Eigenyou Theory

We've talked a lot in the last few years about how software is eating the world, but one thing we haven't really talked about is how software eats you, and I mean you specifically. I don't mean robots taking your job or humans being generally transformed by iPhones. I mean how, in a medium-is-the-message way, the digital environment changes you and makes you a different and, as I will argue, more characteristic version of yourself. And less of an interchangeable part.

Here's one way to think about it. A book is a versatile object. It can be used as a doorstop, torn up to make cones in which to sell snacks, for origami artwork, as kindling for a fire, hollowed out to make a box for holding secret spy stuff, or to make papier-mâché artwork. But there is one use of a book that is characteristic: reading it. This characteristic use, reflected in the intentional arrangement of the atoms that constitute it, is only revealed in a literate society of the right sort, and only to specific readers with the right mindset. Reading is a characteristic use of a book, one that reflects the natural internal coupling between the how and why of its nature. When you read a book, means and ends are in harmony. There's a wonderful German prefix that helps to tag this sort of thing, eigen-, which means something like "characteristic." You could say that reading is the eigenuse of a book.

Humans are not books of course. We do not have fixed, static eigenuses. But there is something it is like to be each of us uniquely. There is an eigenyou, revealed when software eats you. Why? Because when software eats the environment and makes it programmable, you can project your own reality distortion field onto it, reshaping it to be better suited to you. But this means a "real" inner you has to come into being and manifest itself. And therein lies a difficult personal growth challenge. Because the eigenyou is something you have to work to grow into. It's not something latent in your genes that simply expresses itself automatically in some lucky circumstances.

Your eigenvalues and eigenvectors

1/ The easiest way to understand the idea of an eigenyou is via a math metaphor: A circle has no preferred, characteristic orientation, but an ellipse like in the picture, with clear long and short axes, does.

2/ If you hang out around engineers or scientists, you'll often hear them talking about eigenvalues and eigenvectors. Think of an eigenvalue as a measure along a characteristic dimension (like your height) and an eigenvector as the characteristic direction, like "vertical" for you as a biped.

3/ The math of working out eigenvalues and eigenvectors is not trivial, but suffice it to say that if you can approxrimately describe a system using what are called linear equations, you can figure out these eigenvalues and vectors.

4/ An ellipse for example, has 2 eigenvalues/vectors. The eigenvalues correspond to the lengths of the longest and shortest axis, while the eigenvectors corresponds to the directions.

5/ An ellipse is a simple geometric figure, but the idea applies to much more complex systems. For example, for a pendulum, the natural oscillations correspond to the eigenvalues.

6/ Think of the eigenthing as the thing described in terms of its natural structural, functional, and behavioral grain. To have identifiable eigentraits is to be an eigenthing, a thing with a grain. Eigenthings are things with non-trivial, non-interchangeable "identities."

7/ Now, let's apply the idea to people. People are extremely general and complex lumps of matter.  As neural meatbag computers, we share with iPhones and supercomputers the capacity for being programmed in effectively unlimited ways (we are approximate universal turing machines).

8/ But that generality of potential capability does not mean we are featureless abstract subjects, or lacking in preferred orientations/temperamental grains. The essential intuition behind eigenvalues and eigenvectors applies to humans too.

9/ Think about when a thing reveals its eigenthing nature: when structure and function are in harmony internally and in relation to the environment. When it is situated and being used in a way that is most natural for it. In an "along the grain" way.

10/ Another way of saying this is to say that means and ends are in harmony (though not necessarily in a rational way). The thing as a means embodies the ends to which it is being employed and vice versa. And therein lies an idea that comes in dumbass and smart versions.

11/ When it comes to humans, there is a strong temptation to apply this idea via what is called the naturalistic fallacy. This is how paleo lifestyles, heteronormative gender roles, and formulaic "self-actualization" scripts turn into growth-limiting, aestheticized lifestyle fetishes. This is the dumbass version.

12/ The smart version takes a few more steps, just like computing eigenvalues and eigenvectors takes a little bit more work than just measuring some obvious external dimensions (even if you're an ellipse).

13/ Humans have been shaping their environment for tens of thousands of years to make more use of their brains and less use of other parts of ourselves. You could say our brains were vastly overpowered/oversized/over-engineered for the paleo environment, and we've been fixing that.

14/ We are now getting close to an environment where our brains are the "right size" for the complexity of challenges we deal with. Some might even say we've overshot the mark and created an environment that's too complex for us to manage, maintain, and debug. That's a story for another day.

15/ Here's a dramatic way to think about it. Imagine a strange alien planet, Bookaloo, where sentient books grows on trees. The other creatures on Bookaloo are rather stupid, and use books as doorstops, kindling, and papier mache.

16/ The sentient books of Bookaloo decide this sucks, and cleverly drive evolution, by picking out a particular species of promising ape to breed to greater intelligence. Until one day, the apes get smart enough to invent language and begin reading the books. And the books live happily ever after in the culture wars that ensue among the apes.

17/ The point of this story is that there is no necessary reason for an evolved thing to be best adapted an environment in the past or not be better adapted to a discoverable/creatable future. There were great photographers in 5000 BC. Just no cameras for them to express themselves with.

18/ Of course history matters. There are elements of our nature that do reflect the paleo environment. But the point is, the eigenyou does not belong to a particular time, place, or evolutionary niche. As a specific instance of a generalist meatbag computer design, the eigenyou is best defined by a characteristic way of doing things.

19/ To understand this, we have to first stop thinking of human beings (or any general-purpose computing entity) as being fit for a particular purpose. We are not pegs in holes, round or square. That's not how "design" (evolved, not intelligent) manifests when it comes to general intelligences.

20/ In something like a hammer, structure and function are tightly, statically, and rationally coupled. It is a means for an end (banging on nails) in a closed, bounded, efficient sense. It is meaningful to think of optimizing a hammer for the nails it is supposed to bang on. If you do it well, you get a beautiful hammer, fit to its purpose. The nature of the coupling is dictated by the environment of adaptation.

21/ It is possible, but kinda stupid, to think of humans this way. In fact, that is exactly what religions would prefer you to do. That you are "instruments" of a "higher power" fit for a particular purpose. Paleo-fetishists just dispense with the divine in such functionally fixed self-understandings.

22/ A human though, has a nature that is open and unbounded. When you could be chopping wood today, solving differential equations tomorrow, and trolling people on twitter next week, to think of yourself in terms of efficiency of design and fit for purpose is not even wrong.

23/ Nor does "general purpose" mean you are like a Swiss army knife, composed of a bunch of specialized, insect-like capabilities. Your programmability means that is the wrong level to analyze structure-function coupling.

24/ In fact, thinking of ourselves in such finite means-ends ways -- the essence of utilitarian and many modern "rationalist" approaches -- is a recipe for psychological self-destruction, since, as has been long recognized by philosophers, it is a road to nihilism. Another story for another day.

25/ Traditional means-ends reasoning simply does not apply to generalist computing beings. It is meant for thinking about things for which the coupling between structure and function, means and ends, lies in the environment rather than internally.

26/ When both ends and means, rather than just means, are open to selection and design, they must be designed, or rather evolved, together. They are entangled like quantum states. Because you are never just figuring out how to do something, you're also simultaneously figuring out whether to do it at all or do something else instead.

27/ It is this internal entanglement between the how and why of a behavior, before either is fixed, that is the essential characteristic of generalist computing entities, and where you must look for your eigenyou. A panda enters a bar, eats, shoots, and leaves. You're a bit more complex. You enter a bar and get drunk to clarify means-ends entanglements.

28/ This is why I am generally skeptical of "optimal living" how-to advice, and neither accept, nor dispense it. Either offer entangled how-to/why-to ideas, or nothing. A good source of this kind of inspiration is fiction. Fictional characters offer entangled how/why prototypes for you to think about.

29/ To a first approximation, a character in a book or movie is interesting to the degree their motive is strong without being entirely legible through a rational lens. Think Ahab in Moby Dick, Tyler Durden in Fight Club, Phil Connors in Groundhog Day, or Andy Dufresne in The Shawshank Redemption.

30/ When a character is interesting, even if there’s a clear motive, it tends to be nominal (whale, create anarchy, get the girl, escape the prison) and leaves much about the behavior unexplained. You know there's much more to the character than the nominal ends towards which they are striving, and the means with which they are doing so.

31/ The Shawshank Redemption is not a story about tunneling out of prison. It is about the revelation of a pattern of how-and-why entanglement embodied by Andy DuFresne. Groundhog Day is about Phil discovering his eigenyou, not about him escaping the time loop.

32/ Notice that the behavior of such characters tends not to be "efficient" in pursuit of the nominal goal. Groundhog Day is a particularly wonderful example. Phil Connors has suicidal days and "self-improvement" days and everything in between. Exceptions like the Terminator (a robot) prove the rule.

33/ But interesting characters are not "inefficient" either. They just exist at a level of self-definition that exists beneath efficiency considerations. The behavior conforms to the grain of the personality rather than to the means-ends logic of the specific problem (which is generally under-determined enough that there is room for eigenyou-expressive behaviors).

34/ By contrast genre fiction characters efficiently pursue their nominal goals, at least within a formulaic notion of efficiency, tied to a signal virtue/value like "courage". Liam Neeson’s “particular set of skills” in Taken for example define a terminator-like robot character. Genre comedy characters merely flip it to inefficiency; turn robots into comically flawed robots.

35/ A test of whether a character is unique or generic is how they are parodied. A parody of Ahab or Andy DuFresne is recognizably derived from its source material. A parody of a genre character results in a trope like "action hero", with few/no signature traits.

36/ Occasionally, such stock genre characters rise above formulaic efficiency via what we tend to call a stylized presence, like John Wick. They go from character to abstract, but still recognizable, idea. Such characters are somewhere in between genre and literary.

37/ A strong character compresses a how+why coupled motivation into an identity. The can’t just win, they have to win in a characteristic way. They can’t just apply their skills to anything, they have to have characteristic goals. The battles they pick or don't pick reveal things about them. They can't just grow older and wiser through death/resurrection moments. They have to do so in a characteristic way.

38/ Strong characters are eigencharacters (n=1 archetypes). Driven by eigenvalues on eigenvectors, no matter what the mission. Sometimes the traits are superficial. James Bond is a superficially unique eigenpainted stock character. The Sean Connery, Roger Moore, and Daniel Craig incarnations share little beyond a taste for vodka martinis, "shaken not stirred." It's hard to tell Craig's Bond apart from, say, Jason Bourne.

39/ On the other hand Doctor Who has deeper eigenvalues we expect to see retained in each incarnation. If you think Doctor Who is necessarily a man, you don't get the character. If you think s/he is necessarily kind in problem solving, and that the ninth "War Doctor" is the real anomaly, you're starting to get it.

40/ When Jodie Whittaker takes over the role from Peter Capaldi this fall, despite the spotlight being on the gender switch (merely a new eigenpaint job), her challenge will be continuing the recognizable Eigendoctor personality.

41/ One reason Doctor Who is a better character than James Bond is that both means and ends of behaviors vary very wildly, as evident in the far wider variety of super-adversaries. Daleks, Cybermen, Weeping Angels --- it is a dizzying menagerie compared to interchangeable Bond villains in their lairs. The Doctor deals with all of them with his usual eigenflair.

42/ A dead giveaway of a cardboard-cutout character is a generic and fixed goal like “save wife/child” or "defuse bomb". This is not an eigengoal. A more subtle giveaway is a generic skill like martial arts as opposed to Holmes’ eigenskilled abductions/deductions. Adversaries are live giveaways. Professor Moriarty, unlike Bond villains, is a true evil twin to Holmes, an eigenvillain.

43/ What happens when your means and ends are not entangled? When structure and function don't evolve together? When you separate the two? You get a disease I call hedonism-prowess syndrome.

44/ Hedonism-prowess syndrome is a compartmentalized way of thinking about your nature in two pieces. A "means" way, (what you're good at, your skills, at which you can exhibit prowess irrespective of goals) and an "ends" way (what you pursue, seen purely a function of your tastes/values)

45/ Prowess is means optimized without regard to ends. Hedonism is ends valued regardless of means. Money or tribal loyalty often provide the entanglement between the two in psyches broken by hedonism-prowess syndrome. You exercise skills, earn money (or the Master's trust), spend it on stuff.

46/ In both cases, the entanglement of means and ends -- the source of meaning -- is outside of you. If money provides the entanglement, you turn into a mercenary, shareholder-value-maximizing creature of the market when society decays. When loyalty provides the entanglement, the object of your loyalty is the source of meaning, and you turn into an anomie-driven ronin when the Master dies.

47/ The result in either case is a spiraling down into the void. To entangle means and ends within yourself is to create a sort of living quantum aliveness in yourselves, resulting in a system that "lives" in a certain characteristic way and cannot be psychologically annihilated by external changes (you still die if hit by a bus though).

48/ Fun wild metaphor: in physics, one of the implications of the Heisenberg uncertainty principle is that there can be no such thing as a perfect vacuum, because that would mean you'd simultaneously know (say) the components of the electromagnetic field or positions/velocities of atoms (all zero) with perfect certainty.

49/ Let me pull back before I go too wild with that metaphor. The basic idea is that a general purpose entity with approximate universal-computing capabilities is defined by a certain way of doing things, embodied in a context-independent internal entanglement of means and ends that we call "personality."

50/ Call it style. Call it your "way". Call it the "real you" or your "inner beauty." Whatever you choose to call it, the interesting question is, what does it take for that personality to come out and play?

51/ I'll argue that the key mass-enabling condition is a software-eaten world. In a software-eaten world, nobody need be a doorstop instead of a book unless they want to. You can tinker with your environment enough to allow your eigenyou to emerge.

52/ George Bernard Shaw famously said “The reasonable man adapts himself to the world: the unreasonable one persists in trying to adapt the world to himself. Therefore all progress depends on the unreasonable man.” In a software-eaten world, everyone can be unreasonable at low cost, not just kings and queens.

53/ In fact I'd say you must do so, because prowess-hedonism syndrome is no longer a tenable way of being. All external loci of means-ends entanglement, from economies to objects of tribal loyalty, are now highly unreliable and will betray you. To paraphrase Nick Szabo, trusted third-parties are meaning holes.

54/ How do you discover your eigenyou in a software eaten world? The mechanics are well-known and I've been taking about them ad nauseum for years already: trial-and-error, experimentation, turning yourself into a key, trading compasses for gyroscopes, adopting a cyberpaleo ethic, learning to be a precious snowflake, avoiding the adjacency fallacy.

55/ But there's no how-to recipe for this (it would be very ironic if there were). Like Phil Connors in Groundhog Day, you're going to have to flail. Methodological anarchy is your friend. But there's a couple of things you can avoid that will tend to slow you down.

56/ The first trap is the finite game trap. What Clifford Geertz famously called "shallow play" as opposed to "deep play" in his classic, Notes on the Balinese Cockfight (ironically he took the idea from Jeremy Bentham, founder of utilitarianism and inflictor of the disease of means-ends decoupling upon us, leading to an epidemic of prowess-hedonism syndrome in the industrial age).

57/ In shallow play, the game plays you, and you maximize utility in some rational means-ends sense. You play the game as a finite game, the game plays you as a finite being. You play to win, and the game wins you.

58/ In deep play you play the game, and you do so to renew your sense of self, reacquaint yourself with the characteristic nature of your internal how/why entanglement, thereby figuring out how to continue playing. Or eigenplaying.

59/ This is not easy. One of the reasons I am not a reliably productive writer is that I don't have an industrial-age sense of myself in an externally defined set of roles like "writer" or "consultant." So my inhabiting of those roles is unstable.

60/ In my case, my eigenyou sometimes finds expression through writing, sometimes through other behaviors. And sometimes, it gets obscured by the fogs of inner and outer wars and fails to find expression at all. Like everybody, I go through periods of being lost and found to myself.

61/ Finding your eigenyou is not a "win" state you arrive at once and never leave. At least not if you choose to continue living in as full a way as you are able. It is at best a pattern in how you tend to lose yourself and find yourself again, so you can continue playing, continue living.

62/ And with the withering away of the industrial environment, you no longer need to stay in the boxes others put you in, be they loving family members, evil sociopaths, or apathetic institutions. You can program and project your own reality, define your own game, your eigengame. Even if you're the only one who knows you're playing it.

63/ I'll conclude with a set useless fake-math Three Laws of Eigenyou Theory because that's an eigenthing for me to do (which somebody noted is one of my more annoying traits). Incidentally, when I tweeted bits of these ideas earlier, this is where I started, and apparently nobody understood what I was getting at. Hopefully the ideas are clearer and presented in the right sequence now. (And before you ask, the h/2π in the third law is a joke reference to Heisenberg's uncertainty principle)

  1. First Law: An optimal How cancels a rational Why resulting in ∅ and insanity

  2. Second Law: A How is tasteful in proportion to the Why being irrational, ie misaligned with visible incentives

  3. Third Law: An activity is stably meaningful if tastefulness(How)*irrationality(Why) ≥ h/2π

Feel free to forward this newsletter on email and share it via the social media buttons below. You can check out the archives here. First-timers can subscribe to the newsletter here. More about me at venkateshrao.com

Check out the 20 Breaking Smart Season 1 essays for the deeper context behind this newsletter. You can follow me on Twitter @vgr

Copyright © 2018 Ribbonfarm Consulting, LLC, All rights reserved.