Profundis: Zen and The Art of Motorcycle Maintenance

(What follows is a reposting of a few short essays I wrote for Scott Young’s bookclub in response to the perennial classic ‘Zen and The Art of Motorcycle Maintenance’):


Near the end of chapter 3 the narrator makes a number of epistemological and metaphysical claims which confused me for a long time and confuse many people still. In recent years I have resolved them to my satisfaction, and this seems like as good a place as any to elucidate my thoughts.

He writes: “The law of gravity and gravity itself did not exist before Isaac Newton”, then continues “…[w]e believe the disembodied words of Sir Isaac Newton were sitting in the middle of nowhere billions of years before he was born and that magically he discovered these words.”

This nicely demonstrates an incorrect conflation of laws and physical phenomena. Unless you’ve been snorting uncut Postmodernism fresh off the Continent you’re bound to think that gravity existed before Isaac Newton. What he did was distill gravitational observations into formulae by which to describe and predict future observations.

Gravity existed prior to these formulae just like apples existed before anyone named them.

As Alfred Korzybski put it, ‘the map is not the territory’.

Entire planets worth of error can be avoided if you keep this in mind. For example, I’ve seen Gödel’s Incompleteness Theorems cited in defense of the existence of God. The Incompleteness Theorems say, in essence, that formal systems of sufficient power to perform arithmetic or describe the properties of the natural numbers contain enough recursion to ineluctably give rise to paradoxes. There are statements which are true in these systems but which cannot be established by any algorithmic procedure.

Truth, in other words, is bigger than proof.

Put more simply GITs demonstrates that the weirdness associated with a statement like ‘this sentence is false’ is to be found at the heart of mathematics and as a consequence of its deepest nature.

But — crucially! — the limitations of GITs apply only to the formal systems themselves. They tell us nothing about a non-formal system like the universe whose behavior is captured by formal systems we invent. There is a gigantic difference between saying ‘the symbols we use to describe system A have these built-in limitations’ and saying ‘system A itself is subject to those same limitations’.

And I think Phaedrus is making a similar error.


In chapter 6 we learn that Phaedrus believed there to be two fundamental ways of viewing the world. The ‘classical’ view tends to think in terms of underlying components, processes, and interactions, whereas the romantic view thinks in terms of intuitions about immediate, surface appearances.

Below is one answer to that, expanded from a comment left earlier which is worth it’s own spot:

Coming at the classical/romantic idea from a completely antithetical direction, Ayn Rand’s Objectivist philosophy champions an aesthetic merger of the two called ‘romantic realism’. I realize that she is one of those thinkers that splits the world into fervent worshippers and rabid detractors, so I’d like to avoid getting into the whole ‘Ayn Rand debate’. It’s my belief that her claims about aesthetics can stand independently from her other philosophical positions.

Objectivism sees art as being essential to the task of concretizing man’s widest abstractions in the form of perceivable, physical objects. Artists look out at the world and choose some subset of what they see to represent in their art. Their artistic choices — and our emotional responses to those artistic choices — are guided either by an explicit philosophy or by an unarticulated ‘sense of life’. Something very deep is implied by the decision to paint a skyscraper or the decision to paint ruins, and something equally deep is implied by which of these two we find aesthetically pleasing.

As beings whose nature is conceptual we require literature, art, and music to reify our ethical and metaphysical convictions — otherwise they would remain mere abstractions with a limited influence on the actual choices we make day-to-day. By owning and repeatedly responding to a work of art we reinforce a system 1 response which harmonizes with our system 2 judgements.

With time, art becomes like a fuel one burns to keep their motor running. One can fight with more vigor when surrounded by books and paintings which remind them of how the world could and ought to be.

And say what you will about the merits of her writing, I personally find the art it inspired to be gorgeous. Sylvia Bokor and Quentin Cordain both paint in the romantic realist style and NASA’s Jet Propulsion labs just released some excellent Art Deco posters from the future which I liked enough to get framed. Nick Gaetano did a series of iconic covers for editions of “Atlas Shrugged”, “For the New Intellectual”, and “Capitalism: The Unknown Ideal”, all of which inspired the cover of my upcoming book on the STEMpunk Project.

It’s a shame that Rand’s own vitriol has prevented more exposure to the view that art has a cognitive justification grounded in man’s needs qua man. Even if you reject everything else in Objectivism her treatment of aesthetics remains fascinating, original, and profound.


In chapter 10 the narrator makes several jarring criticisms of the scientific method which, if one hasn’t ever considered them before, could very well cause intellectual vertigo and a sense of nausea.

First, we have this:

“If the purpose of the scientific method is to select from among a multitude of hypotheses, and if the number of hypotheses grows faster than the experimental method can handle, then it is clear that all hypotheses can never be tested. If all hypotheses cannot be tested, then the results of any experiment are inconclusive and the entire scientific method falls short of its goal of establishing proven knowledge.”

Let’s call this the Problem of Underdetermination (PU).

He continues:

“…[W]hat seems to be causing the number of hypotheses to grow in recent decades seems to be nothing other than scientific method itself. The more you look, the more you see.”

Let’s call this the Problem of Hypothesis Proliferation (PHP)

Finally, we are told:

“Through multiplication upon multiplication of facts, information, theories, and hypotheses, it is science itself that is leading mankind from single absolute truths to multiple, indeterminate, relative ones.”

This one we call the Problem of Scientific Learned Helplessness (SLH).

I will address the first two problems here. The third I may answer at some point in the future.

PU is a pretty standard strawman of the scientific method, and it’s surprising to see it crop up in such a significant work. Everyone knows that the purpose of science is not to establish irrefutable proven Truth (with a capital ‘T’), but instead to sift through reams of data and establish one or several hypotheses that can predict future data points. Additional criteria, like Ockham’s Razor, are used to temper the forging of hypotheses with considerations of their computational burden. (I can say more about this if necessary)

The fact that evidence *always* underdetermines hypotheses has been an acknowledged problem for as long as there has been a philosophy of science, and it crops up in algorithms (like EBL, KBIL, and ILP) which have to form their own understanding of a data set.

There isn’t an easy solution here, but there are a few things we can note. First, there are a number of ways we can constrain the space of possible hypotheses. Perhaps the most common is by making assumptions which are valid within the prevailing theoretical framework. We assume, for example, that the color of a scientist’s shoelaces doesn’t affect their observation of light from distant stars.

Do we know this for certain? No. Might we someday uncover evidence of a link between shoelaces and light beams? Sure. But without a reason to see a connection now, we assume there isn’t one, and thereby rule out some regions of hypothesis space.

Moreover, until we get to the point at which a paradigm shift is necessary we usually don’t entertain hypotheses which contradict our broader theories. General Relativity says faster-than-light travel isn’t possible, so any hypothesis which utilize FTL are ruled out a priori. If and when someone dethrones Einstein that may change, but until then we don’t concern ourselves with those regions of hypothesis space either.

Even with all this there might still be a number of possible hypotheses which make sense of a given data set. The solution, then, is to hold all of them as possibly true until more data comes in.

The brilliant Nate Soares has discussed a kind of update to science he calls ‘simplifience’. It’s essentially science with some information theory and Bayesianism thrown in. The idea is that one doesn’t hold beliefs about data, one assigns probabilities to any candidate explanations for a given phenomenon. If there are five viable explanations of, say, the Mpemba Effect, then we try to work out how likely each is on the evidence and modify when possible.

Getting Bayesian statistics to run on a human brain is tough, of course, but far easier with a digital mind. Given current trends it’s possible that software scientists will outnumber meat scientists in the future, so maybe this won’t be as much of a problem.

I believe that Phaedrus makes too much out of the PHP. Yes, it’s true that every discovery raises new questions, but I submit that it *answers* far more, such that the net result is an increase in understanding rather than a decrease.

If we hear a rustling in the bushes, there is a near-infinite set of questions we could ask ourselves: is it a human or an animal? If it’s an animal, is it a predator? If so, is it a bear? A Wolf? An alligator? Is it hungry?

Let’s say we then hear the animal barking like a dog. Okay, this discovery makes us wonder about a few additional things: is this dog hungry? Does it belong to someone nearby? Is it friendly? Does it have all its shots?

Phaedrus sees this and says, ‘See! Science doesn’t settle a damn thing!”

But while our discovery that the animal is a dog generates new queries it simultaneously collapses vast regions of possible queries which we needn’t concern ourselves with.

We don’t have to ask if the animal is a bear; we know it isn’t. We don’t have to ask if it’s an alligator (and what in the world an alligator is doing in the Rocky mountains), because we know that it isn’t. For each of these animals we could ask the same set of questions we ask about the dog: is it hungry, etc.

None of that need concern us now.

So our discovery raised ten questions, and obviated the need to ask literally thousands of others.

We have not, therefore, gotten more confused by gaining information.


In Chapter 19 the narrator begins to probe the (in)famous subject/object distinction, postulating that Quality might not only be a kind of bridge between them, but the actual phenomenon giving rise to separate perceptions of self and other in the first place.

But first he must resolve a dilemma. The two horns are: (I) if Quality is objective, then why is it that scientific instruments aren’t able to detect it? (II) if Quality is subjective, then how is it any different from being ‘just what the observer likes’?

After briefly treating (I) and failing to resolve it satisfactorily the narrator turns to (II): ‘if Quality is subjective, isn’t it just what you like?’ If we excise the word ‘just’ we are left with the question, ‘if Quality is subjective, isn’t it what you like?’, which isn’t as sinister.

The assumed problem is that your preferences emerge from a soup of irrational, contradictory impulses which means that they aren’t likely to be much guide to Quality in any useful sense.

This argument breaks down into two related ones, which the narrator dub ‘scientific materialism’ and ‘classic formalism’. They are the claim that ‘what is real is whatever is made of matter and energy and detectable’ and ‘something isn’t understood unless its understood intellectually’, respectively. Scientific materialism is relatively easy to do away with: we can’t detect the concept ‘zero’, and yet it remains objective.

I think it’s possible to formulate a reply to this. ‘Concepts’ are real things, though they don’t exist out-in-the-world the way chairs do. Instead, they are abstractions running on a neural substrate. They have realness in the sense of having a causal impact on the world because, being housed in brains, they change the way agents like humans behave. They might even be measurable, in a way: there may come a time when brain imaging technology is so advanced we can see concepts as activations in neural circuits. (I’m being a little facetious here but I think you see what I’m saying)

Leaving this aside we still have classical formalism, which is harder because it’s more forceful. All it really says is that we should not base our decisions upon our romantic surface impressions but should consider the larger context and the underlying, classical structures involved. This seems sensible enough, but cleaves Quality in two. There is now a surface Quality which appears immediately and a deeper Quality which takes time to understand. People disagree about Quality precisely because they get this wrong. Some people use their surface impressions in their evaluations of Quality and others use deeper ones, and therein lies fodder for argument.

Frankly, I don’t share the narrator’s consternation over this. I’m prepared to say that Quality just is this deeper appreciation; there are not two Qualities, only one, and people basing their Quality judgements on surface understanding are wrong.

But this requires a caveat: there are people with a tremendous amount of talent in a field like music or mathematics for whom surface impressions do seem to count as Quality detection, even though they may have little formal understanding of the classical structures below. We usually call these people ‘prodigies’, and not that much is known about how they function. For most of us, however, the relationship does hold.

With these notes in place the narrator goes on to formulate a position similar to one I’ve arrived at independently: Quality (though I didn’t call it that before) is really a phenomenon occurring at the interface between agent and world. We can illustrate this same principle with a different problematic term: Beauty (with a capital B).

Are some things Beautiful? Yes. Does the term ‘Beautiful’ resist definition? Yes. Is there enough broad agreement to suggest there is something objective underlying the concept? Yes.

How about this: if all sentient beings in the universe were to perish, would Beauty still exist? No. There would still be paint on canvasses, but Beauty presupposes an agent able to perceive the thing of Beauty. It makes no sense to speak of Beauty elsewise.

And I believe Quality is exactly the same.

Music As Communicative Medium

While listening to an audiobook on music history I was exposed to the idea that the Romantic artists believed that music was the highest form of aesthetic expression because it acted directly on the mind without the intermediary steps of words or images.

This was an idea I have been struggling to formulate, and I think it could use a little fleshing out.

Here’s where I’m at:

(1) Considered within the space of possible general communicative media (think: language, mathematics, the visual arts[1]) music is  very high BANDWIDTH. It can download emotions like despair or exultation, which are surely many gigabytes in size, directly into your consciousness.

(2) But music is a very low RESOLUTION medium for all that[1]. It can only make large transfers as complete wholes. It can’t do things like proofs[2], and even something like ‘symphonie fantastique’, which tells a story, can only do so in broad strokes and only with the help of words to communicate the narrative.


[1] I took pains to point out that my analysis is of music compared to all other information-transfer media. A friend pointed out that if it is considered solely as an emotional medium then it is both high bandwidth and high resolution.

[2] Though there is actually at least one attempt to do that; See: David Stutz.

Profiting From The Written Word

– Mentorbox is a new subscription-based service which sends customers a monthly box containing interesting books filled with study sheets, detailed notes, summaries, and the like.

– Alain De Botton’s School of Life has a bibliotherapy service in which people are guided to penetrating works of literature that grapple with whatever problems they’re currently facing. Feeling depressed? — here is a list of ten of the greatest books talking about happiness/meaning/suicide/etc. Oh, and we’re eager to help you apply those messages to your unique situation for $100/hr.

– Bill Gates famously locks himself away for two weeks in an isolated cottage to read books which he believes will add value to his business.

– I once read an article (from The Economist, I think) which opined that businesses should forego generic team-building exercises in favor of having employees read and discuss books as a way of articulatinga shared vision.

Maria Popova famously makes a living reading awesome books and sharing their lessons on how to live well.

– There are entire college curricula geared toward the Great Books. For a long time this was the way of educating a society’s elite.


Surely it should be possible to combine these business models in some way, right? You could have a monthly subscription service which sends you books and notes a la mentorbox, but maybe there could be different ‘tracks’; instead of only receiving books about productivity, you might also opt to receive books about happiness, intentionality, adventure, etc. Each month you could switch your focus depending on how you’re feeling and what your needs are. For an additional fee you could get 1-on-1 coaching, maybe even with the author if they’re still alive.

Offer a special package to businesses interested in a company reading list. Work with the CEOs to devise a company worldview and then have your professional readers build a curriculum on that basis. Have your own space for businesses wanting to do retreats — and charge $10,000 for two weeks, with unlimited individual and group coaching.


I can’t think of a better than job than ‘professional reader’.

Does Biology Drive ‘Pebble Formation’ in Ideologies?

In ‘pebble form ideologies‘ I advanced the idea that over time information topologies degrade into much smaller ‘pebble forms’ which are are recognizably religious in nature, and I stipulated that part of what drives this process are ‘religion-shaped grooves’ worn deep into mammalian nervous systems. Elsewhere a commenter challenged this notion, questioning why I invoked biology as an explanation when I might instead assume this phenomenon to be an artifact of language.

My reply is that humans exhibit a property known as ‘hyper-active agency detection (HAAD)’, which is a tendency to posit agents as the cause for unknown phenomena instead of, say, the interactions of underlying components. If this were not the case, and  language were to blame for HAAD, we might expect there to be cultures whose members don’t exhibit this property. I’m unaware of a single case in which this is true. Even in situations where a mechanistic causal relationship is obvious (i.e. ‘the wall fell down because a rock hit it’) the ultimate cause is almost always assumed to be the will of some god. This is true despite the staggering diversity of the world’s languages.

That’s why I think the reliable degradation of even secular worldviews into quasi-religious ones is rooted in biology and not (solely) in language.

The Ultrapraxists

Thanks to an invite from my good friend Jeffrey Biles I was recently able to participate in a weekend session of Sebastian Marshall’s ‘ultraworking pentathlon‘.

The pentathlon consists of five ‘cycles’, with each cycle broken into two parts: an uninterrupted 30-minute work period followed by a 10 minute break. Before each cycle you ask yourself the following questions:

  1. What do I plan on accomplishing?
  2. How will I begin?
  3. What hazards are present?
  4. What are my energy and morale like?

And after each cycle you ask yourself these questions:

  1. Did I accomplish my goal?
  2. Were there any hazards present?
  3. How will I improve for my next cycle?
  4. What are my energy and morale like?

Additionally, at the beginning of a the pentathlon you ask yourself this set of questions just once:

  1. What’s my first priority today?
  2. Why is this important for me?
  3. How will I know when I’ve finished?
  4. Any dangers present (procrastination etc.)?
  5. Estimated number of cycles required?
  6. Is my goal concrete or subjective?

And of course when you finish you debrief with this list of questions:

  1. What did I get done this session?
  2. How does this compare to my normal output?
  3. Where did I consistently get bogged down? Is this part of a pattern?
  4. What big improvement can I make in future cycles?

So each individual cycle is bracketed with before and after questions, and the whole macro-structure is bracketed with before and after questions. At first glance this may look like it would be very tedious and distracting, but once you get used to it it only requires a few seconds to complete.

Though this is just an adaptation of the familiar Pomodoro method, the questions form a metacognitive framework which confers several advantages:

First, asking questions like ‘why is this important to me’ is a great way to orient towards a task. Once I get focused it’s often not difficult to stay motivated hour-to-hour, but it can be very difficult to stay motivated day-to-day. Reminding myself of why I’ve chosen to work on a project at the start of each session helps to mitigate this problem.

Second, it encourages frequent reflection on the learning process and facilitates rapid iteration of new techniques, preserving those that work and discarding those that don’t. It’s easy to have a great idea for an improvement in your work process but to then to get so absorbed in actually working that you either forget to implement the idea or you implement it but don’t notice whether it actually works.

These are non-trivial improvements. If the mind can be thought of as a manifold with attention and motivation flowing through it like liquids, then we can think of mantras, visualization, and a host of other ‘self-improvement’ techniques as being akin to engineering depressions in the manifold towards which those liquids flow. Simply wanting to form a habit often isn’t enough, for example, so building a mantra stack around the habit is like putting a bowling ball on a trampoline: everything is more likely to drift toward the new default behavior. Ultraworking is a scaffold making this process quicker and more efficient. Good ideas can be tested and their results measured in just a few cycles while you simultaneously probe for larger regularities in your output and get actual work done. It’s great!

The development of the ultraworking pentathlon places Sebastian Marshall squarely in the company of thinkers like Cal Newport and Scott Young (hereafter MYN et al.), whose stratospheric achievement is built on the consistent application of simple, pretty-obvious-in-retrospect techniques. This contrasts with figures like Srinivasa Ramanujan, who was one of the most prodigious mathematicians to have ever lived. By his own account he would dream of Hindu Gods and Goddesses while complex mathematics unfolded before his eyes. Decades after his untimely death people are still finding uses for the theorems in his legendary notebooks.

What can a monkey like me learn from a mind like that? Not much. But I can learn a ton from MYN et al. Because, while these guys are very smart, I’m pretty sure Gods aren’t downloading math into their brains while they sleep, and yet they still manage to write books, run businesses, prove new theorems, have kids, and stay in shape.

As valuable as the Ramanujans of the world are, MYN et al. might be even more valuable. Their bread and butter consists of:

  1. ‘block off as much time as possible to work on hard problems because switching tasks is distracting’
  2. ‘Summarize concepts in your own words because then you’ll remember them better’
  3. ‘Do Pomodoros, but also ask yourself some questions before and after to stay on track’
  4. etc…

Anyone smart enough to read is smart enough to do that. This means that while Ramanujan can do mathematics that twenty other people on Earth can even understand, MYN et al. can raise the average productivity of tens of thousands of people, maybe by orders of magnitude in extreme cases.  I’m not positive that makes their net positive impact bigger than Ramanujan’s, but I wouldn’t be surprised if it were.

Any group of thinkers that important should have a name, and here’s my proposal: The Ultrapraxists. ‘Praxis’ comes from Latin and refers to ‘action’ or ‘practice’ (think: orthopraxy). I kicked around a few different ideas for this title, but since Sebastian Marshall calls his technique ‘ultraworking’ and Scott Young just published a book on ‘ultralearning’ I settled on ‘ultrapraxist’.

Read ultrapraxy and learn from it; 2017 could be the most productive year of your life!

Literary Criticism as Applied Apophenia

Growing up I had far more books than friends, and have been writing regularly since I was about seventeen. In high school I was a voracious reader of “the classics”; with the lamp on late into the night I’d turn the pages of Hemingway and Dickens, not caring to wait for the English class in which they’d be taught. Owing to some high test scores I started college studying masterpieces of world literature with more advanced students, which necessitated much in the way of paper writing and classroom debate.

So it may be a surprise to learn that I’ve never had much patience for literary criticism. Upon hearing someone say “the author is using the bridge as a metaphor to…” or “the lion’s jaw is clearly an expressive vehicle for…”, I would think to myself, how could anyone possibly know that? Yes, a bridge could be a metaphor, but it could also just, y’know, be a bridge.

Now, literary criticism is a vast field and I admit to having explored little of it. But I have had many friends who enjoy literature and film, a nontrivial fraction of which were themselves steeped in the relevant theory. In an honest effort to understand I’ve often asked them about the basis of their interpretations, but they’ve rarely provided answers which I found satisfactory.

But with time and experience I’ve learned much. This essay is an attempt to answer my younger self’s skepticism by providing two different mechanisms which can justify the literary critic’s perception of metaphorical significance.

Semi-permeable cognitive membranes

I’ve written before about the fact that human introspection is shallow and much of what’s going on between our ears must be inferred. If we envision the mind as a kind of machine then many of its components are submerged under water and can only be understood indirectly. Further, the cognitive processes utilized for things like crafting a story are not cleanly partitioned from each other.

A corollary of the foregoing is that layers of meaning and metaphor can creep into a work even if the author fails to realize this. I see two ways this could happen, the first being through what may be called “leaky empathy”.

As an author tries to model characters and situations they may themselves begin to drift into corresponding emotional states. The process of writing about a group of horribly oppressed villagers preparing to travel through the forest surrounding their town could well give rise to feelings of despair or anger, albeit probably mild versions. If so, when the author conjures up an image of the forest their brains will be more likely to produce one that is dark, caliginous, and perhaps vaguely sinister.

The setting has become a metaphor for the internal states of the characters even though the author may not be remotely aware of this dynamic.

Second, and for basically the same reason, a work might reflect an author’s convictions and knowledge despite being ostensibly unrelated to the work through what may be called “leaky concepts”.

Imagine an author has just spent a year thinking about how Communism is/isn’t the greatest/worst idea anyone has ever had. When the same author sits down to design a world and plan out a story arc, is there any serious chance they’ll be able to keep these political beliefs from influencing their depictions of kingdoms, economies, and states?

Of course many authors write with the explicit purpose of promulgating a worldview or exploring some complex theme. But even if an author fails to see the lessons implicit in their work, that does not mean that the lessons aren’t there.

Reflective patternicity 

There was supposed to be some rational explanation to justify the mumbo-jumbo. Left-hemisphere pattern-matching sub-routines amped beyond recognition; the buggy wetware that made you see faces in clouds or God’s wrath in thunderstorms, tweaked to walk some fine line between insight and pareidolia. Apparently there were fundamental insights to be harvested along that razor’s edge, patterns that only Bicamerals could distinguish from hallucination.

-Peter Watts, “Echopraxia”

Another corollary to the shallowness of human introspection is that you may be surprised by the contents of your own consciousness. Sometimes the only way to explore your mind is to twist dials until lights start coming on.

Everyone has had the experience of being unusually moved by a song they’ve heard many times before. If a loved one has just passed away, then heightened emotional sensitivity is to be expected. But this isn’t always the case; sometimes, life is progressing as normal and a snatch of conversation, the light of the sun reflected in the glass windows of a skyscraper, or a memory from childhood grabs hold of you and stops you dead in your tracks. Besides being profound and worth experiencing for their own sake these moments also hint at a range of emotional states which most people don’t realize they’re capable of.

If you’re tempted to resist my claim that you don’t know yourself as well as you believe, read through this characteristically thoughtful post from Scott Alexander. It relates the story of a boy who lived his entire life without a sense of smell and didn’t realize it until his late teens. This despite the fact that he used all sorts of olfactive expressions like saying fresh bread “smells good” or teasing his sister by telling her she stinks.

But how was he to know that his sensory experience was different from anyone else’s? He can’t borrow someone else’s nose. He can’t just open a neural command line, run ‘$ grep feelz.txt’, and get back a schematic of his perceptual apparatus, complete with a little blinking cursor in the spots where there are gaps.

Put more plainly: there are numerous facets of your own mind that you aren’t aware of, so it’s worth reading poetry, listening to new music, and going to art museums, just to see how you react. Likewise, it can be useful to try and interpret a piece of literature just to see what your brain comes up with.

This first began to dawn on me in a major way while I was living in South Korea. I had just re-read Dancing With the Gods and it came to my attention that an unusually careful and prolific neopagan scholar had taken up residence in a town not far from mine. We spent a day hiking and discussing all manner of recondite issues in philosophy and religion.

It was a blast.

Near the end I half-jokingly made a disparaging remark about tarot cards. He calmly pulled a deck from his backpack and told me he always carries it with him. During the return walk he made a compelling argument for the utility of reading cards which was rooted entirely in a secular, non-mystical understanding of human psychology.

His reasoning was that superimposing an interpretive framework over cards as they come out can yield genuinely useful information. The mental dots being connected were there all along; the cards emphatically do not provide access to knowledge of the future. But, in the same way that you can agonize for weeks over an important decision and then realize that the answer is obvious after a five minute conversation, sometimes you just need that initial spark.

This is the key point behind Scott Alexander’s essay “Random Noise is Our Most Valuable Resource“. He specifically mentions tarot cards as a source of noise which can help break us out of our mental ruts. Vivian Caethe has tried to leverage this for profit by inventing a tarot deck calibrated for aspiring authors stricken with writer’s block. Both of these are examples of outwardly-focused processes which can also usefully be turned inward.

And when viewed a certain way I think literary criticism can be a similar sort of introspective scaffolding. Whether or not you believe that the author intended the lion’s jaw as a metaphor, seeing how your brain interprets it metaphorically can be akin to performing a literary version of the Rorschach test. I imagine that, as with tarot cards, doing this long enough will yield an increasingly subtle familiarity with the folds and wrinkles of your psychology.

It’s important not to get too excited about this. Just as people can form incorrect hypotheses about physical data, they can form incorrect ones about introspective data; all the usual rationalist warnings apply. But I have come to believe that this sort of “applied apophenia” can be a tool in the arsenal of those wanting a better understanding of their phenomenological field.


Machine Ethics is Still a Hard Problem

I have now read Digital Wisdom’s essay “Yes, Virginia, We *DO* Have a Deep Understanding of Morality and Ethics” twice, and I am unable to find even one place where the authors do justice to the claims they are criticizing.

With respect to selfishness they write:
“Followers of Ayn Rand (as well as most so-called “rationalists”) try to conflate the distinction between the necessary and healthy self-interest and the sociopathic selfish.”
This is simply untrue. The heroes of Atlas Shrugged work together to bring down a corrupt and parasitic system, John Galt refuses to be made an economic dictator even though doing so would allow him limitless power, and in The Fountainhead Howard Roark financially supports his friend, a sculptor, who otherwise would be homeless and starving.

Nothing — nothing — within Objectivism, Libertarianism, or anarcho-capitalism rules out cooperation. A person’s left and right hand may voluntarily work together to wield an axe, people may voluntarily work together to construct a house, and a coalition of multi-national corporations may voluntarily work together to establish a colony on the moon. Individuals uniting in the pursuit of a goal which is too large to be attempted by any of them acting alone is wonderful, so long as no one is being forced to act against their will. The fact that people are still misunderstanding this point must be attributed to outright dishonesty.

Things do not improve from here. AI researcher Steven Omohundro’s claim that without explicit instructions to do otherwise an AI system would behave in ways reminiscent of a human psychopath is rebutted with a simple question: “What happens when everyone behaves this way?” Moreover, the AI alarmists — a demimonde of which I count myself a member — “totally miss that what makes sense in micro-economics frequently does not make sense when scaled up to macro-economics (c.f. independent actions vs. cartels in the tragedy of the commons).”

I simply have no idea what the authors think they’re demonstrating by pointing this out. Are we supposed to assume that recursively self-improving AI systems of the kind described by Omohundro in his seminal “The Basic AI Drives” will only converge on subgoals which would make sense if scaled up to a full macroeconomic system? Evidently anyone who fails to see that an AI will be Kantian is a fear-mongering Luddite.

To make the moral turpitude of the “value-alignment crowd” all the more stark, we are informed that “…speaking of slavery – note that such short-sighted and unsound methods are exactly how AI alarmists are proposing to “solve” the “AI problem”.”

Again, this is just plain false. Coherent Extrapolated Volition and Value Alignment are not about slavery, they’re about trying to write computer code which, when going through billions of rewrites by an increasingly powerful recursive system still results in a goal architecture which can be safely implemented by a superintelligence.

And therein lies the rub. Given the title of the essay, what exactly does our “deep understanding of morality and ethics” consist of? Prepare yourself, because after you read the next sentence your life will never be the same:

At essence, morality is trivially simple – make it so that we can live together.”
I know, I know. Please feel free to take a moment to regain your sense of balance and clean up the blood loss that inevitably results from having such a railroad spike of thermonuclear insight driven into your brain.

In the name of all the gods Olde, New, and Forgotten, can someone please show me where in the voluminous less wrong archives anyone says that there won’t be short natural-language sentences which encapsulate human morality?

Proponents of the thesis that human values are complex and fragile are not saying that morality can’t be summarized in a way that is comprehensible to humans. They’re saying that those summaries prove inadequate when you start trying to parse them into conceptual units which are comprehensible to machines.

To see why, let’s descend from the rarefied terrain of ethics and discuss a more trivial problem: writing code which produces the Fibonacci sequence. Any bright ten year old could accomplish this task with a simple set of instructions: “start with the numbers 0 and 1. Each additional number is the sum of the two numbers that precede it. So the sequence goes 0, 1, 1, 2, 3, 5, 8…”

But pull up a command-line interface and try typing in those instructions. Computers, you see, are really rather stupid. Each and every little detail has to be accounted for when telling them which instructions to execute and in which order. Here is one python script which produces the Fibonacci sequence:

def fib(n):

    a,b = 1,1
    fib_list = []
    for i in range(n):
        a,b = b, a+b
    return fib_list

You must explicitly store the initial values in two variables or the program won’t even start. You must build some kind of iterating data structure or the program won’t do anything at all. The values have to be updated and stored one-at-a-time or the values will appear and disappear. And if you mess something up, the program might start throwing errors, or worse, it may output a number sequence that looks correct but isn’t.

And really, this isn’t even that great of an example because the code isn’t that much longer than the natural language version and the Fibonacci sequence is pretty easy to identify. The difficulties become clearer when trying to get a car to navigate city traffic, read facial expressions, or abide by the golden rule. These are all things that can be explained to a human in five minutes because humans filter the instructions through cognitive machinery which would have to be rebuilt in an AI.

Digital Wisdom ends the article by saying that detailed rebuttals of Yudkowsky and Stuart Russell as well as a design specification for ethical agents will be published in the future. Perhaps those will be better. Based on what I’ve seen so far, I’m not particularly hopeful.