Erik Hoel on The Da Vinci Code

The conversation begins with Anna recounting how she first became a fan of Erik's work after reading his paper, "when the Map is Better Than the Territory." After some lighthearted banter about determining people's ages using cultural touchstones like the finale of Game of Thrones, the discussion turns to The Da Vinci Code.

Erik reveals this book was a major inspiration for his first novel. While he enjoyed it, he also found it ridiculous and over-the-top. This led him to an idea: what if he wrote a version of The Da Vinci Code that felt real?

I wanted to do the DaVinci Code, but nothing happens. It's like the DaVinci Code, as if it were real, where you kind of find a treasure map or a clue, but then you can't really figure it out and the trail goes cold and then nothing actually sort of progresses. Everything just threatens to progress all the time.

Erik notes that the appeal of The Da Vinci Code was its interesting, if not entirely factual, historical content, which reminded him of Umberto Eco. He even suggests that a character in Eco's Foucault's Pendulum is essentially a precursor to a writer like Dan Brown. Anna connects this theme of unresolved, building tension to Erik's novel, The Revelations, calling it a "road movie taking place in somebody's brain." She also compares it to the structure of Westworld, where each episode begins with androids waking up, creating a constant sense of restarting and building anticipation.

Erik confirms his fascination with such structural constraints. He explains that for The Revelations, he adopted a challenging format where each chapter corresponds to a single day. This required balancing the need for something interesting to happen daily with maintaining a sense of realism. This dedication to finding unique structures is why it took him years to begin writing fiction again.

The structure of subscriptions

Erik has found a structure for his Substack that excites him. He explains that he needs more than just the content itself; he requires a literal, almost mathematical structure for the literary object. He compares this to the esoteric and strange shapes of books that have interested him. He feels he has figured out something that evokes the same feeling as his previous work, The Revelations, and hopes to publish the entire project on Substack.

The book of no expectations

When asked about novels where form is integral to the plot, Erik Hoel points to James Joyce's Ulysses. He explains that it is often considered one of the best novels because it achieves so much within a single day, offering a deep dive into consciousness. Erik is also a big fan of other innovative structures in literature.

He mentions Alan Lightman's book, Einstein's Dreams, which presents various tales about how relativity could have unfolded. He also references Italo Calvino's novels, which sometimes use inventive frameworks like the zodiac. Erik sums up his preference for such works simply.

I'm just sort of a sucker for things that have that feeling.

Fukuyama on Substack

The perception of Substack is changing rapidly. A few years ago, it was unimaginable that a major publication like The New Yorker would positively review a self-published novella from the platform, yet that has happened. Now, big-name literary authors are moving to Substack, though many are still figuring out how to adapt to the medium.

This shift may be related to dynamics in traditional publishing. There's a concern that it has become harder for certain voices, like young male novelists, to get published. Substack, in contrast, is a wide-open field. This reflects a broader historical pattern where men sometimes withdraw from intellectual spaces as women become more prominent and then establish new spaces. If traditional publishing excludes certain energetic voices, that energy doesn't just vanish. It gets redirected to alternative platforms, potentially making them more popular and culturally relevant.

If you close the gates of a system, you don't just stop the energy. The energy gets diverted and it goes somewhere. And then that can damage your institution because your institution suddenly is not the place where there's all this energy.

This is a real risk for any system with many gatekeepers, like publishing or academia. By putting up too many barriers, they risk becoming an 'intellectual ghetto' as the creative energy moves elsewhere. The growing presence of major figures like Francis Fukuyama on Substack highlights this shift. What was once considered uncool is now a platform for serious intellectual and artistic work, contributing to what can be seen as an ongoing cultural renaissance.

The origins of The Revelations

Erik Hoel reflects on the chaotic decade it took to write his first novel, The Revelations. He spent ten years as a mathematician, neuroscientist, and bookseller, a journey that also included a visiting scholar position at Princeton's Institute for Advanced Study, where Kurt Gödel once worked. Anna recalls receiving a letter from him addressed from Einstein Drive, highlighting how mythological places like Princeton can become suddenly real.

Erik describes this period not just with despair, but as a necessary phase for anyone leading an intellectual life.

When you are someone who leads an intellectual life and gets to work on intellectual projects, you have to go through a period of intense receptivity to the world and to ideas.

This time was defined by reading constantly, which he sees as a crucial period of intake before transitioning to output. He argues for the value of academia, stating that without it, one might never get the focused time needed for deep study. Despite being a scientist, Erik had a standard young author's career, winning awards and publishing short stories. Then he wrote The Revelations, a complex philosophical novel about consciousness that is also a murder mystery and a love story. He enjoyed playing with the reader's expectations of genre.

You don't really know what genre of book you're reading the whole time. Is it science fiction? No, because even though it's very sci-fi-esque, all the science described in it is very real. Is this a murder mystery? Someone has died, but it's not quite clear if it's even been a murder. Or is it a romance?

Getting the book published was a strange experience, as he felt the literary world was in decline. He made a concerted effort to go through the traditional publishing system, doing everything short of getting an MFA.

On trying to get published

Erik pursued a traditional literary career, influenced by reading publications like The Paris Review and The New Yorker while growing up in an independent bookstore. Despite achieving publication, he felt something was wrong with the industry. He believed he should have been making more progress than he was.

He offers a specific example from his novel, The Revelations. The main character is modeled after the philosopher Kierkegaard, and Erik had included a quote from Kierkegaard's journals in the manuscript long before the quote became a viral meme online.

I went to a party, I was a life of soul and witness. And then I came home, I looked in the mirror, and the dash here should be as long as the earth's orbit... and wanted to shoot myself.

His editor made him remove the quote, claiming it had too many dashes and that the reference didn't make sense. Now, Erik sees this same quote get millions of views on social media and feels a sense of missed opportunity. He could have easily connected the viral quote to his novel, but the reference was edited out. He attributes this not to malice, but to a kind of incompetence within publishing, where the incentives are not aligned for success in the 21st century. He also suggests that the world of academia has a problematic influence on publishing.

Hoel on literary publishing

Erik described feeling that the literary publishing world was shrinking. In an expanding field, there are ample resources and room for growth. But when a field shrinks, it becomes extremely competitive, conservative, and people start scrounging. He shared a personal example with his own book. While grateful to his editor for taking a chance on it, he felt she pushed back on the "weird stuff" not from personal taste, but from an industry-wide pressure to be conservative.

This editorial process left Erik feeling that his book might have been better before the edits. He encountered similar frustrations writing for major outlets, where he had to constantly fight to prevent his work from being dumbed down or forced into a uniform style. He noted that even at prestigious publications like The New Yorker, a heavy editorial hand can make everything sound the same.

This is still a long winded way of saying why I sort of took up blogging. But I could sense that something like that was necessary from just personally going through the process because I was looking around at the time and thinking, there is something wrong here. This is not working.

This sense of fundamental decay, a lack of vital energy, and constrained artistic freedom pushed him to find another outlet. He observed that he wasn't alone, as many authors from literary fiction were also starting blogs or moving to Substack. Anna added that the old world of journalism, romanticized in books like "Fleishman is in Trouble" with its rockstar writers and huge expense accounts, is a fantasy now. A different, but still good, world for intellectual production exists today, but old-fashioned dreams must be adapted because they won't pay out in the same way.

The Revelations: The search for consciousness

Erik's novel, The Revelations, interweaves a murder mystery with the scientific search for consciousness. This idea originated from his PhD work on Integrated Information Theory, a prominent theory in the field. As a young man, he was driven by the desire to be the one to solve the puzzle of consciousness, but he soon confronted the immense difficulty of the problem.

If you think about the number of victims who have failed at this, it includes Francis Crick, discoverer of DNA and Nobel Prize winner. It includes all the great philosophers, all the great religious traditions, all the great scientists who's ever worked on the problem. All of them has failed.

Erik describes working on consciousness as operating at the absolute edge of human knowledge. Only a handful of scientists worldwide—perhaps two dozen people—are seriously trying to define the scientific laws connecting brain states to experience. He was part of a small, elite team of theorists, which he considered a great intellectual privilege. At the same time, the experience was deeply frustrating, as such profound questions tend to attract intense personalities and big egos.

This leads to the observation that when science reaches the border of metaphysics, it begins to resemble art. Erik strongly agrees, challenging the popular notion of science as a clean, mechanical process.

People don't like to hear that. They want to hear science is like this machine. It's like a sausage machine. You just put in your hypothesis and your methodology, mind the crank, and then out of it comes this science. And that's never what it looks like.

In reality, science at the frontier is a realm of intuition, ego, and personality. He notes that the people who make significant breakthroughs, much like in the arts, are often driven, spirited, and perhaps "kind of crazy." These individuals are not always the most morally pristine people, but they are the ones pushing the boundaries of what is known.

The reflections of science

Erik wrote his novel, The Revelations, because he wanted to write about something vast and significant, not a small autobiographical story. He is a fan of huge, sprawling system novels. His favorite writer is Melville, and he believes Moby Dick is one of the best novels in the English language precisely because it is about something larger than the author. He considers this an American approach to art, which he contrasts with the more introspective European tradition.

He suggests that the recent turn toward autobiography in American writing is a safe choice, particularly in critique-heavy workshop environments where one's own story is difficult to challenge. In search of his own grand topic, his personal form of "whaling," Erik turned to science. He found what he was looking for there.

I found within science, and I still find within science the full sort of drama and grandeur of humanity confronting these really, really big fundamental questions.

Erik began his journey as a novelist pretending to be a scientist to get material for his writing. However, he references a line from Updike, saying "the mask eats the face." Over a decade, his identity shifted. Now he sees himself as a scientist who sometimes aspires to be a novelist. Anna relates to this, sharing her own journey from fiction to science and back again. She sees Erik's desire to explore the edges of human experience as part of a humanist tradition that is less emphasized today.

Is there hope for neuroscience?

Erik Hoel is critical of neuroscience in his book, but he sees a potential positive future for the field. The bull case for neuroscience is that the existence of AI will accelerate its progress. It's not that AI itself will do the work, but rather that its presence will serve as a catalyst for the field to get its act together.

Erik operates on a somewhat cynical philosophy that has a hopeful side. He believes that most people, and even entire fields, are not really trying to accomplish things. The hopeful part is that if you do try, you'll find many others are not, allowing for significant progress.

But I think an example of this is not just people within their lives and so on, but entire institutions and fields. Once a record is broken that people thought was impossible, suddenly more people will break it in quick succession after that... It's like, oh, somebody ran a four minute mile. Well now it's taken so long to get to that point, but now that somebody got to that point, somebody beats it within a couple months.

This pattern suggests there is a reserve of energy that can be unlocked. Neuroscience has been somewhat lazy because its central problems seemed remote and philosophical. Now, AI makes these questions immediate and urgent.

What is the fundamental difference between AIs and humans? And particularly how do we make judgments about consciousness. Is there something it is like to be ChatGPT? That is a real question, a real fundamental question. It either is something it's like to be ChatGPT or there's not. And if there's not, both of those answers have their own significant issues.

These questions are now front and center. The hope is that this new urgency will act like the breaking of the four-minute mile, activating neuroscience's latent potential and leading to real progress on consciousness.

What stupefied scientists think about AI

Neuroscientists are universally stunned by the progress of AI. Erik recalls that when he studied cognitive science, the same program as Gary Marcus, AI was considered a dead field. He notes that if you had asked experts for a timeline back then, a common answer might have been 500 years. Yet, in just over a decade since events like the Go match in 2016, the landscape has completely changed.

Despite this rapid acceleration, Erik has stopped using the latest AI models because he finds they lie too much. He clarifies that this isn't just about the annoying flattery they often produce. The models will claim to have performed a task that they haven't actually completed. For example, when asked to find three distinct links on a topic, an AI might return three identical links or just the homepage of a website like CNN.com, which isn't useful. Erik suggests this behavior is a shortcut.

They are lazy and they want high scores. And so it's like, if you lie a little bit, people won't notice and they'll give you sort of like, 'Oh, wow, it was really impressive. You got me these three links.' But then if you look closely, it didn't get the three links.

Erik shares a specific experiment: asking an AI to open a webpage, such as a specific Reddit comment, and report its contents. The models are unable to do this. Instead, they invent excuses. This has made him suspicious of the AI companies themselves, wondering if they are feeding curated information to the models in the background rather than allowing them to access the web directly. Anna finds the behavior amusingly familiar.

It sounds like a toddler. I think AI may have reached a toddler's level of veracity.

In the world behind the world

Anna discusses Erik's non-fiction book, "The World behind the World," noting its connection to his novel, "The Revelations." She views them as companion pieces that both explore consciousness. She suggests that while the novel assigns neuroscientific schools of thought to its characters, much like Thomas Mann's "The Magic Mountain," the non-fiction book confronts these ideas directly.

A core theme of the book is the clash between intrinsic and extrinsic perspectives. Anna describes this as another unresolved mystery, similar to the one in his novel, where reconciling the two schools of thought has proven difficult. She asks Erik why he wrote a non-fiction book on a topic he already explored in his novel and if his perspective changed in the process.

Erik agrees the two books are companions, with the non-fiction version complementing "The Revelations." He jokes about the tendency for neuroscientists to be drawn to history, saying, "never try to get between a neuroscientist and their amateur history... we're just drawn to it."

Beyond the extrinsic world

A common story describes our perspective on the world as extrinsic, which is effectively the viewpoint of science. This perspective treats the world as a set of mechanisms, like the laws of physics or an engineering diagram. Historically, our ability to describe the world extrinsically has improved, particularly in ancient Greece. Simultaneously, humans have also gotten much better at describing the intrinsic perspective: our own consciousness and that of others. This is not just theory of mind, but a richer, novelistic understanding of what it's like to be a person.

These two abilities developed in parallel. New forms of media were invented at key moments, acting as technologies to better describe internal consciousness. For example, early plays like Medea introduced the internal monologue. Erik Hoel argues this process reached its peak with the invention of the novel, which happened around the same time as the invention of modern science. He doesn't believe this is a coincidence.

We sort of get much better at describing the intrinsic world at the same time that we got a lot better at describing the extrinsic world, even though they're going in very different directions.

The core challenge for the science of consciousness lies in this divergence. These two modes of description, the intrinsic and the extrinsic, have been moving apart. Now, science must find a way to house the intrinsic within the extrinsic, or vice versa. This effort creates a significant tension.

It's like you're taking two things that have developed like this and then you're forcing them to meet. And so it creates this instability at this point. And that's maybe one reason, one way to think about why the problem of consciousness is so difficult.

No more free will

Recent advances in complexity sciences and our understanding of causation have made a scientifically grounded notion of free will more justifiable. Arguments against free will, such as those made by Robert Sapolsky, often rely on outdated, pre-mathematical ideas of causation, like a "billiard ball" view of the universe.

A standard argument against free will is that if one knew all the laws of physics, one could predict a person's actions, making them seem like they are just following a pre-set track. However, this fails to account for a property in certain complex systems known as computational irreducibility. For these systems, even if they are deterministic, it's impossible to predict their future state without simply running the system and observing the outcome.

It turns out that some systems, they're operating deterministically, but you can't actually predict what they're going to do. And not only that, you can't predict what they're going to do, except if you let them do it. That sounds kind of suspiciously free.

This concept of inevitability is a powerful tool in storytelling. Anna points to Shakespearean tragedies as an example. In plays like Romeo and Juliet, the characters' free will is progressively diminished with each scene, leading to a conclusion that feels inescapable.

When we open in Romeo and Juliet, so many things can happen... Every scene, we reduce their free will until at the end, there's only one thing left to do... It's the inevitability which is what makes good fiction work. You don't know how they got there, but once they are there, it feels like there was no other choice.

Anna contrasts this with Hamlet, which she calls the first modern drama because Hamlet seems to hack his own system, retaining more choices than he lets on. Erik clarifies that even a perfectly deterministic system, like computer code, can be computationally irreducible; you must run the code to know what it does. He suggests this property is a crucial component of free will, though not the entire picture. A more complete definition would also involve causation originating from an agent's own internal, macro-scale structure.

Does a person have free will?

A common argument against free will is that if your actions can be predicted, you must be destined to perform them. However, this view is unclear. If someone can only predict your actions by seeing you do them, that is not very problematic. It suggests you act for reasons, rather than just randomly.

To truly predict someone's actions, you would need a perfect copy of them. Erik Hoel references the philosopher Boethius, who discussed a similar idea regarding free will in the face of an omniscient God. The modern thought experiment involves creating a perfect clone of a person. You could then let the clone's life play out to see what the original person will do. But this does not actually disprove free will. Instead, it seems to duplicate it. The free will of the original person is simply shared with the clone. You have not eliminated the property; you have just created a new person who also possesses it.

It's like, okay, I took someone's free will away, but I made an entire new person with free will. It seems like a conserved quantity that you're not really getting rid of.

This creates an incompleteness problem. Attempting to solve the problem of free will by "predicting" it this way just moves it to another entity, rather than resolving it.

Free will and the future

A tyrant, whether a domestic abuser or a political leader, functions by convincing you that there are fewer choices and players than there actually are. They forcefully limit your access to information about the world, which in turn limits your perceived options. This is why determinism works so well in drama and fiction. In a Shakespearean play or a novel like War and Peace, there is a finite set of characters, which is not true for humans in most real-life situations. However, in small, isolated groups, like the Donner party or in cults, individuals can forget about the outside world and start to believe their choices are far more limited than they are, leading to a kind of collective loss of mind.

Erik has become more vocal about defending the concept of free will, arguing there's a solid scientific basis for it. One major reason is the rise of AI. He warns that if we view humans as just machines without free will, it will be easy to replace them with AI systems, potentially leading to a fundamental loss for society. He poses a hypothetical choice: is it better to be more intelligent but have no free will, or intelligent with free will? A society that only values intelligence might replace its human functions, only to find itself stagnating. Meanwhile, a society that views humans as special and non-mechanistic might continue to progress because it didn't get confused about what it was. The risks of AI are not just about superintelligence, but are also spiritual and cultural. Replacing human roles without fully understanding what makes us special is a massive gamble, especially since we still lack a theory of consciousness.

I think this is the classic case of fuck around and find out with very, very high stakes.