your total lifetime fecundity — this is a score that is only counted up after you as an individual are gone. You learn from things that happen to you, from trial and error. You learn from trying. So what are you trying to do? 

The genes of an organism, if we can say that genes learn, they learn as a lineage — across generations, not within them. The domain of learning is the persistence of alleles within a gene pool — the population of alleles (which is different versions of a gene) shifts over time based on natural selection. But for an individual to learn, to add something to the program that can be inherited, that’s functional, it has to be for a different reason. This is the important part: There has to be a different primary purpose if individuals learn. They cannot possibly be trying to survive and reproduce. They must be trying to do something else. 

This key unlock is where our first ABI tests become relevant. Because teleology is actually readily testable, all we have to do is control rigorously for natural selection. Some theorists impute a lineage-level purpose to life, survival and reproduction, inheritance. But there is another possibility. 

So my guess as to what they are trying to do is this: We, all living cells, are all trying to understand causation. We are looking to see what wiggles what. Even bacteria in this view of life are like natural philosophers. They are like human children. To some tiny degree they are trying to increase their own causal knowledge, their understanding of what causes what around them. This is the possibility that being is for knowing. Survival and reproduction may be accidental. It doesn’t mean they don’t happen, just that they are an accidental side-effect of knowledge. That would explain why new mutations in bacteria have sometimes been shown to be random with respect to reproduction. 

I suspect that organisms are trying to know something about the worlds they live in based on the bodies they have. This is the concept of epistolution. It’s a word I invented that combines epistemology and evolution, and means something like “universal cellular learning aimed at causality.” This would mean that the directionality of variation, the organizational force that makes a trait a trait, that makes DNA functional, is the drive toward knowledge, understanding.

You see truly random variation in an organism looks like fatigue, or aging, or disease. A novel trait is already a coordinated thing. It’s already organized. It’s an aesthetic, creative change. That is what is being presupposed in Neodarwinism, that a change equals a new trait rather than just disintegration into a pile of mush. If these changes are created to understand the world, then they are indeed random with respect to survival and reproduction, but they are morally and aesthetically creative and valuable. 

The possibility of epistolution changes everything. It explains what we see in nature in a much deeper way than the theory we have. And the amazing thing about it is that it is testable. We can test that basic assumption. The big unlock is to accept that teleology is an intrinsic part of biological explanation and then empirically test between the two teleologies of inheritance and epistolution. Epistolution is a new possible teleology that is agential. It integrates body and mind into one system. 

Beyond that there are tests that can be carried out that would really open up this field, and there are several plausible theories of cognition to test between that are compatible with the teleology of epistolution, as well as others to test if the correct teleology turns out to be inheritance.

For example, we can look at heritable stress responses in lineages without natural selection. We can study memory and learning in the other parts of the cell, such as the membranes, the structured water, and the microtubules, and understand what is holding information in an analog way. We can study assemblies of cells that have no evolutionary history, like Michael Levin’s anthrobots and xenobots, and see what they can discover without natural selection. We can build models of Bayesian network Hebbian learning or harmonic resonance that correspond with cell-level dynamics and see if these models do in fact have an epistemic curiosity — to see if they scale intelligence. This project can constrain the models we have of cell behavior with a totally new way of interpreting biology that makes more sense. This could lead not only to much better therapies, but to engineering true understanding: intelligence outside of biological material. 

You will notice that none of this addresses the question of consciousness, because consciousness is just one aspect of mental phenomena. The rest of the tree of life, presumably, is conscious in very different ways from our own, and we are conscious very differently when we are asleep, knocked out, meditating, or on drugs, yet we are alive in all these states. We cannot possibly know much about the problem of consciousness until we study teleology in a rigorous scientific way. But in order to do that, we have to look at nature in a new way.

This way of looking at nature gives us a new picture of our own motivations and links them to the living systems around us. It portrays all beings as moral beings, as sentient, as aware, as active and purposive. Morality is a natural phenomenon, and so is technology. Because we come from biology! This agential point of view does away with the intellectual separations that hold us out of the theory of life and put us on a pedestal. We are not on the throne above nature, we are just parts of the living system.

PART 3: A Critical Test

Let’s talk about the actual critical tests that ABI wants to do. What do we actually plan to do in the laboratory? 

A critical test in science forces a choice between two possible hypotheses; if the test comes out one way the first hypothesis is disproven. If it comes out the other way the second hypothesis is disproven. Nothing can justify a hypothesis as the final truth. Every theory is just a guess about the world, but a critical test gets you to a better guess than the one you had before.

The first thing is to make sure your theories are in a rigorous, testable form. People have a tendency to modify their theories by adding in observations that don’t follow from the actual rules of the theory. In this case, many people are saying, “well sure genes are the instructions, but we know there is some epigenetic inheritance too, so our theory must allow for that.” So now the theory has “genes are the instructions, plus there is a little bit of agency, too.” This observation doesn’t follow from the Neodarwinist theory. The whole idea of genes is that for four billion years, everything in biological systems has been organized by the tendency of DNA to replicate itself. So everything that a biological system does is only explained if it tends to cause DNA replication. In the Neodarwinist theory, an organism is only a vehicle for what its genes want to do – metaphorically. This idea will be familiar to those who have read Richard Dawkins’ The Selfish Gene, which is an admirably rigorous account of the consequences of Neodarwinism. 

The whole concept of agency is that organisms can have minds that change how they behave, based on what they learn. In trying to match these observations to Neodarwinism, people have said, “well of course, the organism has evolved to have this capacity to adjust to the circumstances, to go this way or that, because it has to do that in order to survive, and if it didn’t survive, it couldn’t replicate genes.” In other words, in the Neodarwinist inheritance teleology, inheritance is the only drive that matters, and everything is either random or pointed toward inheritance. 

Contrast this with what we are introducing, which is the possibility that organisms may want to know about the world, and gene replication and inheritance is just an accidental side effect. So in the epistolution teleology, everything iseither random or pointed toward learning.

How could we test for learning in single cells? Isn’t it hard to tell if a single cell has a worldview? This is where we must use a proxy that is more obvious. The idea here, which may sound like a leap, is that learning — understanding the world — makes your body and mind more resilient and durable as a self. We are arguing that this is the actual source of the self. Understanding is the process that makes a self happen. So learning should make the individual more resilient to the stresses it encounters. 

The trick is to find a condition where we would expect resilience as an individual to be improved, and inheritance as a lineage to be worsened. In other words, to test our teleology, we have to find an experiment that we think will force organisms to learn something useful but go against the interests of their genes. That would disprove the idea that inheritance is the teleology of life. If the test doesn’t have that effect, and instead organisms are faithful to their DNA instructions even when we think they should not be, then epistolution will be disproven. 

So here’s the test: First, we set up several lineages of organisms that can carry on with no natural selection. Every difference that is inherited in this experiment has to be acquired from learning, not selection. So for example, we have a mother who creates one daughter, and then that daughter grows up and by herself creates one daughter, who grows up and by herself creates another daughter, and so on. None of the daughters are removed by selection. In bacteria this would be easy because when the cell divides by mitosis, you just flip a coin and one of the two daughters carries on the lineage, and the other can be sacrificed.

Second, we apply a stress condition to some of the lineages and not to others. Some are controls and some are treatments. Say, for instance, we put some of the bacteria in high heat – a constant high heat for many generations – and others in normal temperatures. If epistolution is the right teleology, the bacteria in the heat condition will eventually adapt to be able to tolerate the heat better over time even without any natural selection. If any of the bacteria die in the process before being killed in a measured way (to test their resilience), we have to throw out that lineage, because that constitutes selection. Say over thirty generations, we expect that after the mother cell has reproduced and made her daughter, it will take more heat to kill her. So we expect to see the bacteria learn to deal with the heat in a linear, progressive adaptation even without natural selection. This would show that variation in the lineage is nonrandom.  

Now here is the key part of the test. After a long time adapting –  many generations –  we compare how well the treatment lineage vs the control can compete with one another to reproduce in the heat condition.  We also expect that, since reproduction isn’t part of the epistolution teleology, that over time, even though the bacteria are getting better as individuals at living in the heat, they are changing in ways that are no better than random at reproducing their DNA in the heat. So we would expect that even as they become more resistant to being killed by the heat (After they reproduce) , when they are put back into a common population with those that have never had any heat at all (a naive strain), they will be  outcompeted in reproduction by those naive strains, even in the heat condition. 

So if these experiments go the way we expect them to, the lineages of bacteria that never experience heat stress will actually reproduce better in the heat than bacteria who have evolved to be harder to kill with heat. That’s the weird result that we expect and epistolution teleology requires. That is the perverse, unlikely prediction that will be the critical test. 

You might think the naive bacteria would die from the heat, but the constant heat condition is at a survivable temperature for the naive bacteria, otherwise the test would not have gotten going in the first place and then lasted thirty generations. But why would the naive bacteria reproduce better than the heat-adapted ones in the heat? Because this experiment looks at adaptation without natural selection. We are only looking at acquired changes, not selected changes. Adaptation without natural selection isn’t about reproduction. It’s random with respect to reproduction, which means it usually will not favor reproduction but rather diminish it. It can’t be about reproduction because the changes come from an individual making choices for their own reasons, not a lineage winnowed by natural selection. That’s why this is a critical test. 

In principle, epistolution applies to all organisms and all stress conditions, so in principle we could do this to any taxa and use any type of stress. Occasionally the test should fail, because of course sometimes learning does cause enhanced reproduction, otherwise every lineage on earth would be extinct. But usually it should not. Eventually over enough generations we should be able to make any lineage of any type of organism evolve away from its reproductive interest while simultaneously evolving individual resilience. That should be impossible if inheritance teleology is true, but possible if epistolution is true.  

You might ask why this overturns anything, since all we would be showing is that survival is a priority of life over reproduction. But this is exactly the point: a new theory, in addition to proving the old theory wrong, always has to show why the old theory appeared to be true for so long. In this case, because we have never rigorously tested whether all life had individual agency, it may have looked like agency was just a proxy for inheritance and replication of DNA. You see, biology isn’t supposed to care whether an individual lives or dies, it is only supposed to care whether that individual reproduces its genes. If the fundamental life process promotes the individual at the expense of the lineage, then there is a logical requirement for a new explanation. 

As well as the principles involved, we have two other sources of empirical evidence that make us expect this perverse result. For one, there have been many prior trials of novel mutations in bacteria that have shown that novel mutations incurred by stress are not typically correlated with greater reproductive advantage. That has been a major source of evidence for those who have argued that mutation is random, but this is a mistake; it only shows it to be random with respect to reproduction.

Second, we have an enormous natural experiment going on right now in the human species. If all biological agency served the interests of genes, then the species with the most agency should logically be using it to produce more genes. Instead, despite unprecedented peace, medical care, and food, the majority of the human species has fallen below the replacement rate of fertility, even in countries like India and Mexico. We humans have used our biological agency to increase individual lifespans while reducing our fecundity as a lineage. This fact is not explained by Neodarwinism. Humans are caused by biology. We can’t have a fundamental theory of life in which a major species disobeys the rules of the theory.

What will we do if these experiments go as we expect them to do? If we can safely assume that the inheritance teleology is critically flawed, we can look for the specific cognitive theory that will give us the best mechanism of epistolution. There are several fundamental ideas that are candidates: For example, Karl Friston’s free energy principle may be one. Robert Rosen’s MR anticipatory systems theory may provide another. Richard Watson’s natural induction may be another. I personally suspect a fourth version is correct, an idea I call Popperian double resonance. Michael Levin is developing another concept involving platonic forms that may soon provide another testable alternative. This is when our investigation becomes much more concrete, and we have to develop universal tests of memory, learning, and curiosity. 

The reason we have such a broad-based interdisciplinary team is that these universal tests have to show that the mechanism of cognition works in all sorts of different domains. Richard makes computational models, for instance, Josh Bongard makes swarm robotics, Marco Cavaglia measures lipid membranes, and Michael Levin makes novel groups of cells. Any candidate mechanism will have to work in all these different contexts in order to pass our tests. We are trying to truly create an integrated, universal theory of biological agency. 

PART 4: Deep Optimism

So what does all this mean for how we should think about life? How should it change our hearts? What is all this saying about the status of our scientific understanding of what human beings really are?

The scientific possibility that being is for knowing rather than survival and reproduction is a major change to our worldview. This is my personal passion, and I hope to inspire you to be optimistic about the possibilities here. I think that an integrated scientific theory of bodies and minds should open up our hearts to one another and to other forms of life. It should make us more compassionate and respectful about what others are trying to do. From this point of view, all humans are trying to learn. The fungi in the soil of the forests and the plankton in the oceans are trying to learn. Even the tiniest of all bacteria or archaea may be trying, to some small degree, to improve the world they find themselves in by learning and changing themselves to some extent. 

First let me point out a few ways that the Neodarwinist theory is deeply embedded in our culture. The idea that life forms are trying to survive and reproduce is not just a trope used in every modern nature documentary, it is a deep anchor for all our political, social, and environmental sciences. In political science, the evolutionary history of humanity is used to explain why we are violent, why we posture for social dominance, and why our societies are hierarchical. In environmental studies, we assume the drive to reproduce imposes a mandate on us to strip the earth of its vital riches before competing lineages, other nations, remove them from our grasp. The nature vs. nurture debate is a long-running discussion about whether we are able to somehow transcend an inherently primitive human nature baked into our DNA.

Intentions from the Neodarwinian point of view are in the realm of theology, of psychology, and of art. They are disconnected from the natural sciences like physics and chemistry and biology, and merely presupposed in disciplines like economics, law, politics, and business. Throughout the Western philosophical tradition, minds and souls have been supposed to be immaterial. We suppose ourselves to be free to make up our minds as we wish, independent of any physical influence, and even those wishes are supposed to come from ethereal sources. 

Ironically this scientific denial of agency leads us to overestimate its power. Without a scientific theory of agency we cannot measure its causal impact. This leads us to pretend that we are complete masters of our own destinies. If I’m a typical person, I think that I in part decide the direction of the events I participate in, that I have important decisions to make about my life, and these decisions impact the world profoundly. I likewise suppose that as a human I control nature, that I am a guardian of the earth, and I have free choices about whether to destroy or sustain its resources, which I see as largely irreplaceable. And the question of what exactly is the “I” in this sentence is never addressed. 

Morality is the name we have for our intentions in the abstract. As a result of this separation we see our moral values as immutable laws of the universe, revealed to us through religion or through rational discourse. We treat morality like mathematics, as an intrinsic set of truths discovered about right and wrong. Since, like mathematics, it has no physical basis, we assume there is no process of testing or learning from evidence involved in developing moral values as a society or as a species. We suppose our values to be eternal features of the world, outside of history. We pick good guys and bad guys in history, raising up heroes and villains according to today’s standards of behavior. We always assume that if we were born in a past era we would have surely been good guys, and that the bad guys surely should have known better.

All of our plans for the future as a society are based on these assumptions. We are looking at the tremendous technological changes of AI, crypto, drones, climate change, and nuclear fusion through this lens, and through this lens we are rightly terrified. We suppose that the resources we need to survive as a species will soon be destroyed by violent competition, and that our great-great grandchildren will inherit a disastrously polluted and damaged, almost unliveable planet. We see the process of technological development as highly dangerous and uncertain. We assume that we will need colonies on Mars or elsewhere as an insurance policy against the destructive power of our runaway technologies. The only hope we have in this context is an irrational, desperate hope, an illogical wish that there may exist some small group of heroes with good intentions that might salvage a small victory from the apocalyptic forces aligned against us. And such is the plot of every drama we can imagine about the future.  

The possibility of an epistolution teleology changes all that. The idea that our intentions come from a natural process is an unfamiliar notion to us that reframes nearly everything. Ironically, identifying a physical process as the source of the self actually should change our image of living beings from selfish to unselfish. 

The creation of reproducing lineages is a zero sum game in which competition forces life to kill other life or else die out. The creation of knowledge, on the other hand, the epistolution teleology, is not zero sum at all but its opposite; it is a positive sum game.

The more you win, the more others win. 

Let me repeat that. The creation of knowledge is a game in which the more you win, the more others win. That would mean that the resources we need to live as human beings are not really specific materials, but rather the knowledge necessary to make use of whatever is available. 

Knowledge creates resources out of things that were once useless. So for example, we once feared that we would need almost unlimited farmland to feed the world’s population. But we discovered how to use energy to make fertilizer literally out of air; it’s called the Haber-Bosch process, it’s how we fix nitrogen to make chemical fertilizer. Now, in most developed countries like the US, Europe, Australia, and the UK, huge swaths of land that were once farms have grown back into forest because we have learned to produce far more food with far less land, and those countries have huge food surpluses. Even natural ecosystems around the world now largely depend on this artificial process for the initial source of much of their essential nitrogen.  So the growth of the vast continental forests of America and Europe would be impossible without an industrial fertilizing process to sustain them.

Technology is the name we use for changes in what we do with our moral intentions. The reason we build spaceships and computers and feature films is because we, each of us, decide what to do with our lives based on the obligations and the constraints that we face. We have made these things because we thought it was morally acceptable or even imperative to make them, and others have agreed. 

Our current economic theories do not contemplate any change of moral intentions, and only deal poorly with the growth of technology. We have assumed that these constraints were imposed by the scarcity of money, of time, of space, of food, of clothing, and that we were doing these things to promote our individual interests over those of other people. This is the competitive theory of Adam Smith, the invisible hand, of supply and demand. But these economic theories are based on Neodarwinist assumptions, too. 

Epistolution teleology raises a possibility that in the future, there may be a much larger role for what economists call anti-rivalrous goods. These are things that by using them become more valuable to others. The internet is one of these things, some others are AI, transportation, and national security. The more we use the internet and the more we query LLMs, the more valuable information gets stored within these systems, and the more useful they are. These kinds of goods make it faster, cheaper, and more efficient to satisfy material needs, because they increase knowledge.  

When we imagine the human future, we basically extrapolate today’s problems unchecked because we cannot imagine the solutions to them yet and we don’t trust people to solve them in the future. But the epistolution teleology suggests that not only humans, but all living beings are adaptive problem-solving systems and that we all basically have the same interests and intentions. Of course the knowledge that a fungus has is hardly useful to us, but sometimes, it is. For example, penicillin was invented when a scientist noticed that a fungus had evolved an anti-bacterial compound. Penicillin was “co-invented,” you might say, by yeast and humans. 

There is no scientific reason to insist that the entire living system of the earth has intentions itself, or that Gaia has our best interests in mind. It is not necessary to claim that nature itself is like a benevolent God or cosmic protector. But epistolution shows that biology may have the feature of possibly being an interconnected informational network, a network of meaning, and this raises the possibility that there are some self-balancing features to biological systems.

Problems always arise before solutions, but living beings focus on solving the problems they think are worst with the most urgency. Solutions always come with unintended side-effects. This means we are continually trading old problems for new problems. 

For example, two side-effects of efficient food production using the Haber-Bosch process are climate change and obesity. But as bad as these problems are, most of us in our hearts would be morally satisfied knowing that we had prevented the acute starvation of those we love, of our grandparents, and children, by creating a slightly warmer world and some extra fat on the bodies of the survivors. Thus we may have traded worse problems for better problems. There is good scientific reason then to assume that on the whole, this may be the future that we can expect; increasing complexity, but also better and better problems. Better problems means less urgent, less disastrous, less irreversible problems.  Not a perfect world, but a less bad world. 

That is the possibility I call Deep Optimism. It’s what I consider to be the cultural and social implications of the epistolution teleology that is currently being tested by the ABI. I hope this idea inspires you a bit. I hope it restores your faith in the future of humanity and the Earth to some small degree. 

PART 5: Intelligence

I hope so far in this explainer series I’ve convinced you that agential biology is worth pursuing for its value as a scientific explanation. But what difference does it make? What will be able to do or build when we have a thorough physical theory of agency? If we can make an agential system from scratch from abiotic materials, how will that change the world?

I want you to come away from this explainer with an urgent message. Our medical progress as a whole society is blocked by the problem of agency. People are suffering from deadly and painful medical conditions because we have not figured this out.

We have made enormous progress in medicine in the past century, but the progress largely falls into two buckets: On the one hand, the most lifesaving medical breakthroughs have come about because we have learned how to kill organisms and viruses that should not be in or on our bodies. All the antibiotics, anthelmintics, vaccines, and personal hygiene products fall into this bucket. The vast majority of life extension in the 20th and 21st century has come from simply cleaning the environment with indoor plumbing and sanitation, and cleaning the body of internal parasites and pathogens. Many of the infectious diseases that have been defeated are like this: smallpox, cholera, yellow fever, measles, typhus, tuberculosis, malaria, plague, etc. The second bucket are medical devices and surgeries. These are basically mechanical engineering in the body. The treatment of trauma, joint problems, circulatory blockages, and nerve connections has been transformed by our detailed knowledge of anatomy and physiology and our incredibly advanced technical products and services. Cardiac bypass, pacemakers, dentistry, pulmonology, brain surgery, orthopedic devices, and so forth. 

But what are the remaining problems? What medical ailments have we failed to solve? What pathologies have in fact grown worse? 

There are three large and growing classes of problems that stem directly from the body’s agential relationship with itself and with the outside world. First, the immune system is the aspect of the body that detects self vs. nonself and reacts accordingly. This system is increasingly going awry for reasons we do not entirely understand. Food allergies, celiac disease, IBS, arthritis, colitis, asthma, autism, and chronic inflammation are all on the rise, and no treatments are conclusively effective. Autoimmunity is a rising epidemic. Gluten, peanuts, shellfish and other historically common food components are now the source of deadly sensitivities in nearly every preschool. This was unheard of just a few decades ago. The blame has been placed on food companies, but testing has not linked these maladies to any specific foods or food additives. Obesity is increasingly understood as a food-related hormonal imbalance, but since it causes craving for foods that also exacerbate it, its sources and remedies are still obscure. 

Second, cancer may have become more, not less common. As the population grows older and testing for cancer is more widespread, we would expect an increase in the rate of deadly cancers. But cancer has also increased in children and young people. Cancer, like autoimmunity, is a disease of the cellular recognition of the self. A cancer is a cell or group of cells that operate as if the human body is an external environment. Instead of cooperating with the rest of the body, cancer cells go rogue. Cancer treatments have been characterized as “burn, poison, or cut.” We either attempt to radiate cancer, treat it with toxic chemotherapy, or excise it with surgery. Despite advances in these three techniques, and despite much greater public awareness of the cancer-preventative effects of good diet and exercise, and despite lower rates of smoking, cancer is a larger problem than ever before. 

The third major class is mental illnesses. Depression, anxiety, addiction, PTSD, eating disorders, ADD, ADHD, and other psychiatric illnesses have become routine problems for developed-world families. These illnesses are, as we have discussed, literally diseases of an entity — the mind — that has at best a questionable ontological status in Western medicine. For example, all FDA-approved medical treatments and substances are still subject to placebo-controlled trials. The logic of placebo control is simple; mental effects do not count as medicine, therefore effective medicine must be evaluated absent mental awareness of having been treated. Consonant with that logic, the chemicals that we have developed for treatment of mental illness have had only questionable effectiveness. Since few depressions go medically untreated in developed countries, even these blockbuster drugs have been seriously questioned as possibly no better than placebo at preventing longterm mental suffering or suicide. Meanwhile depression and other mental illnesses continue to increase. 

These three classes of problems share a common source: agency. The genetic makeup of the human species has not shifted meaningfully during the past fifty years, so we can’t blame our DNA for these problems. A change in genetics would take thousands of years to take effect. We attribute all the increase in chronic disease to environmental changes, but what specifically is it about the environment that we should restore to its previous condition? In order to identify, for example, the layer of information that is necessary to change in order to dissolve a tumor or a blood-borne cancer, we would need to have a theory of what mechanism is making normal cells cooperate with one another in the first place. 

Once we have such a theory, manipulating cells is, in principle, unlimited. If agential signalling can change cells from a human embryo into an adult spleen or pancreas or brain, then surely signalling could in principle reverse cancer, autoimmunity, and mental illness. All biological systems are intrinsically regenerative. This is why wounds heal and homeostasis continues —- it’s why life stays alive. For example, our research partner Michael Levin has mapped the bioelectric layer of a planaria, writing to the cells a new program that causes them to develop two heads rather than one. 

What would unlimited reading and writing to cells look like? It would look like the unlimited power to engineer new tissues and organs and integrate them into our bodies when the original ones have become aged or diseased. It would look like treatments that restore the whole system in a healing way like fasting, exercise, and sleep. It would look like explicitly designed medicine to enhance core spiritual and relationship health. It would look like actual cures for cancer. It would look like restorative compounds that re-order immune reactions without causing inflammation. It would look like the indefinite extension of healthy life. 

Drug development today is plagued by contrary indications and side-effects. It takes decades to develop effective drugs. Imagine if in order to alter an airplane design, we had to blindly fly thousands of airplanes and see which ones crashed and which ones made it closer to the target? This is much like how drug development has traditionally proceeded. With a physical theory of agency, we could design personal drugs and interventions that altered just the very aspect of our living system that needs altering, while affecting all the others positively rather than negatively. 

So what else could we do with a physical theory of agency? Mind is, of course, the original source of what we call intelligence.

I think there is a misunderstanding going on in our culture right now. It is living beings that invent symbols that represent things because we have a hunger to know about the world. Lately we seem to be unable to distinguish this from the act of indexing those symbols so they can be queried. AI, although it impersonates a human, is just an indexing technique. Recently people have been worrying that AI will soon take over the world and extinguish humanity. Strangely, they never made this mistake about previous indexes of symbols like dictionaries, telephone books, train schedules, or technical manuals. Now that the indexing technology produces the answer in a lifelike, comprehensible, friendly voice, many commentators are breathlessly wondering if these indexes could replace humans entirely, go rogue and destroy us, or even make humans lose their creativity. I think these worries show a deep misunderstanding of the problem of biological agency.

Computers are logic machines. They follow strict orders. If one bit is out of place in a vast series of precise instructions, the computation stops. Today’s advanced codebases contain various failsafes and duplications, but these are also strict logical structures, incapable of the plasticity, regeneration, and adaptation that is characteristic of cellular agency. Today’s AI doesn’t have the capacity for purposive trying, for agency. It follows its instructions. Goals are set by the programmers, the data, and the input from the user. If the system contains the answer, it produces it. It has no choice.

The goals of an epistolution-driven agential system come about through an entirely different process. Its goals are determined by the state of its knowledge. What it wants to do is generated directly by what it already knows, contrasted with the situation it’s in. I once coined the term “expect-actions” to describe this. Its expectations ( “expect-actions”) rub against the world, creating sore places, contradictions that need to be worked out and investigated. Knowing is a moving process of intentional change, testing causality. This means inventing concepts and testing them against the world. There is no way to be this kind of mind without also being a body. The epistolution teleology explains why a body needs a mind, and a mind needs a body: they are the same thing. The only way to extract meaning from the world is to be an ongoing process of trial and error. This kind of system is a restless thing, never completely satisfied.

What would that type of agential system be like if it was scaled up into a vastly complex structure, more complex than humans? What would it be like to be with a system that contained genuine understanding and wisdom far beyond our capacity for knowledge? What would an exponential explosion in wisdom look like?

Today’s AI has an astonishing breadth, and can answer questions about any subject in its training data, which is now nearly the entire internet. But the answers are no more profound than the people who put them there, and since they are a sort of average, typical response triangulated from many sources, it stands to reason that they will never be as insightful as the very best people. 

The Zen philosophers might tell us that great wisdom looks like great foolishness. Perhaps the agential systems would be clowns, like Shakespeare’s Tom in King Lear, mocking and teasing us into observations that we are not capable of the self-knowledge to appreciate in their direct form.