Links and thoughts (May 2022)
A smattering of readings on AI-art, the brain, dating, psychosis, and time
I’ve been thinking about purpose as a verb rather than a noun. As something that emerges out of your actions day to day rather than being decided on ahead of time.
I like to think of myself as someone who reads a lot, but that hasn’t been as true recently as I’d like. I was that guy who read 40 books in 2020, but then I read just 12 in 2021, and halfway through 2022 I’ve read 4. It’s become clear to me over time that a big part of my wellbeing is getting the chance to understand something new every day. I’ve been trying to make that feel like my purpose again.
In that spirit, here are some readings and ideas that I’ve found interesting recently.
AI-art isn’t art by Erik Hoel
Erik Hoel argues that art generated by AI is not real art because no conscious beings created it. Art is ultimately about expression—a connection between two points of consciousness. Per Tolstoy, “the aim of works of art is to infect people with the emotion the artist has experienced.” If you find a potato that just happens to look like a face, that's not art because no emotion or intention went into the making of the potato.
It wasn’t until reading this piece that I realized I’d always held a deflationary view of art, as Hoel calls it. If you’d asked me what art was before this piece, I’d have said: anything that someone looks at and finds beautiful. But then “beautiful” and “art” become synonyms:
Without taking into account the consciousness of the artist, the word “art” loses all meaning, becoming merely a synonym for “beautiful.” We may find something pretty, or interesting, or striking, or pleasing, but none of these mean that it is art. Without the intentionality of the artist taken into account, the definition of “art” is bled of all meaning, all differentiation, all usefulness as a term. This is why deflationary theories of art, like how art is “whatever is in an art gallery” or “whatever anyone says” or “whatever pleases the senses” are all unsatisfying as definitions, for they strip the word “art” of all capacity to do the job of a word, which is to differentiate.
The job of a word is to differentiate; art and beauty are differentiated by the fact that art requires intentionality, while beauty does not.
A commenter makes a good counterargument that AI-generated art actually does involve human consciousness, making it no less real than art made by humans:
If you actually dive into the creation of Dalle-2 — the crude attempt of apes to recreate the part of themselves that conceives, the way the neural nets are forced into desperate (mathematical) free-fall through millions of human-labelled images — there is in fact intrinsic substance behind Dalle. Pollock drips, cavemen place a hand on the wall, Pistoletto smashes mirrors, and AI researchers force mathematical constructs to reverse the diffusion of pictures. You see Dalle as devoid of humans, like a rock, but I see it as fiercely, farcically, tragically, human, like a cathedral.
Interesting to think about “life stories” as an object of study — how do people frame their life stories? What are common patterns and themes? How does it differ for people with mental illness? How do life stories correlate with well-being?
People with schizophrenia tend to have disjointed, incoherent life stories. But over time with certain cognitive therapies, they rebuild a life story centered on redemption and agency. Very uplifting stuff.
The agony of eros by Ava
A good relationship is about good communication!!
When I was in my late teens or early 20s I would sometimes be in the early stages of dating someone and feel like there was a glass wall between us, I just didn’t really get them, I didn’t know what they wanted from me, and it was so hard for me to have any clarity about who they really are. But I would be like, well, objectively this person is hot and smart and cool. So why wouldn’t I want to date them? Answer: because you can’t really talk to them, dumbass.
Sometimes you like someone and you can’t tell whether they like you back. This article made me realize that it’s not attractive or cool or mysterious that you don't know what they're feeling. It's literally just stupid.
The inability to communicate is not an indication of some flaw of yours or some hidden incompatibility that you can't figure out. The inability to communicate literally is the incompatibility.
Deconstructing Dzogchen by Christofer Lövgren and Jake Orthwein
Despite having become acquainted with Buddhism and meditation for a few years now, I only just learned from this podcast the categorization of different philosophies/practices:
Sutric view: renunciation is the goal. attachments are the source of suffering. we must drop attachments.
Tantric view: renunciation is not the goal, transformation is the goal. attachment is not the problem. you’re supposed to be “in and of” the world.
Dzogchen: there is nothing to be transformed. the mind is perfect as is. (per Jake and Christofer, this view is technically true, but not helpful for someone just starting out.)
Jake and Christofer land on a helpful definition of the elusive concept of emptiness. To the untrained mind, our conscious experience seems to have this solidity that you can grasp; emptiness is about the un-graspability of experience. It's like trying to grasp air.
In this sense emptiness seems very analogous to the nebulosity stuff that David Chapman talks about. Nebulosity is why we can't "pin down" categories in reality, which is a kind of grasping.
Meditation teacher Shinzen Young, when asked by a student, "what if there are things I don't want to let go of? e.g. projects, my marriage, my kids”, responded: “the only thing we are ever called to let go of is our present moment sensory experience.”
One of the appeals of rationalism—of believing that the truth and the categories of the world are definite—is that it enables control and prediction.
The less you cling to free will, the more agency you have.
Nomen est omen by Roger’s Bacon
Nominative determinism is the view that “people that people tend to gravitate towards areas of work that fit their names”—consider e.g. Usain Bolt (the fastest human ever), or William Wordsworth (poet), or William Headline (journalist at CNN).
I agree with Roger’s Bacon conclusion that the effect of one’s name, if it does exist, is small at most. Your name is not your destiny. But I was struck by his insight that our practice of assigning permanent names at birth is itself a cultural practice, not something to be taken for granted:
The virtually universal practice of assigning a permanent name at birth (which only exists so that governments can more easily tax us) has caused a species-wide shift towards a more narcissistic and egotistical mode of being.
Maybe this goes too far (or not far enough…), but the more general idea I’m trying to raise is that personal nomenclature is never value neutral and never without cultural and psychological ramifications. In other words, we shouldn’t be surprised that our culture has become sterile, stagnant, and individualistic to a fault when those values are embedded in the way we name ourselves.
Ideas aren’t getting harder to find and anyone who tells you otherwise is a coward and I will fight them by Adam Mastroianni
There’s a widespread meme out there that we’re nearing the end of our discovery of new knowledge and it’s gonna get harder and harder to come up with new ideas. That there’s a small set of true or useful theories out there and we’ve already encountered the vast majority of them.
Mastroianni makes many strong arguments against this view:
We’ve believed this many times before (e.g. in the 1890’s leading physicists believed that physics was “basically over”, before it was revolutionized 30 years later by Einstein).
New ideas always look easy in retrospect but were actually hard to come up with at the time (e.g. Darwin’s theory of evolution).
Discoveries beget more discoveries. Ideas compose a graph, not a fixed linear sequence of discoveries. Each new idea opens the door to more possible ideas.
New ideas compress and synthesize previous ideas. You don’t need to learn everything from scratch to understand the modern theories of a field.
I’m puzzled by the claim that scientists must labor under an ever-increasing burden of knowledge. The author of that paper writes: “If one is to stand on the shoulders of giants, one must first climb up their backs, and the greater the body of knowledge, the harder this climb becomes.” This suggests that if you peek into PhD programs, you’ll see lots of students bent over their books, desperately trying to learn everything that’s come before so they can start their own projects. “I can’t do any physics yet,” you might hear them lament. "I’m only up to Huygens!” Instead, you’ll see PhD students doing original research from Day 1—and often long before. Indeed, many students start doing more interesting work once they stop looking at lots of previous work, as it finally frees them from imitating other people and searching for “gaps in the literature,” two strategies that are unlikely to yield anything interesting.
The Pseudo-Time Arrow by Andrés Gómez-Emilsson
Andrés Gómez-Emilsson has spent years putting together a formal theory of the structure of conscious experiences. In this piece he describes how our conscious perception of time (phenomenal time) may be related to physical time (the ticking of the universe’s clock).
He contrasts two philosophies of time: presentism (the view that only the present “actually exists” at a given moment, that the future and past are abstractions) and eternalism (the view that all moments of time exist equally, and the present only looks like special because of our limited conscious perspective).
While most of our theories of physics seem to suggest eternalism (in general relativity, time is just an additional dimension on an existing 4-dimensional spacetime thingy), our conscious perception strongly suggests presentism.
Gómez-Emilsson paints one theory for how our conscious experience of time is created: our sensory systems take a bunch of snapshots in physical time, and each individual “moment” of experience is really a bunch of these snapshots superimposed on each other, with the most recent snapshot having the highest weight. (Kind of like Apple’s Live Photos.)
When time feels slow, what’s happening is that a larger number of snapshots are being blended together. The “after-images” of recent inputs last longer. When time feels fast, the opposite is happening.
Much of neuroscience over the past few decades has consisted of map-making: charting out the connections between regions in the brain and figuring out which parts are responsible for what. But a map is not an explanation. Creating detailed maps is just the first step towards coming up with an actual theory of how our brain works.
Case in point: the C. elegans nervous system has been mapped out completely—we know every connection between all 302 neurons the worm has. And yet, we still don’t understand how any of this generates the worm’s behavior. (For scale, humans have ~100 billion neurons.)
David Poeppel is interested in language specifically. How does our brain parse the unstructured sound waves that enter our ears into structured, grammatical, meaningful communication? We still have no clue, but Poeppel has conducted an interesting experiment showing that brain waves are coupled with the high-level rhythms of speech.
Neuroscientists have known that your brain “breathes”—we can measure (with an EEG) rhythmic patterns of neural firing that span large swathes of the cortex. These waves occur at a number of different frequencies: some are faster, some are slower. What Poeppel has found is that there are brain waves that seem to match the cadence of incoming speech, at multiple hierarchical levels:
there are brain waves that are coupled to the individual syllables we hear
there are other brain waves coupled to the word phrases (e.g. “dry fur”, “big cat”)
and then there are brain waves that are coupled to entire sentences (e.g. “dry fur rubs skin”)
They ran a number of different experiments to confirm that the brain waves are actually coupled to the structure of the speech rather than just the rhythms in the sound itself. The details are in the lecture. The takeaway, though, is that our brain must be generating and computing over abstract structures that aren’t there in the raw stimulus. (Assuming that our interpretations of the brain wave data are correct, of course.)
There’s a higher-level question here about how the brain works: does the brain build discrete abstract units, that interact with one another based on rules? Or is it like a giant machine learning system, with continuously shifting weights, inferring statistical relationships between inputs and outputs? (Are these mutually exclusive?) Poeppel’s experiment is a nudge in the former direction: the brain does create abstract structures that are not just derivative of statistical relationships in the input. Poeppel’s hunch is that the brain is materially different (no pun intended) from modern deep learning systems:
Remember that humans can learn things on experimental trial number one, learn complex ideas from sparse data, extrapolate insights that were never in the training data, and say things that they’ve never heard before. And remember that humans have generativity and compositionality. So scientists working on human cognition might look over and see what can be learned from huge statistical models, and we do learn things from those statistical models, but it simply doesn’t follow that that’s somehow the answer to how human brains are organized.
Time will tell. When I look at the state of modern neuroscience I can’t help but think that there is so much understanding ahead of us; we’re where Newton was with physics, or Hooke with biology. This is just the beginning.