Deep Learning and Abstract Orders

It is well known that Friedrich Hayek once rejoiced at Noam Chomsky’s evolutionary theory of language, which stated that the faculty of speaking depends upon a biological device which human beings are enabled with. There is no blank slate and our experience of the world relies on structures that come from the experience in itself.

Hayek would be now delighted if he were told about the recent discoveries on the importance of background knowledge in the arms race between human beings and Artificial Intelligence. When decisions are to be taken by trial and error at the inside of a feedback system, humans are still ahead because they apply a framework of abstract patterns to interpret the connections among the different elements of the system. These patterns are acquired from previous experiences in other closed systems and provide with a semantic meaning to the new one. Thus, humans outperform machines, which work as blank slates, since they take information only from the closed system.

The report of the cited study finishes with the common place of asking what would happen if some day machines learn to handle with abstract patterns of a higher degree of complexity and, then, keep up with that human relative advantage.

As we stated in another place, those abstract machines already exist and they are the legal codes and law systems that enable their users with a set of patterns to interpret controversies concerning human behaviour.

What is worth being asked is not whether Artificial Intelligence eventually will surpass human beings, but what group of individuals will overcome the other: the one which uses technology or the one which refuses to do so.

The answer seems quite obvious when the term “technology” is related to concrete machines, but it is not so clear when we apply it to abstract devices. I tried to ponder the latter problem when I outlined an imaginary arms race between policy wonks and lawyers.

Now, we can extend these concepts to whole populations. Which of these nations will prevail over the other ones: the countries whose citizens are enabled with a set of abstract rules to based their decisions on (the rule of law) or the despotic countries, ruled by the whim of men?

The conclusion to be drawn is quite obvious when we are confronted with a so polarised question. Nevertheless, the problem becomes more subtle when the disjunction concerns on rule of law vs deliberate central planning.

The rule of law is the supplementary set of abstract patterns of conduct that gives sense to the events of the social reality in order to interpret human social action, including the political authority.

In the case of central planning, those abstract patterns are replaced by a concrete model of society whose elements are defined by the authority (after all, that is the main function of Thomas Hobbes’ Leviathan).

Superficially considered, the former – the rule of law as an abstract machine – is irrational while the latter – the Leviathan’s central planning – seems to respond to a rational construction of the society. Our approach states that, paradoxically, the more abstract is the order of a society, the more rational are the decisions and plans that the individuals undertake, since they are based on the supplementary and general patterns provided by the law, whereas central planning offers to the individuals a poorer set of concrete information, which limits the scope of the decisions to those to be based on expediency.

That is why we like to state that law is spontaneous. Not because nobody had created it -in fact, someone did – but because law stands by itself the test of time as the result of an evolutionary process in which populations following the rule of law outperform the rival ones.


A long read

According to Instapaper this article at Wait But Why is a “139 minute read.” And it was time well spent.

It’s about a new Elon Musk venture, Neuralink, but there’s plenty non-Musk stuff in there of interest. I’m agnostic on whether Elon Musk is or isn’t the next coming of the (anti-) Christ. What’s really interesting is the background material this article gives, building up a highly entertaining natural history of knowledge. The section below really captures the main thrust of that story, but it’s worth reading anyways.

minimal tribal knowledge growth before language

And this:

That leads into a discussion of how brains work, that “soft pudding you could scoop with a spoon.” Here are some excerpts:

I’m pretty sure that gaining control over your limbic system is both the definition of maturity and the core human struggle. It’s not that we would be better off without our limbic systems—limbic systems are half of what makes us distinctly human, and most of the fun of life is related to emotions and/or fulfilling your animal needs—it’s just that your limbic system doesn’t get that you live in a civilization, and if you let it run your life too much, it’ll quickly ruin your life.


Which leads us to the creepiest diagram of this post: the homunculus.

The homunculus, created by pioneer neurosurgeon Wilder Penfield, visually displays how the motor and somatosensory cortices are mapped. The larger the body part in the diagram, the more of the cortex is dedicated to its movement or sense of touch. A couple interesting things about this:

First, it’s amazing that more of your brain is dedicated to the movement and feeling of your face and hands than to the rest of your body combined. This makes sense though—you need to make incredibly nuanced facial expressions and your hands need to be unbelievably dexterous, while the rest of your body—your shoulder, your knee, your back—can move and feel things much more crudely. This is why people can play the piano with their fingers but not with their toes.

Second, it’s interesting how the two cortices are basically dedicated to the same body parts, in the same proportions. I never really thought about the fact that the same parts of your body you need to have a lot of movement control over tend to also be the most sensitive to touch.

Finally, I came across this shit and I’ve been living with it ever since—so now you have to too. A 3-dimensional homunculus man.17

This is too far outside my area of specialization to say, but it’s certainly an entertaining read that seems to fit with what I know about these topics (although the section on neurology could be made up as far as I know).

From there it builds up to the moon shot idea Musk apparently has in mind: building the core technology for a high bandwidth mind-computer interface. This would be the ultimate logical extreme of a trend towards better interfaces that’s been going on since before punch cards. If you think over the natural history of knowledge, it becomes clear that this idea is ultimately just a few dozen steps further down a path we’ve been on for billions of years.

And the implications of taking it that far are profound. The pros and cons of that power are huge. Consider how much more powerful your brain is with paper and pencil than without. Or a computer. Or a computer with a GUI and a copy of Excel. Once you can plug into your computer Matrix-style, all those awesome hot keys that let you zip through your computer like a pro will be like roller skates next to a rocket sled. And the two-way link would mean we could genuinely exercise some self control… for example,  by running a computer program that zaps you when you eat too much chocolate cake.

Readers of this blog will hear Hayek warning you! Such a device gives you a lot of power to manipulate a complicated thing in ways we may never be able to understand.

But this type of personal power might be a necessary bulwark against government or corporate power. Network externalities have already locked us in to Google and Facebook. A Byzantine government has created rent-seeking opportunities that puts enormous power in the hands of the politically connected. The NSA is terrifying. And machine learning will continue to get better, giving those entrenched players even more ability to understand and manipulate large numbers of people. (I’m not endorsing this forecast, just listing it as a possibility.)

In any case, even if Neuralink is just an April Fool’s joke I missed out on till now, this article provides a theory of knowledge that’s well worth reading.

In the near future I’ll argue why you need such a theory of knowledge. Stay tuned.

The Political is about to disrupt the crypto-currency scene -or at least they say so.

According to this Financial Times report, Bitcoin is at the verge of a critical decision.

The implications of the chosen terms (“existential crisis,” “decisive leadership,” “political flaw”) are not casual. It looks like the market that crypto-currency had carried from the beginning contain the germ of its own destruction. As in an Escher’s drawing, Bitcoin has unraveled its political strand and its whole existence is, now, dependent upon a moment of decision of the sovereign: the assembly of miners. The decisionist narrative would be fulfilled if the political decision had to be taken by acclamatio instead of voting.

Nevertheless, the decision by acclamation would be still possible: the ones who want “Bitcoin Core” might follow one direction and the other ones, who choose “Bitcoin Unlimited,” might follow their own way. After all, no existential crisis can be solved by voting.

So, which is inside of which? Is the market framed in a system depending upon a political decision of the sovereign? Or does every decision need to be taken inside a spontaneous framework of rules?

We are used to praising Bitcoin for its independence from any political factor: Bitcoin supply depends on a set of rules which allows the public to form expectations about its value with a high degree of probability of proving to be correct.

Taken in isolation, Bitcoin emulates the market. Nevertheless, being independent of political institutions is not enough for being “the market.” The attractiveness of Bitcoin is that it operates in an open system of competition of currencies. In this system, there are many other crypto-currencies, and there might be several variances of Bitcoins as well –in esse or in posse.

Imagine, for example, that Bitcoin effectively splits into Bitcoin Core and Bitcoin Unlimited. Which of the two will prevail over the other? It does not matter. What really matters is that there will be several variances of currencies in competition. The factors that determine the selection of the prevailing currency depends upon a higher level of abstraction that impose an absolute limit to our knowledge.

So, is Bitcoin in an existential crisis? Does a political decision need to be made? Maybe.

But that does not imply that “The Political” will take over the reins of the crypto-currency market. Moreover, opposite political decisions are the linkages which the spontaneous selection process -in this case, of currencies- is made of. In this sense, “Bitcoin Core” and “Bitcoin Unlimited” are attributes of a competitive system and the final prevalence of one variance among other alternatives will not be the result of a deliberate decision but of an abstract process of evolution.

Why do we teach girls that it’s cute to be scared?

I just came across this fantastic op-ed while listening to the author being interviewed.

The author points out that our culture teaches girls to be afraid. Girls are warned to be careful at the playground while boys are expected… to be boys. Over time we’re left with a huge plurality of our population hobbled.

It’s clear that this is a costly feature of our culture. So why do we teach girls to be scared? Is there an alternative? This cultural meme may have made sense long ago, but society wouldn’t collapse if it were to disappear.

Culture is a way of passing knowledge from generation to generation. It’s not as precise as science (another way of passing on knowledge), but it’s indispensable. Over time a cultural repertoire changes and develops in response to the conditions of the people in that group. Routines, including attitudes, that help the group succeed and that are incentive-compatible with those people will persist. When groups are competing for resources, these routines may turn out to be very important.

It’s plausible that in early societies tribes had to worry about neighboring tribes stealing their women. For the tribe to persist, there needs to be enough people, and there needs to be fertile women and men. The narrower window for women’s productivity mean that men are more replaceable in such a setting. So tribes that are protective of women (and particularly young women and girls) would have an cultural-evolutionary advantage. Maybe Brandon can tell us something about the archaeological record to shed some light on this particular hypothesis.

But culture will be slower to get rid of wasteful routines, once they catch on. For this story to work, people can’t be on the razor’s edge of survival; they have to be wealthy enough that they can afford to waste small amounts of resources on the off-chance that it actually helped. Without the ability to run randomized control trials (with many permutations of the variables at hand) we can never be truly sure which routines are productive and which aren’t. The best we can do is to try bundles of them all together and try to figure out which ones are especially good or bad.

So culture, an inherently persistent thing, will pick up all sorts of good and bad habits, but it will gradually plod on, adapting to an ever-changing, ever evolving ecosystem of competing and cooperating cultures.

So should we still teach our girls to be scared? I’d argue no.* Economics tells us that being awesome is great, but in a free society** it’s also great when other people are awesome. Those awesome people cure diseases and make art. They give you life and make life worth living.

Bringing women and minorities into the workplace has been a boon for productivity and therefore wealth (not without problems, but that’s how it goes). Empowering women in particular, will be a boon for the frontiers of economic, scientific, technical, and cultural evolution to the extent women are able to share new view points and different ways of thinking.

And therein lies the rub… treating girls like boys empowers them, but also changes them. So how do we navigate this tension? The only tool the universe has given us to explore a range of possibilities we cannot comprehend in its entirety: trial and error.

We can’t run controlled experiments, so we need to run uncontrolled experiments. And we need to try many things quickly. How quickly depends on a lot of things and few trials will be done “right.” But with a broader context of freedom and a culture of inquiry, our knowledge can grow while our culture is enriched. I think it’s worth making the bet that brave women will make that reality better.

* But also, besides what I think, if I told parents how to act… if I made all of them follow my sensible advice, I’d be denying diversity of thought to future generations. That diversity is an essential ingredient, both because it allows greater differences in comparative advantage, but also because it allows more novel combinations of ideas for greater potential innovation in the future.

** And here’s the real big question: “What does it mean for a society to be free?” In the case of culture it’s pretty easy to say we want free speech, but it runs up against boundaries when you start exploring the issue. And with billions of people and hundreds (hopefully thousands) of years we’re looking at a thousand-monkey’s scenario on steroids… and that pill from Flowers for Algernon.

There’s copyright which makes it harder to stand on the shoulders of giants, but might be justified if it helps make free speech an economically sustainable reality. There’s the issue of yelling “Fire!” in a crowded theater, and the question of how far that restriction can be stretched before political dissent is being restricted. We might not know where the line should be drawn, but given enough time we know that someone will cross it.

And the issue goes into due process and business regulation, and any area of governance at all. We can’t be free to harm others, but some harms are weird and counter-intuitive. If businesses can’t harm one another through competition then our economy would have a hard time growing at all. Efficiency would grow only slowly tying up resources and preventing innovation. Just as there’s an inherent tension in the idea of freedom between permissiveness and protection, there’s a similar tension in the interdependence of cooperation and competition for any but the very smallest groups.

What sort of discipline is women’s studies?

Some of the central tenets of women’s studies — and gender or multicultural studies — of patriarchy, intersectional oppression and social constructionism are, as noticed by Toni Airaksinen, unprovable and unfalsifiable. (We’ve had some discussion of Popperian falsifiability elsewhere; maybe this is another opportunity.) Social constructionism, I would argue, stands as a legitimate scientific theory: it can be either confirmed or refuted by biological evidence (Cf. John Dupré, Ian Hacking, Nancy Cartwright, etc.). The other two tenets, however, don’t work with the dominant model of scientific hypotheses, and don’t fit nicely as philosophical, sociological or political theories either. If they are considered philosophical theories, it has to be recognized that they began with their conclusions as premises; ergo, they are circular, and only confirmed by circularity. Neither conjecture has even the loose falsifiability to belong to a social science like sociology, and their refutation (were it possible) would mean the closing of their scientific branch, so they cannot be (relevant) sociological theories. Finally, very few theories that fall under the branch of “political” are fundamentally political; usually, they begin in another, more atomic field and are only secondarily responsive to the political realm. So, calling them political theories begs the question. It makes the most sense to classify theories like patriarchy as quasi-theological conjectures instead of philosophical, sociological or political ones.

To demonstrate the point: firstly, schools like these posit an original sin: some of us are born with privilege, and only through reparations or race/gender-denunciations can we overcome it. They also, again like Christianity, possess a disdain for the current, real state of things: where Christians posit a celestial heaven for the afterlife, progressive idealists embrace utopian visions materially impossible to accomplish (whether through problems with central planning or otherwise), or at least humanly unrealistic. To fuel the utopianism, historicism or a disregard for enlightened economic, historical or sociological analysis comes with the politics. Another tenet of religion is its typical weak exclusivism (van Inwagen, 2010): religions take themselves to be logically inconsistent with other sects (that is, if two belief systems are logically consistent, one is not a religion), and hold that, for people in the typical epistemic state of its adherents, it is rational to accept that religion. This mild exclusivism is very obvious for movements like third-wave feminism, so far from Steinem; it is also easy to see that stronger exclusivism not only follows from weak, but is applicable to the leftist ideologies as well: proponents of a religion must find opponents that possess the same epistemic certifications to be irrational. Also, the same exceptionalism, and infiltration into politics, is familiar to religions (like Christianity and Islam) as well as feminist theorists that seek to distort the law into beneficial means, beyond its legitimate jurisdiction.

Finally, Ludwig Feuerbach wrote in the 1840’s that theology was truly anthropology: Christianity was an appraisal of man, and the story of mankind. Gender studies sees this reversed: what might euphemistically be termed social science or anthropology, sociology, etc. is discovered to be instead a new sort of theology. Facts are subordinate to belief and orthodoxical obedience, and the probing essence of reason is dismissed for the docile, hospitable nature of faith. It seeks to see God, or masculinist oppression, in everything. This is another instance of its discontent for anything formerly satisfying; until the tenets of women’s studies are part of mandatory classroom cirricula, its students will consider themselves forever oppressed. Creationism’s proponents wrestled fruitlessly as evolution replaced their faith in American middle schools. Feminists will try tirelessly to invade grade school as well, until faith can again triumph over critique.

The selfish meme

There are two senses in which to consider the phrase.

  1. The sense in which memes enter or exit our minds.
  2. The sorts of behavior encouraged by our memes.

For those who don’t know what I’m talking about:

A meme (/ˈmm/ meem)[1] is “an idea, behavior, or style that spreads from person to person within a culture”.

Richard Dawkins introduced the idea in his famous book: The Selfish Gene. The bulk of his book discusses examining the gene as the basic unit of analysis in evolutionary studies. He introduces the idea of the meme as a different form of replicator. Both genes and memes will only be reflected in the outcome of biological and cultural evolution if they exhibit fitness–if they are able to survive.

So the cultural traditions that helped hunter-gatherer societies survive droughts or harsh winters tended to survive and spread. Over time a culture accumulates this sort of practical, tacit knowledge. (Side note: this week’s Econtalk has Cesar Hidalgo who does really interesting work trying to indirectly measure the presence of such tacit knowledge in market economies.) And if culture is made up of memes, the same way organisms are made up of genes.

Looking at genes as the unit of analysis (as opposed to the organism) explains some otherwise mysterious behavior. It provides a plausible explanation of altruism: we care for our children more than anyone because 50% of their genes are our genes. A nephew is still precious, but not as important to us because the expected ratio of shared genes between the two of you is 25%. A gene that prompts you to protect your children is likely to survive longer than a gene that doesn’t prompt you. (And genes that hang around with such kin-protects genes are also more fit than their competition.) A gene that prompted you to be kind to neighbors makes sense when you live in small groups. But a gene that prompted you to be kind to total strangers might be a liability in a world where strangers were dangerous.

Cultural evolution certainly makes sense as a gradual mutation of different cultural practices merging together to make what is called (and perceived) as a unique body of culture. It’s a complex of knowledge, ideas and basic assumptions, social interface protocols, and it’s deeply embedded in how we engage in the world. (Perhaps we can’t remember our infancy because we didn’t have a cultural lens through which to reference anything to anything else…) One thing that I’m sure we’ve all noticed is that it can be almost painful to have to reject a cherished belief. It’s even difficult to see one of these memes challenged.

Now genes don’t have to be small bits of genetic code. They can be something simple like “make this enzyme when you get a chance.” But as a unit of replication, you should consider the smallest discrete chunk of genetic coding that replicates itself. If a particular pattern isn’t fit, it will leave the gene pool, while the fit collections of genetic instruction spread. So you might end up with long complex strings of genetic material akin to a computer program. Initially simple scripts might gather as successful collections of genes that work well together. The “produce stomach acid” gene works well with the “produce a stomach” gene and soon the two are virtually inseparable. They’ve become a simple script: “Do this, then that, then maybe this other thing.” Scripts gather into multi-cellular organisms with different functions that can respond differently to different stimuli. Soon you’ve got a complex set of code as your replicating unit.

More complex genes are necessary to prompt more complex behavior. It’s worth noting that Dawkin’s theoretical framework sometimes looks like a hyper-rational economics model. Evolutionary Stable Strategies are a Nash Equilibria that are robust to invaders. They occupy a niche and survive. But this evolution is happening in the context of increasing complexity. The system is learning*. This isn’t an instantaneous process**, but it is gradually becoming more sophisticated.

A complex gene will get bugs due to random mutation, but as long as it’s still generally fit, it will survive. And over time, more subtle and sophisticated programs identify new niches. And we get plant genes surviving by filling the “eat sunlight” niche and animals in the “eat plants” niche, and bacteria co-evolving with animals’ digestive systems.

Slowly working through this long, blind, random process genes surviving this hostile environment develop behaviors that help them flourish (the “four F’s of evolution: Fighting, fleeing, feeding, and reproduction”). Gradually they stumble into opportunities, and an important one was social behavior.

More and more complexity, round and round, until we start to get our first little bits of sentience. I’ve been watching a chinchilla hop around my apartment for a couple weeks now and I’m astonished by how much effort she puts into genuinely exploring her world. She tests objects for structural integrity and learns what she can and can’t jump on. She tests boxes with her teeth, I don’t know what for. She’s distinctly learning and not merely existing or surviving. She’s comfortable and does not know fear (I’ve seen her scare one particularly wussy cat). That sort of behavior requires a great deal of complexity which requires a great deal of genetic material.

I’m noticing as I write this that the biggest gene (i.e. discrete, replicated set of genetic code) must be that very large collection of genetic patterns that must come together in order for a one’s offspring to simply be the same species.*** I’ve heard that humans and chimps share 94 percent of our genetic material. That overlap tells me that some larger percentage than that is what makes us actually a human. The difference between any two individuals, then, must be among a very small portion of their total genetic makeup. This small portion is where genetic competition occurs in the arena of sexual reproduction.

In any case, our first memes (behaviors) seem to be transmitted biologically. Later, with more complex genes, we are able to replicate more complex behaviors. Eventually, we get complex enough to build up a sense of consciousness****.

A complex enough gene might have a subroutine that sets off an error; something like the pain our consciousness experiences when things are going poorly*. And likewise for a meme. Though more likely is that the error is being returned by our psychology. (If our genes are assembly language, our psychology is the operating system, and culture is the mess of basic programs that makeup our desktop environment.)

When we think of memes as self-replicating units, interesting questions arise: what sort of patterns will be robust to competition? Which will occupy what niches? What happens when incompatible memes come together in one mind? What sort of eusocial behaviors are possible? How much do our memes govern our behavior? (This is where nature and nurture overlap.)

Obviously one possibility is a “selfishness meme,” or a culture that hits an equilibrium of distrust. But there are many others, and how they combine matters. At this level we’re essentially asking questions about psychology, culture, and institutions. The fodder of all the social sciences comes together here. Different memes will be transmitted in different ways (which is perhaps what defines the disciplines), but any of these memes may be complex enough to have a defense mechanism that involves activating various processes (including other memes, perhaps) and perhaps making people feel anger and related emotions when someone questions our beliefs and may even push people to fight with their life for their memes.

*We’re computers, markets are computers, societies are computers, the ecosystem is a computer, Earth is just a big giant computer. It processes data and creates new data.

** The next Hayek rap should include the phrase “it’s spontaneous order, not instantaneous…”

*** I could imagine it as made up of some set of smaller genes in some complex, rather than one monolithic gene but I don’t have the language to communicate that idea concisely.

**** And it must be noted that this consciousness is built out of parts designed for the poop-and-panic machines that were our evolutionary ancestors. It’s like building a super computer* out of a truck load of Pez dispensers and a warehouse full of chainsaws. And yet, how else could it be done.

Hume and Humboldt

Divergent dichotomies are not unusual to be found in Hayek’s writings. Besides the essay “Two Types of Mind”, we have his 1945 lecture “Individualism: True and False” on the difference between the British Enlightenment and the Continental Rationalism. Grounded in Edmund Burke’s Appeal from the New to the Old Whigs, Hayek traces the origin of true individualism to Bernard Mandeville, David Hume, Josiah Tucker, Adam Ferguson, Adam Smith and Edmund Burke himself. The XIX Century adds Lord Acton and Alexis de Tocqueville to the list. On the other hand, Hayek states that Jean Jacques Rousseau exemplifies the Rationalist individualism, which postulates isolated and self-contained individuals –whereas, for the former, the individual is determined by his existence in society. The “true variant” of individualism is the notion of “subject” of Hume’s philosophy: the outcome of repetitions, expectancies and habits. Finally, Hayek concludes his lecture with the censure to the German type of individualism, rooted in Wolfgang v. Goethe and Wilhelm v. Humboldt: the individualism expressed in the original development of the personality and defended in J. S. Mill’s On Liberty.

Notwithstanding in this 1945 lecture Hayek claims that this German individualism of self-development has nothing to do with what he regards as true individualism and it is “an obstacle to the smooth working of an individualistic system,” much later, in Law, Legislation, and Liberty, he will restate his opinion on Wilhelm v. Humboldt’s legacy.

This reconsideration of the value of liberty as the development of the unique and particular character of an individual will be acknowledge not only regarding legal theory but as well in his 1976 proposal of denationalization of currency. In his late writings, Hayek will endorse the development of the originality of character as an important trait for the competition to work as a discovery process.

The key to understand his shift onto this new type of individualism is closely related to Hayek’s involvement into the ideas of cultural evolution. The “true individualism” was important to state how a society can achieve certain order. The “Humboldt’s individualism” is needed to explain the dynamic of the evolution of that order. Hume’s notion of subject is related to the ideas of integration and convergence, to how an order may emerge. Humboldt’s ideal of self-development of the unique and original character of each individual implies differentiation and divergence. These two traits are the key to the adaptation to the changes in the environment that defines the notion of blind evolution. A social and political system that assures the development of differences has keen aptitudes to survive to the changes in its environment. At the level of the “true individualism”, individuals are made of institutions, repetitions and expectancies. But at the level “Humboldt´s individualism”, successful institutions are made of differences, divergent series of facts and adaptation.

(Originally published in