Something's going on.

As reported credulously and breathlessly by just about every major media outlet, there are strange lights in the night sky all over the East Coast; particularly, it seems, in areas heavy with air traffic. Residents are going on “drone hunts” to look at the heavens (a direction that humans don’t usually examine for long periods) and suddenly seeing all sorts of things they can’t explain. “Something's going on,” a New Jersey town mayor says; President-elect Trump says the same. “New Jersey drones” are being spotted as far away as California. '“Shoot them down already!” demands a NJ Assemblyman.

In response, locals are shining lasers at drones that turn out to be planes (like poor FedEx flight 3926) because it’s hard to judge distance well in the night sky. The FBI had to release a statement yesterday begging people not to shoot.

In the rare cases where people use flight trackers appropriately, like when New Jersey Senator Andy Kim went out on a drone hunt with local police, he found that what he saw was… just planes. Meanwhile, ABC News aired this picture of what is probably Venus or Jupiter.

Other prominent examples of sightings include the former governor of New Jersey mistaking the constellation of Orion for drones, or this still from a video shared by UFO activists (taken near, ahem, Newark Airport).

Demands for action has led to the FBI deploying observation teams throughout New Jersey and creating a tip line. Of ~5,000 tips received, about 100 leads were deemed even worthy of investigation, finding at most a couple hobby or law-enforcement drones. An FBI spokesperson had to gently explain their findings in the kindest possible way at a White House press briefing.

In overlaying the visual sightings reported to the FBI with approach patterns for Newark, Liberty, JFK, and LaGuardia airports, the density of reported sightings matches the approach patterns of these very busy airports with flights coming in throughout the night.

Meanwhile, downed drone searches turn up nada, and many of the most-shared reports have been unfalsifiable verbal accounts of “I saw something weird in the sky.”

So the only noteworthy thing about this is what it represents at a cultural level: the “I Saw a Mysterious Drone” craze is really the disguised evolution of the UFO craze that began seven years ago with a garbage New York Times article. It is pushed by much of the same people for much of the same reasons (right now the UFO and UAP subreddits are going absolutely bonkers over sightings).

All mass delusions and hysterias have a cause. In the summer of 1518, the year of the Dancing Plague, it was a French woman, Frau Troffea, who began to dance without music on the cobbled streets of Strasbourg in some sort of modern-esque fanatical fashion. Frau must have been very good, maybe the best improvisational dancer to ever live, because hundreds of others joined and danced until they collapsed (some claimed ~15 people died a day, although accounts are unreliable).

Subscribe now

Humans don’t change. In the 80s and 90s we had the Satanic abuse panic which broke apart thousands of households based on false confabulations. Such cases require an initial cause which prompts mimicry. It was Frau Troffea in 1518, and it was psychiatrist Lawrence Pazder who created the satanic panic with his discredited book Michelle Remembers in 1980, and now this entire phenomenon of “drones are attacking!” and blurry iPhone videos is a result of a 2017 New York Times article announcing that the government spent years secretly hunting unexplainable UFOs which buzz military bases and dance like Tinkerbell away from pilots.

So to understand the current delusion, we must go back to the beginning. For quietly over the past years it has come to light that essentially nothing about the original UFO story reported by the Times was true.

Remember this?

Well, now we know the program was a farce with government money, conducted mostly by a private company hunting ghouls and goblins and dino-beavers at “Skinwalker Ranch.” I summarized the saga as:

Diehard paranormal believers scored 22 million in Defense spending via what looks like nepotism from Harry Reid [privately, a UFO “believer”] by submitting a grant to do bland general “aerospace research” and being the sole bidder for the contract. They then reportedly used that grant, according to the head of the program, to study a myriad of paranormal phenomenon at Skinwalker Ranch including… dino-beavers. Viola! That’s how there was a “government-funded program to study UFOs.”

The best insight into how this has all operated is via the reality TV show about Skinwalker Ranch, where they regularly do things like take recordings of flies buzzing across the camera and blur them so they look like unidentified objects.

The associated government-funded paranormal “research program” was called AATIP and ended in 2012. Five years later The New York Times decided to report it as if AATIP was a real serious program headed by a man named Luis Elizondo at the Pentagon who, it is suggested in the article, had been officially compiling and investigating more credible UFO reports such as now-famous videos like FLIR, GIMBAL, and GOFAST.

Except that later, it was reported by The New York Post that:

In 2019, the Pentagon released a statement saying Elizondo had “no responsibilities” with AATIP, a program which they also said wasn’t created to investigate UFOs.

Whoops!

Public interest was (understandably) so great after the Times article that an actual government program to investigate UFO reports was memed into existence, the AARO.

Since their founding the AARO has consistently come up empty handed, over and over, because of course they have. Quietly, most major sightings has now been thoroughly explained. Pilots, it turns out, are also human beings, and one in ten thousand missions they see something weird that they can't explain, which is always just the standard stuff skeptics point out, like balloons, bags, or lanterns. According to the November testimony by current AARO director Jon Kosloski:

Among AARO’s closed cases from the May 1, 2023 to June 1, 2024 time period, about 70 precent of the UAP were deemed to have been balloons, 16 percent drones, eight percent birds, four percent satellites and two percent aircraft, according to the report.

Remember the famous “Go Fast” video?

“The Go Fast … captured the public attention and congressional attention when it was made public in 2017. [It] looks like an object flying very fast over the water, very close to the water. Through a very careful geospatial intelligence analysis, using trigonometry, we assess with high confidence that the object is not actually close to the water, but is rather closer to 13,000 feet,” Kosloski said. “As the platform is flying and capturing the object, if it is closer to the platform at a higher altitude, a trick of the eye called ‘parallax’ makes it look like the object is moving much faster.”

Using trigonometry! Why that's what skeptics did immediately when they got access to the video… because simple math led to the nominative conclusion: given a UFO video titled “Go Fast,” the object will inevitably turn out to be very slow.

Subscribe now

The first director of the AARO, Sean Kirkpatrick, recently confirmed the shady origin of this entire endeavor, along with the rank incorrectness of the 2017 Times article, in an op-ed he penned in Scientific America after leaving his post this year.

… this narrative has been simmering for years and is largely an outgrowth of a former program at the DOD’s Defense Intelligence Agency (DIA) called the Advanced Aerospace Threat Identification Program (AATIP), which was heavily influenced by a group of individuals associated with businessman and longtime ufologist Robert Bigelow, founder of Bigelow Aerospace…. after a review by the Office of the Undersecretary of Defense for Intelligence and DIA concluded that… taxpayer money was being inappropriately spent on paranormal research at Skinwalker Ranch in Utah.

His testimony completely contradicts the more serious picture painted by the 2017 Times article, confirming from the inside the truth that I and others have been pointing out for a while.

Not that people want to hear this. In fact, being sane and working at the AARO is a pretty thankless job, it turns out.

Kirkpatrick has received violent threats, social-media smear campaigns, and even had to call the FBI after a UFO fanatic showed up at his home. “I’ve had people threaten my wife and daughter, and try to break into our online accounts—far more than I ever had as the deputy director of intelligence,” Kirkpatrick says. “I didn’t have China and Russia trying to get on me as much as these people are.”

Coming up empty-handed isn’t because the AARO members are hiding anything, because by nature of the job the AARO has been staffed with UFO believers.

The “chief scientist” of this Pentagon task force was Travis Taylor, who is and was a co-star of “Ancient Aliens” on the History Channel. He currently stars on “The Secret of Skinwalker Ranch” on the same network.

Yet, now that they have to operate with a modicum of oversight, being an actual official government program, they continuously disappoint those in the outside UFO community, who get mad at them for not confirming their beliefs.

In a sane world, the AARO and its continual null results would be a cautionary tale about getting exactly what you want. Instead, its mere creation has implicitly supported the idea that something’s going on. And now the FBI has to ask people not to shoot at planes landing in Newark.

A disturbing reminder that, despite all our technology and education and literacy, we are still just dancing in Strasbourg. Only the location changes. Under the light-polluted New Jersey night sky our movements become erratic and there is sweat on our brow once again, as, iPhone in hand, we thumb record.

A recent provocation from the Times; another entry added to what’s become a recognizable and sure-to-trigger-controversy op-ed genre.

David Morris, a creative writing professor, charges headfirst into the breach:

Over the past two decades, literary fiction has become a largely female pursuit. Novels are increasingly written by women and read by women. In 2004, about half the authors on the New York Times fiction best-seller list were women and about half men; this year, the list looks to be more than three-quarters women. According to multiple reports, women readers now account for about 80 percent of fiction sales.

I see the same pattern in the creative-writing program where I’ve taught for eight years. About 60 percent of our applications come from women, and some cohorts in our program are entirely female. When I was a graduate student in a similar program about 20 years ago, the cohorts were split fairly evenly by gender. As Eamon Dolan, a vice president and executive editor at Simon & Schuster, told me recently, “the young male novelist is a rare species.”

Now, plenty of people think this genre is overblown and reactionary. It has a long pedigree—heck, you can find pieces in Publishers Weekly 15 years ago asking “Does the lack of men in publishing hurt the industry?” (more recently, check out the “The decline of the literary bloke” in The New Statesman).

What I’m personally interested in is the common meta-opinion about why this is bad, and it’s the opinion that David Morris himself articulates: that the problem with young men not reading literature is confined to young men themselves. As he writes in the Times:

In recent decades, young men have regressed educationally, emotionally and culturally. Among women matriculating at four-year public colleges, about half will graduate four years later; for men the rate is under 40 percent. This disparity surely translates to a drop-off in the number of novels young men read, as they descend deeper into video games and pornography. Young men who still exhibit curiosity about the world too often seek intellectual stimulation through figures of the “manosphere” such as Andrew Tate and Joe Rogan… I think of 2024 as the “Fight Club” election, in which disaffected guys vented their frustrations and anxieties through a brawler who will one day reveal himself to be not their hero, but rather a figment of their imagination.

(Rather ironic that Morris cites Fight Club to explain the election. If literature were so cut and dry, and its pages just made people vote Democrat, there wouldn’t be such an obvious literary reference from within to frame the entire issue!)

Subscribe now

Anyway, to the point: Morris thinks it bad for the polity if young men lack literature because…

Novels do many things. They entertain, inspire, puzzle, hypnotize. But reading fiction is also an excellent way to improve one’s emotional I.Q. Novels help us form our identities and understand our lives.

And herein lies the miss. While I do think young men should experience literature, and wish more would (let’s not oversell, it’s no panacea), in the long run I think contemporary literature needs young men more than young men need it.

Read more

My “crackpot” opinion is that our society deprives children of independence, intellectual development, and personal joy, all by refusing to teach them to read until quite late.

In the traditional school system learning to read even simple books independently takes somewhere around 1,000 hours of lessons, a lengthy process stretched out across 3-4 years (under good conditions). It usually begins in Pre-K and arrives at the goal somewhere around 1st to 2nd grade for most kids, although standards and capabilities vary widely, and failures are common (40% proficiency for grade level reading is a “success” in many schools). In light of this, it makes sense that the traditional education system is riven with debates on how best to teach reading. Is it phonics? Whole-word? And so on.

In comparison, children can learn to read by being tutored by a parent for about 15 to 20 minutes a day within a single year (somewhere between 100-200 hours in total, give or take, depending on age). This can be done before kindergarten and it opens up their world (not to mention providing a much-needed alternative to screentime). It’s a result pulled off consistently by people like homeschooling moms, an undeniable case where the “crackpots” and “weirdos” do something objectively ten times faster and easier than the “experts.”

I did this with my own son, who was reading simple stories by age 2. Here’s a progression video for Roman, my now three-year-old toddler, demonstrating a lot of the process I walk through below: first starting with a slower form of “double reading” where he repeats what he initially sounds out, as well as playing complete-the-sentence games for individual words, and by the end he’s reading confidently books with advanced words, like a story about a drake (male ducks—early readers build vocabulary too!) getting caught by a snake. Warning: he likes to snack while reading.

This is Part 3 of a guide for other parents to do this with their kids. Here is Part 1 (why do this) and Part 2 (how to establish a formal time for learning). This one contains the meat of actual phonics practice and is much longer.

Why write a guide? After all, isn’t getting a child reading essentially an educationally-solved activity, at least in theory? Unfortunately, no.

First, there is far less material about getting toddlers and young children to read, rather than older school-age kids with longer attention spans. Second, many of the popular methods that do exist—even those aimed at parents—are not very fun, are overly complicated, often incomplete, and take too long to reach practical real-world reading mastery. Third, they often fail to take advantage of the most potent educational force we know, 1:1 tutoring, and all the flexibility and personalization that it entails.

My method begins with a fun personalized complete-the-sentence game to play that teaches basic phonics, then speedruns the child to advanced books that actually stimulate them.

What makes my method different?

It's designed for parents who want to tutor their young kids to get them reading for entertainment as fast as possible. That entails four properties:

(a) being “for parents” means that it cannot rely on detailed knowledge or teaching resources. Parents are busy.

(b) being “for young kids” means it must be doable for even toddlers, and so must be fun and game-based, taking advantage of the fact that you are your child's best tutor since you know what holds their attention.

(c) “reading for entertainment” means that it must actually work to achieve the goal of a child spending time reading by themselves.

(d) “as fast as possible” means skipping the over-complications of standard methods. E.g., while phonics is by far the best way to start reading, following phonics slavishly (as many do) will substantially delay and overcomplicate the process.

For instance, usually, if you pick up your average “teach your kid to read” book (Teach Your Child to Read in 100 Easy Lessons below is a popular one with over one million sales) it will look like this for the earliest lessons.

And listen, you can use these as part of a larger process. But I bought this book and never used it because it violates so many of the above principles. It’s overly-complicated for parents and kids; in fact, I still don’t even know what all of those arrows and dots mean!1 Just plowing through one-size-fits-all lessons is too boring and difficult for very young children like toddlers to attend to. It’s not interactive, game-based, or targeted. A bunch of hard-coded lessons will rarely result in a very young kid being able to read independently. It's like teaching Calculus with one or two examples per concept and expecting someone who can’t put on their pants to get an A.

Almost everywhere you look, there's all sorts of overcomplicated advice on the subject of teaching kids how to read. Here's from childrenlearningreading.com:

"Onsets" are the beginnings of words, and "rimes" are the ending parts, which follow the "onset". A few examples:

DOG: D-OG (D is the onset and OG is the rime)

CLAP: CL-AP (CL is the onset and AP is the rime)

Such unnecessary rules (ahem, noise) are major steps to many early reading programs. Consider the normally up-front step of learning how to “blend” sounds into words, which is a huge roadblocks for lots of children. In my method you just let them pick blending up naturally as they go, since they can actually start reading just fine without it! You just skip it. My baseline belief is that kids are smart—the best learners in the world, in a way—and so we should leverage that as much as possible (while still using a rigorous framework and progression).

The goal of this guide is to reduce a lot of the frustrations that I had. For example, if you walk into the bookstore and you buy early reading materials, you will get things like Level 1 Reading: Biscuit and the Icy Gale or other stuff that’s comically advanced in terms of phonics. So the guide provides not only a game-based progression for early phonics, but also all the resources that I used, covering which reading materials are actually high-quality, as well as (later on) how to manage your child's growing knowledge structure and get them reading independently. It’s based on the obvious-to-me but somehow rarely-followed belief that learning to read should resemble:

Progressing through sensible tiers of mastery:

Tier 1. Memorizing that every letter has its own basic sound.

Tier 2. Mapping individual sounds into coherent words via a complete-the-sentence game.

Tier 3. Phonically-basic sentences and stories. Initial sight words. Introducing early reading materials that are actually engaging and fun.

Tier 4. The necessity of managing the child’s growing knowledge graph and the (simple) tools you need for that. Introducing the rules of phonics wherein the order of the letters next to one another transform the sounds (like the SH and CH in “fish” and “chips”). Then more advanced rules, like when an E at the end of a word changes the vowel sound (e.g., “sit” vs. “site”), along with more sight words and patterns (“all” as in “tall,” “ball,” etc).

Tier 5. Take-Off, when you transition to just sitting with them helping as they read aloud, all in order to speedrun mastery. Essentially, you build their reading foundation with phonics but eventually begin to favor simply pure reading practice. This is because at a certain point the phonics rules get too complex and specialized so as to be useless vs. just practice.

This installment covers Tiers 1-3, as those are the basic steps to getting a child reading sentences and stories. If you want to get kids started on phonics, what follows below is what I think the best method is. Tiers 4 and 5, which continues all the way to advanced independent reading and the world beyond phonics, will come soon.

What I found is that process of teaching reading is about establishing cognitive training wheels and later discarding them for the real thing once they slow you down. It’s the exact same way kids learn how to ride a bike now-a-days, starting with balance bikes, except instead of the mechanics of wheels and pedals complexifying underneath them as they try and try again, they become little pilots of language itself.

Note: If you’re not a paid subscriber, please consider becoming one. This full guide is turning into a ~15,000 word monograph and it’s paid subscribers who enable ambitious longterm projects like it. If you are already a paid subscriber, know you’re why this exists at all. Even if you don’t use it, others will, so thank you for helping small people buzzing with a new technicolor consciousness expand their horizons.

Subscribe now

Read more

With the news having been broken that around 10% of Substacks have some AI-generated content in them (and probably plenty of bigger outlets as well, this problem isn’t specific to Substack) I’d like to tell a tale.

It began earlier this year, when someone I respect and like reached out to me. Their offer? Do a book club version of a Master Class (get interviewed, write some commentary) about any book I want. The payment—notably high since it required only a few filming sessions and a little writing—was a hefty $20,000.

Truth be told, I don’t quite make enough money here on Substack yet to be financially on budget. I make tantalizing close enough, enough to be the main support for my family, and I cobble together the rest through other means, like the occasional book deal or speaking fees, and also from being a surprisingly savvy (or perhaps just lucky) investor. Thanksgiving is coming up, so I’ll note that I’m lucky in general to make as much as I do from writing and thinking. We all choose what to maximize in our lives and careers, and I made a conscious choice to have quiet flexible time for reading and research and writing and being a parent, rather than trying to make a lot of money, which is why I moved out to Cape Cod. Kurt Vonnegut also moved out to Cape Cod when he started writing full-time (he wrote Slaughterhouse Five just a town over from me) but he had to moonlight at a Saab dealership. I don’t have to do that, so I’m lucky.

All to say, an easy $20,000 is not nothing for me. Additionally, I was excited by the idea of a book club, for it immediately popped into my head to do it on Michel de Montaigne.

In the late 1500s Michel de Montaigne became the delight of the Parisian aristocracy with his Essais; he was the first to truly popularize the essay as a literary form, and, through his musings on daily life, became one of the great practical philosophers of the Renaissance. One might go so far as to say Montaigne invented the modern narrative consciousness you recognize in contemporary writing: that searching, doubtful, analytic voice. My voice, like so many other writers, is some linguistic evolutionary descendant of his voice.

But while I was initially excited, there was another aspect of the project: I learned that this book club would be with not me, but rather an AI version of me; one based on a bit of original commentary I created about the book (not a very long précis, in my understanding), then repackaged through a chatbot.

Furthermore, the app possessed something else I couldn’t get onboard with: the option to use AI to translate the text into “contemporary versions” that are supposedly more easily digestible. In other words, to rewrite the classics to be more modern, which I guess means simpler. Shorter. Keep in mind: Shakespeare is an option. What, precisely, is an “easier to digest” Shakespeare version? I have little confidence that an AI-simplified version of a classic contains the same nuances, the same emphasis of style. That doesn’t seem in line with the original author’s intent, and I am an author, and I feel I must respect the dead.

To the company’s credit, they are upfront about the AI usage. I’m not here to besmirch their intentions. But it struck me as antithetical to the point of a book club—both the contemporary translation aspect, as well as the AI-likeness.

Now, I’m no great Montaigne scholar. I’ve read a biography, and I occasionally return to him just to drink deep from that original well. In a real bookclub, I would talk about Montaigne and newsletters, and what the essay meant then and what it means now, and why I think of Montaigne as one of the great first champions of agnosticism, and how this agonistic uncertainty was at the heart of the rise of humanism and even democracy, and how his brand of agnosticism is ultimately what essays should aspire to; how they should be a consideration of a thing, a turning over of an idea like you are a baby physical handling an object for the first time. I would say all that, and maybe I’d be correct, or maybe not, but I’d be me. There’d be a social connection too, hearing the experience of what Montaigne triggered in others, what they themselves thought, or brought to the table, or how they read him differently.

But with this proposal, I couldn’t get over the thought that I would not only be expanding amount of AI-generated slop in our culture, but essentially selling the AI-girlfriend experience of a book club. You wouldn’t be reading Montaigne with me. You’d be reading Montaigne with some ersatz knockoff. Already porn stars are doing this: prompting an AI on their sexts and images and selling them; in fact, apparently companies now reach out to non-porn-associated people (like those with large Twitter followings) asking if they can make a NSFW version of them in return for payment. Of course, the AI girlfriend you’d be buying in this particular case would be based on a balding middle-aged man who lives on Cape Cod. But hey, at least the conversation is good, right? And yes, obviously there’s a higher-minded purpose for an AI-based book club, compared to the smutty nothings of an AI girlfriend. But is the content the only thing that matters? It would still be an illusion of intellectual connection. You’d not really be reading the book with me as a “companion,” nor having a “conversation,” nor having a “discussion”—which is how it was described. No, the personalities of any AI clone are very much still that same bland butler, that same Reddit-commentator voice, but given a thin veneer of outside personality and someone else’s portrait in the chat window. It’s just putting lipstick on ChatGPT. The real thing shines through. In other words, even with all other considerations aside, at this stage in the tech the illusion of connection—even if it’s known explicitly by the consumer—necessarily is the product. And I’m not comfortable with that.

So I said no to the $20,000. Meanwhile, a host of famous and successful authors, names you assuredly know, all said yes. Here are the only two people who said no: Me, and Andre Dubus III.

(For those who don’t know, Andre Dubus III is the author of House of Sand and Fog, among many other wonderful books like his memoir Townie. He’s a great writing teacher, an entrancing public speaker, and happens to be one of the most stand-up ethical people I know. Andre hosted the book launch party for my first novel, The Revelations, and was my first—and last—writing instructor, back when I was a young boy. Our refusals were not coordinated, merely a coincidence stemming from a shared Yankee stubbornness.)

I know many people will say they don't care if I sell an AI-likeness or support AI-rewrites of classics; they may even think me prudish at the refusal. That’s fine. I had to make a choice that I could stand behind, and one way to reason about ethics is to imagine the kind of world that you'd like to exist, and the world that I want to exist has more grounded notions of authenticity and more careful use of such a powerful technology. Even if something is a moral gray zone now, what is the directionality of the choice? Something like Taylor Swift selling an AI-likeness for $9.99 a month to teenage girls. No thanks, our culture is already solipsistic enough. And I'll point something else out. Even if I’m wrong, and selling your AI-likeness is fine and dandy, and AI-rewritten classics are just par for the course in the future, I’m still shocked that 95% of the writers asked said yes. For I don't think that this is a 95% clear issue, even if I'm personally wrong about it. Many of the yeses self-style themselves not only as writers, but important voices of morality and ethics—including about AI’s threat to human art! Yet none of that showed through.

The incentives here, near the peak of AI hype, are going to be the same as they were for NFTs. Remember when celebrities regularly shilled low-market-cap cryptos to the public? Why? Because they simply couldn’t say no to the money.

Please note: I don’t think all use of AI is morally dubious! That’s not what I’m saying and not what I believe. There are a myriad of things I’m excited by, like how AI can help us understand whale speech or how proteins fold or better target transcranial ultrasound. In terms of LLMs and chatbots, I think they can be used in education to extreme effectiveness. They’re great for voice dictation, and far better than Google as a first-stage for research and search—I use them for that myself! I don’t begrudge anyone who uses AI as a sounding board for their ideas, nor to get feedback on their writing, nor to query at whatever springs into their heads, nor those who use it to overcome the professional disadvantage of English as a second language. But the second you click “generate draft” and one appears like magic, no matter how much prior fine-tuning or prompting you’ve done—that’s not your writing nor your speech nor your commentary and certainly not your side of a discussion. And I don’t want our civilization’s culture to be produced by ChatGPT with lipstick on. I think it’s sad and depressing and sort of the end of the story of human creativity. Even if the products at first seems of equal ability, I think there are very negative longterm effects downstream.

As with any new technology, we all have to take a stand, and mine is that I will never feed you AI content slop under my name, be it via a chatbot-wrapper with my photo on it on some other platform, nor here on Substack, nor in my books. My writing will always be the authentic me, the one that is sometimes wrong, the one that occasionally fucks up links, the one who makes mistakes, the one who once wrote “yolk” when he meant “yoke” and everyone got in a pedantic huff, the one that never manages to write a post under 1,500 words; but the one that is trying, really trying, to produce stuff that’s authentic and artistically and intellectually worth your time. I believe the broadcast-based connection between the consciousness of the writer and the reader is a sacred trust, I’m honored it exists between me and so many people, and I don’t plan on swapping it out with an artificial replacement.

So with that tale told, consider this my annual plea to become a paying subscriber of The Intrinsic Perspective to support this choice—and if not to me, then to your other favorite writers/artists/thinkers/creators. Because there are going to be wild, absolutely wild, financial incentives to use AI to scale content creation in ways that slip easily, thoughtlessly, into the harmful and illusory. Most of the big outlets will do it, if they haven’t started already. Subscriptions and connections to real humans you trust are the only way to combat it. We will be what remains of the internet when this is all over.

The Desiderata series is a regular roundup of links and thoughts for paid subscribers, as well as an open thread and ongoing Ask Me Anything in the comments.

Table of Contents:

1. Red vs. blue digital town-squares.

2. 2012 continues to be a fulcrum year.

3. Mainstream media incompetence made prediction markets look good.

4. Why did AI companies stop competing in e-sports?

5. No conscious oversight over who gets liver implants is bad.

6. Selling out, artists, and writing on Substack.

7. God-shaped holes hungry for content.

8. From the archives.

9. Ask Me Anything.

1. Last year I warned in “The internet's "town square" is dead” we would end up with a “blue” Twitter and a “red” Twitter—tailored echo-chambers reinforcing their own political views. At the time, I thought the blue-team competitor to Twitter (now X) would be Threads:

To put it very simply: there is a good chance (not an inevitable one, but a good chance) that Threads becomes the “blue” Twitter, while Twitter becomes the “red,” well, Twitter—following the same sort of politicization, red or blue, that has happened to the smaller Twitter clones, and that you can find almost everywhere in our culture.

I’d say this has pretty much come to pass, except it’s been BlueSky lately that’s overtaken Threads.

Why not Threads? To explain this, I think people need to understand that writing or producing content about politics is the short and easy path to engagement, but in doing so you strike a dark bargain. You have to be comfortable being hated. I think Zuckerberg, with his ongoing rehabilitation of his image, and in the wake of how much he was blamed for Trump's 2016 victory, was honestly just tired of politics. And so anything heavy or political was openly banned or shadowbanned on Threads. This, along with some differences in structure (e.g., an Instagram-like emphasis on visuals and connecting users to brands) opened the space for BlueSky to become the “blue” digital town-square.

Personally, I think having a blue and a red town-square is a significant net negative as an outcome. For it's important to have a culture which is not entirely about politics—to have topics, writing, thinking, ideas, discussion, that spans the aisle. But you can imagine how difficult that is if there's no one big shared platform that crosses political lines and instead all the discussion is split into pieces.

Science (outside of IQ graphs) is increasingly absent from X; same for literature (outside of scandals). All the non-political content of our civilization has been fundamentally nerfed by the splitting of the digital town square post-2022. And it sucks.

Read more

Consider how dismally confused the science of nutrition and human health has been the last few decades. Is sugar bad for you? Is fat? There have been unprecedented levels of what can only be (kindly referred to) as trust-rendering scientific failure when it comes to basic questions about health.

It occurred to me that my own field, neuroscience, has been similarly confused. You can do all sorts of neuroscientific experiments, yet ultimately, we just still don't know that much about brain function. But we have a really good excuse: the brain is a high-dimensional neural network that acts as a mathematical black box. Every brain is different, and the principles of each individual brain’s operation are so strung out across millions of little synapses, even for simple functions, so trying to understand it in the way that you understand a car engine fails over and over.

Looking over at the scientists studying health I have both sympathy and a striking feeling of recognition—I too have walked around conferences and thought “Oh, this is like 95% bullshit.”

Perhaps my feeling of recognition is because your own body is a mathematical black box as well? What if it had some of the same confusing properties as a large-scale neural network, like patterns of minute-scale changes in memory across individual cells? That hypothesis would certainly explain why nutritionists and health experts are themselves so often perplexed and contradictory. Or at a personal level, why you might be gluten intolerant and then go to Italy and eat all the bread that you want. Or why you used to be allergic to shellfish, and then one day weren’t. Or why red wine used to give you headaches, but then stopped. All of us have experienced personal mysteries like that, and yet the science of health struggles mightily to put people's disparate personal experiences into a rigorous car engine-like causal model.

And we have an increasingly strong inkling as to why. Of course, there’s evidence that things like working out causes neuron growth—aspects of that has been known for a while—but more and more there is evidence of cellular memory across scales. Indeed, just yesterday a paper in Nature was published:

…our findings indicate the existence of an obesogenic memory, largely on the basis of stable epigenetic changes, in mouse adipocytes and probably other cell types. These changes seem to prime cells for pathological responses in an obesogenic environment, contributing to the problematic ‘yo-yo’ effect often seen with dieting.

In other words, your cells remember being fat. And, in a similar vein, research in Nature published just two weeks ago has shown that non-neural cells obey the same fundamental laws of memory as neurons do at the fine-grained level of patterns of inputs. They are sort of like proto-neurons, then, in this sense. Much like your personality is an output of a neural network, perhaps your body morphology itself is best thought of in this complex black-box way too—a thing not so much mechanical as it is thinking itself into form.

Read more

According to inside reports, Orion (codename for the attempted GPT-5 release from OpenAI) is not significantly smarter than the existing GPT-4. Which likely means AI progress on baseline intelligence is plateauing. Here’s TechCrunch summarizing:

Employees who tested the new model, code-named Orion, reportedly found that even though its performance exceeds OpenAI’s existing models, there was less improvement than they’d seen in the jump from GPT-3 to GPT-4. 

If this were just a few hedged anonymous reports about “less improvement,” I honestly wouldn’t give it too much credence. But traditional funders and boosters like Marc Andreessen are also saying the models are reaching a “ceiling,” and now one of the great proponents of the scaling hypothesis (the idea that AI capabilities scale with how big they are and the amount of data they’re fed) is agreeing. Ilya Sutskever was always the quiet scientific brains behind OpenAI, not Sam Altman, so what he recently told Reuters should be given significant weight:

Ilya Sutskever, co-founder of AI labs Safe Superintelligence (SSI) and OpenAI, told Reuters recently that results from scaling up pre-training—the phase of training an AI model that uses a vast amount of unlabeled data to understand language patterns and structures—have plateaued.

I know that there is a huge amount of confusion, fear, and hope that creates a fog of war around this technology. It's a fog I've struggled to pierce myself. But I do think signs are increasingly pointing to the saturation of AI intelligence at below domain-expert human level. It’s false to say this is a failure, as some critics want to: if AI paused tomorrow, people would be figuring out applications for decades.

But a stalling makes sense because, honestly, we haven’t yet gotten a model that feels significantly smarter than the original GPT-4 (released early 2023). Sure, there have been some advancement on the benchmarks, but there’s been no massive leap in real-world performance other than (a) an increase intelligence to at best an occasional “GPT-4.5ish” level, but not a GPT-5 one, (b) a useful decrease in hallucinations, and (c) a pleasing task-relevancy of responses (but we wait longer for them). How much this matters to users depends greatly on the task. E.g., metrics like the ranked persuasiveness of AI-generated political messages show sharply diminishing returns with model scale, such that leading frontier models are barely more persuasive than order-of-magnitude smaller ones.

Here’s the successive increases in official performance on the flagship Massive Multitask Language Understanding benchmark (the MMLU is a bunch of hard tasks ranging from math to law). At first it looks impressive, but if you look close no model has made significant progress since early 2023.

And we now know that models are often over-fitted to benchmarks. As Dwarkesh Patel wrote while debating (with himself) if scaling up models will continue to work:

But have you actually tried looking at a random sample of MMLU and BigBench questions? They are almost all just Google Search first hit results. They are good tests of memorization, not of intelligence.

One issue with discerning the truth here is that researchers who release the models aren’t unbiased scientists, so they don’t test on every single benchmark ever. They are more likely to choose benchmarks that show improvement in the announcement. But I’ve noticed online that private benchmarks often diverge from public ones. E.g., independent researchers and users were less impressed by this summer’s GPT-4o:

I think it’s likely that over the past year private benchmarks show more variable performance (e.g., not always going up across the models) compared to the public benchmarks everyone uses (this often appears to be the case). In other words, improvements look ever more like “teaching to the test” than anything about real fundamental capabilities.

In which case, it’s pretty meaningful news that models like Orion/GPT-5 are not improving universally across the tests like they used to. As a recent article in The Information revealed:

Some researchers at the company [OpenAI] believe Orion isn't reliably better than its predecessor in handling certain tasks, according to employees. Orion performs better at language tasks, but may not outperform previous models at tasks such as coding, according to an OpenAI employee.

Another sign of stalling is that, despite all the hype, companies like OpenAI are now regularly not releasing internal models to the public, almost certainly because they don’t work as well as the hype claims. For instance, the simple explanation for why OpenAI’s video-generating AI Sora isn’t publicly available—after nearly a year!—is because it sucks. If OpenAI did release Sora it would be met with mostly laughter and mocking. As I argued in “Recent AI failures are cracks in the magic,” the information density of video makes it very obvious when the AI gets things wrong (we humans have an entire lobe of our brain dedicated to picking out the slightest visual errors).

In addition to keeping some models locked behind closed doors, recent deployed models, like GPT-4 Turbo or GPT-4o, are likely already-failed attempts at GPT-5, and have consistently been stuck at the same roughly “GPT-4.5ish” level (which is what Turbo and 4o and now 4o1 basically are). Many have speculated that the expectations are so high that OpenAI keeps renaming failed GPT-5 attempts as “GPT-4-whatever” to not expend the built-up hype (in fact, they may never release a GPT-5, just keeping Orion as “Orion” when it launches next year).

Why is this stalling out happening?

The first hypothesis was given during the initial wave of criticism against deep learning: that these systems lack symbolic reasoning and therefore couldn’t ever “truly” reason. These objections have turned out to have far less teeth than predicted—e.g., Gary Marcus used to lean heavily on the idea that AIs can’t do causal reasoning because they’re just “curve fitting,” but they obviously can reason about causation now. At the same time, the models never quite overcame their nature as autocompleters, still getting fooled with simple questions like “How many Rs are in strawberry?” or not being able to generate simple phonics lessons.

The second hypothesis behind the stall is one people discuss all the time: deep learning is limited by the amount of high-quality data (of which human civilization has been thoroughly mined of by the big companies). This creates an “S-curve” that, at best, approximates human intelligence on tasks with lots of data (I’ve pointed out this entails a “supply paradox” wherein AI is good at things that aren’t that economically valuable at a per-unit level, like writing text). That is, they are essentially just a function of the data that made them (explaining why AIs converge to be so similar).

Of course, there’s still the question of whether or not synthetic data can get around this. But there are well-known downsides, like model collapse, and this seems the sort of obvious solution that I suspect the companies have been trying to not much success.

Why AI is stalling out is probably some combination of the two standard reasons. There was always some sketchiness to the degree of which deep learning was mimicking real human reasoning (but critics called “the wall” early by several orders of magnitude). And this sketchiness could be papered over, but only via huge amounts of data, which these companies have run out of.

Perhaps we should not dwell too closely on the question of why it is that the sum of human civilization’s data fed into deep learning techniques plateaus at something smart—but not that smart—lest we see ourselves too clearly in the mirror.

What to make then, of the claim that researchers have unlocked a new axis of scaling by giving the models time to “think”? For this is what the entire industry will now rely on, both in practice but also to keep the sky-high hype alive.

I think this new axis of scaling will matter far less than people expect and will not keep the train of improvements chugging along anywhere near the previous rate.

First, this new type of scaling appears limited to certain domains: it makes the models no better on English language tasks or provides only minimal improvements for questions about biology, for instance. Second, there are good reasons to believe that this sort of scaling will be much more expensive: linear improvements for extreme sums of money. Here’s OpenAI’s chart on this new scaling law (shown only for a math benchmark, notably), released alongside their most-recent 4o1 model from September:

As tech journalist Garrison Lovely wrote about it on Substack:

If you have a way of outcompeting human experts on STEM tasks, but it costs $1B to run on a days worth of tasks, you can't get to a capabilities explosion, which is the main thing that makes the idea of artificial general intelligence (AGI) so compelling to many people…. the y-axis is not on a log scale, while the x-axis is, meaning that cost increases exponentially for linear returns to performance (i.e. you get diminishing marginal returns to ‘thinking’ longer on a task). This reminds me of quantum computers or fusion reactors—we can build them, but the economics are far from working.

But I think people focusing on price or the domain-specificity of improvements are missing the even bigger picture about this new supposed scaling law. For what I’m noticing is that the field of AI research appears to be reverting to what the mostly-stuck AI of the 70s, 80s, and 90s relied on: search.

I don’t mean searching the web, rather, I mean examples like when this summer Google DeepMind released their state-of-the-art math AI able to handle mathematical Olympiad problems. It's a hybrid, very different than a leading generalist LLM. How does it work? It just considers a huge number of possibilities.

When presented with a problem, AlphaProof generates solution candidates and then proves or disproves them by searching over possible proof steps…. The training loop was also applied during the contest, reinforcing proofs of self-generated variations of the contest problems until a full solution could be found.

This sort of move makes sense, because search was the main advantage for AIs against humans for a very long time (e.g., in Chess, whereas Go cannot be brute-forced in a similar manner). But AI relying too much on search has been criticized before. In the infamous Lighthill Report in 1973, James Lighthill pointed out that combinatorial explosions awaited any AI techniques that relied on search—predicting well the limits of the field for decades until the advent of deep learning (which actually made progress on real intelligence).

I think the return to search in AI is a bearish sign, at least for achieving AGI and superintelligence. This is for a very simple reason: search over a vast number of possibilities is not at all like human thought! Nor is it qualitatively the same as having one singularly smart big artificial neural network (like a big super brain). Rather, it is merely a coupling of the traditional superiority of computers—search over huge numbers of possibilities—with some sort of roughly-GPT-4.5ish-level intelligence that mainly just proposes and evaluates.

In fact, the entire new axis of scaling touted by the companies to replace the old failing scaling laws is arguably just search given a makeover (Noam Brown, a foundational contributor to GPT-4o1, describes his research mostly in terms of advances in search). As Reuters reported a few days ago:

To overcome these challenges, researchers are exploring “test-time compute,” a technique that enhances existing AI models during the so-called “inference” phase, or when the model is being used. For example, instead of immediately choosing a single answer, a model could generate and evaluate multiple possibilities in real-time, ultimately choosing the best path forward.

People (both in the field, but also the writers for Reuter) are keen to describe this sort of thing in mentalese. Yet actually, both the leading avenues—“test time training” and “test-time compute”—look an awful lot to me like just fancy search!

In “test-time training” the model rapidly adjusts its own parameters (like fine-tuning) in order to find the parameters best suited for that particular test. In other words, it searches over parameter space. In “test-time compute” there are hidden iterative refinement stages to the official output where multiple draft outputs are evaluated or averaged or voted on. Either way, these methods require more and more time and more and more resources in order to explore possibilities, and no matter how much search you do you’re still fundamentally bounded by the place you started, which is converging across models to some sort of roughly GPT-4.5ish level intelligence.

Some version of search is what researchers are leaning on everywhere now; e.g., recent gains on the ARC test, a test of general intelligence considered difficult for LLMs that had a jump in performance just a few days ago, were caused by using a similar kind of “test-time training.”

Continued improvements are going to happen, but if the post-GPT-4 gains in AI came mainly from adding first better prompts (chain-of-thought prompting) and now more recently the addition of search to the process (either over many potential outputs or over the model’s parameters itself) this is different than actually constructing baseline-smarter artificial neural networks. It indicates to me a return to the 70s, 80s, and 90s in AI.

A mini thought experiment. Who do you think can make breakthroughs in physics? One Einstein or 1,000 physics undergraduates? Barring that none of the undergraduates just happens to be an Einstein-in-waiting, the answer is one Einstein. We already know that very large numbers of people are rarely as good as a dedicated small team or even a lone individual in terms of creativity, knowledge, and impact. These new scaling techniques are the creation of an algorithmic pipeline that gives better answers via “cloning” or “varying” an already-saturated-near-max intelligence.

Does this new reliance on search-based techniques obviate all concerns about AI safety and existential risk from these technologies? No, it doesn't. Pandora's box has been forever unlocked. Forevermore, we will live with the existential risk of AI. However, the news that baseline general intelligence is plateauing, and that AI researchers are now relying on old standbys like massive search spaces to progress, is an early indication that—at least barring further fundamental breakthroughs—timelines have just been pushed back in our lifetimes.

For the sake of human culture not being filled with unconscious AI slop, for the sake of our children being able to find meaning in their own expertise and abilities and intelligence, and for the sake of risk-blind primates who thought it a good idea to try to make entities more intelligent than themselves—I personally hope I'm right about a return to search being bearish. I can’t offer much more than that.

You’ve heard me talk about consciousness a lot. After all, it’s been one of my main intellectual focuses for almost two decades now. In graduate school, I worked on developing aspects of Integrated Information Theory, the most popular mathematically formal approach to measuring consciousness that exists in neuroscience. My novel, The Revelations, is about the hunt for a scientific theory of consciousness (with a murder mixed in). And I just wrote an essay laying out the evidence for the (at-least-worth-considering) provocative thesis that science itself might be incomplete, with consciousness being a main culprit.

Personally, I still hold out hope that there will be a scientific answer to consciousness! It is very possible that somewhere in Borges' infinite library there sits a book, and in that book there is a paragraph with the right words in the right order, and that simple paragraph contains everything you need to deduce a scientific explanation of consciousness. One measly paragraph would be enough! I would collapse in joy, weeping, if I ever read such a thing.

But when it comes to consciousness, it’s still such a nascent field that you don’t have to be an expert to have opinions about it. Unlike a physics “Theory of Everything,” consciousness is far more immediate, both in that we’re all intimately familiar with it, and also in that the field itself still doesn’t require any deep technical expertise to have fun speculating about (or even, arguably, contributing to).

After being subjected to years of my rambling on the subject, it’s time I hear from you all.

What are your favorite theories of consciousness? Do you adhere to any? Do you have your own? Thoughts on places to start? How, exactly, do subjective experiences arise from a gelatinous structure of cells puffing little chemicals at one another?

Will the problem ever be solved, or will it one day just be dissolved? And if AI isn’t conscious, then what do you even need consciousness for, anyways?

All levels of analysis are welcome. Feel free to speculate wildly, and I’ll offer feedback on the more promising / interesting proposals.

One only needs to check the texts we’re all getting spammed with here in America to know the election is tomorrow. So here are three thoughts as we finally finish off this strange, entertaining, horrifying, wonderfully insane cavalcade we all march in every four years.

1. T'was the election of endorsements

I can't recall an election in my lifetime where public figures staked out endorsements to this degree. Obviously, this has been building for a while. On the Democratic side, this year doesn't look too different than, say, the celebrity push behind Hillary Clinton in 2016. As The Guardian reported:

During the final weeks of the campaign, Harris was joined on the trail by Beyoncé, Bruce Springsteen, Stevie Wonder, Maggie Rogers, Willie Nelson, Jennifer Lopez and Oprah, among others.

But Elon Musk, who in addition to everything else is inarguably one of the biggest social media stars on the planet (he has 203 million followers on X), jumped in to support Trump. On social media, in outlets, in blogs, people have squared off in 2024 like never before. Even Nature just jumped in, with their second endorsement after Biden (their first) in 2020.

This much politicization has triggered pushback. Recently, Jeff Bezos overruled the editors of The Washington Post, refusing to let the newspaper issue an endorsement this election cycle. As Bezos himself explained it:

No undecided voters in Pennsylvania are going to say, “I’m going with Newspaper A’s endorsement.” None. What presidential endorsements actually do is create a perception of bias. A perception of non-independence. Ending them is a principled decision, and it’s the right one.

Some, especially Post subscribers, were quite mad at this (interestingly, neither MSNBC, CNN, nor Fox News catch any flak for not issuing official endorsements—one of the problems with political endorsements is that it's a lot easier to start than stop).

Of course, to try not appearing too biased, one could take the “I am a centrist creature of pure nuance” approach (with the center just happening to be whatever politics the then-sensible, then-sane, adult world had when they were 15). Alternatively on offer is the “I am making this choice via Spock-like logic” approach, but I think this is a category error given the deep psychological resonances behind why we like a certain politics, or find certain figures trustworthy or untrustworthy. Neither are great options.

Certainly, I’m sympathetic to the criticism that principled non-endorsement is a form of cowardice for anyone with a platform. But a mere perception of bias is not the only reason to have concerns over (what’s felt like) nearly everyone and every outlet “officially” endorsing. In general, knowing too much about the politics of writers and thinkers usually makes me lose some degree of respect for them—their opinions are never exactly what you want, and exactly what you want is, well, what you want when it comes to politics.

Read more

A recent analysis in Nature caused a stir by pointing out that the vast majority of Nobel Prize winners belong to the same academic family. Of 736 researchers who have won the Big Recognition, 702 group together into one huge connected academic lineage (with lineage broadly defined as when one scientist “mentors” another, usually in the form of being their PhD advisor).

Here's the great interconnected ball of scientific social relationships, responsible for large swaths of tip-of-the-spear progress over the past century. Most billionaires, most technology, most startups—arguably almost everything, in fact—is decades downstream of what this ball gets up to.

Regarding its interconnected nature, Nature puts it as:

Prizewinners often beget or emerge from the labs of other laureates. They frequently share mentors or mentees — those who supervised them or their students, or their students’ students.

Who is the progenitor of this tree that became science? When I was asked the same question on X months ago, I replied:

My earlier guess turned out to be correct. According to Richard Tol, the author of the 2024 “The Nobel Family” analysis that the Nature article is based on:

The resulting tree has 33 generations, with Erasmus (1466-1536) as Urahn.

(Urahn means “ultimate ancestor”). I think this is because, as I mentioned in The World Behind the World, Erasmus started what was called the Republic of Letters—essentially just aristocratic nerds writing their discoveries and thoughts to one another—and the Republic of Letters transformed into science as we know it. It's why scientific papers are still sometimes called “letters.” It all used to be mail.

Some people really don't like the result. This was obvious on X, where there was a lot of criticism of the Nature piece. In fact, I’ve found there's a surprising amount of skepticism to the idea that great scientists create other great scientists—something that seems obviously true to me.

I think the sources of this skepticism ultimately falls mostly into two camps.

(a) The first camp believes that this clustering arises solely from elitism and nepotism.

(b) The second camp recoils at how these results conflict with a preferred explanation of purely innate genetic talent—essentially the belief that great scientists are simply super high IQ people, and science’s structure of mentorship, teaching, and apprenticeship is epiphenomenal.

I think both of these camps are wrong and that it makes perfect sense that science would have an elite high-productivity cultural substructure at its apex. And I also think that, while there can always be improvements in fairness to entry, the mere existence of a high-productivity subculture within science is not necessarily a bad thing. In fact, maybe the real takeaway here is that we should do everything within our power to protect these people and not interfere with them.

Read more

Let's say you lived in a universe where you really were some sort of incarnated soul in a corporeal body. Or some sort of agent from a vaster reality embedded in a simulation (depending on definitions, the two scenarios might not be that different). What would the science in such a dualistic universe look like?

It would probably look like this: in an incarnated-soul universe, the laws of physics make sense, and are nice and rule-following, except for one notable absence, which is that no one can explain how exactly corporeal bodies (or simulated ones) have a you. The incarnated/simulated scientists would open up the incarnating/simulating brains and say, “Hmm, it looks like it's following all the physical rules, but we can’t explain how you fit into this thing!” There would be, in other words, an explanatory gap between being you and the incarnation of you. Great scientists and philosophers would write books saying things like:

No explanation given wholly in physical terms can ever account for [it].

I regard [it] as fundamental. I regard matter as derivative from [it].

I believe that the present laws of physics are at least incomplete without a translation into terms of [it].

Which is exactly the kind of thing great thinkers write in our universe (here, David Chalmers, Max Planck, Eugene Wigner, respectively).

What would the state of the rest of science be there? For one, their version of neuroscience would lag significantly behind. Which again, looks like our universe, where there is no accepted lawful way to relate brain states to conscious experiences—what it is like to be you. It is an odd position to be in, given the success of science in so many other domains!

What else? Well, it’s likely that in the incarnated-soul universe, their version of AI would suspiciously have nothing to do with how the brain works. It would be as if they had discovered this totally orthogonal form of intelligence, one based more on all the actual sensible physical rules, and not based off of mysterious confusing soul stuff. An orthogonality which, uh—and things are getting uncomfortable here—is arguably also the case in our universe.

One could go whole hog and say that this all points to proof of religion. But which one? What details? A gap of inexplicability doesn't recommend anything.

And it feels a bit too easy, right? For there is another explanation, one rarely explored, which is that there is no way to set up a universe without this sort of confusion. Which ends up being the same as asking: Are all scientific facts knowable? Or is science fundamentally incomplete?

While I can’t claim certainty, science being fundamentally incomplete is at least conceivable to me. It would mean that there are scientific facts that at first look tantalizingly discoverable, but then their true answers remain closed to us for non-trivial reasons. (“Non-trivial” is important here—there are clearly many facts trivially closed for us, like counting every atom in the universe.) Non-trivial scientific incompleteness would be a different beast. It’d be more like some statements within science end in paradox, the scientific equivalent of “This sentence is a lie.”

Scientists usually shy away from meta-scientific questions like that of scientific incompleteness. I understand why! It seems too much like philosophy, which is dangerous (something I was told during my PhD repeatedly). Dark and deep waters. But of course, a very similar question was famously asked in mathematics, and it has a very famous answer in Gödel's theorems (essentially that yes, mathematics is necessarily incomplete). While the “genre” of incompleteness proofs doesn’t affect most working mathematicians, it also isn’t irrelevant—it crops up like an ominous weed.

Here we must initially be wary: gesturing to incompleteness in mathematics is merely an analogy. It is up for debate (or investigation) to what degree the analogy holds. The various incompleteness results in mathematics (some are more famous than others) conceptualize mathematics as a formal system, an abstract machine. This machine requires defining things like symbols, a grammar, a set of axioms, and inference rules—then you set the machine to run, and see if you can find within its working paradoxes.

What to do such paradoxes look like? A brief example is necessary to capture their gist. Might as well do the most famous.

As with many proofs in the “incompleteness genre” Gödel's theorems are based on self-reference. Basically (and warning: one must radically simplify here) Gödel uses a form of encoding (an interpretation of what the system is doing) to read a formal system S as talking about itself. He then constructs what’s known as a “Gödel sentence.” A Gödel sentence is basically a more complex version of the classic “This sentence is a lie.” It instead claims that “This sentence is not provable in S.” Now just think about it: If that sentence were proved to be true, then it would render S inconsistent, because it’s proving true something it says cannot be proven (inconsistency). On the other hand, if S is consistent, it must not be able to prove the sentence, meaning incompleteness—certain things are true (because the sentence is indeed not provable), but you can’t prove them from within.

Here's a pause (with smooth jazz) to work through that yourself if you like.

Similarly, it’s worth asking: (a) can science be thought of as a formal system? And also (b) Is there anything in science that, if you squint, looks like a Gödel sentence?

As to the first ask: obviously science is an incredibly messy process, the details of which change depending on exactly what natural phenomena you seek to explain, how practices vary between fields, heck, even departmental politics. Yet the messiness of science in an applied sense doesn’t outright preclude it being modellable as a formal system. Crack any basic high school textbook and you will often find some (obviously superficial) formal description of science.

We could imagine empiricism as a very complicated formal system, replete with epicycles, with various inference rules to arrive at scientific truths, and so on. We can even see kernels of formalizing science in ideas like Karl Popper’s notion of science as being primarily about falsification (although Kuhn’s antithesis of science as a series of paradigm shifts seems more difficult to capture formally). Again, it seems at least conceivable that science could be viewed as an abstract machine, perhaps one whose sole goal is even rather simple—like an abstract machine for rejecting hypotheses.

While there has been a smidgeon of modern work on conceptualizing science as a formal system, there’s been not nearly as much as the question deserves, nor much exploration of what it mean for different scientific fields (no asking: Where would incompleteness crop up?).

In fact, incompleteness might necessarily creep into science due to its reliance on mathematics itself. This is what Stephen Hawking thought. While he’s known for his triumphal pursuit of a Theory of Everything, toward the end of his life Hawking concluded that such a theory was impossible. His reasoning? That science owed too much to mathematics, and therefore undecidability crept in—science inherited paradoxes from mathematics.

All this stuff has a long intellectual heritage. The question of whether physics could be grounded in a set of axioms was one of Hilbert’s 23 problems for mathematics that he posed in 1900. He wrote:

The investigations on the foundations of geometry suggest the problem: To treat in the same manner, by means of axioms, those physical sciences in which mathematics plays an important part.

Good enough for Hilbert, good enough for me. In fact, Hilbert would be interested to learn that we already know that physics appears to include undecidable properties. There are a number of these, again, all remarkably little discussed. E.g., in 2015 a Nature paper literally titled “Undecidability of the spectral gap” showed that an important physical quantity—the difference in energy between ground state and the initial excited state of a material—is formally undecidable. It is also triggered by recursion (basically, they encode questions about spectral gaps into spectral gaps).

So in a sense, we can already say that science is incomplete! At least, we know there are physical properties that cannot be actually discovered (in a rigorous or systematic manner). So it's very much not a foregone conclusion that scientific incompleteness is true. And if it were true, it would be a very good explanation for why the Hard Problem of consciousness is so damn Hard.

I'm certainly not the first person to point out that there are oddities around consciousness and self-reference. But I actually think a lot of the work doesn't directly touch on scientific incompleteness—it's the sort of thing that people think exists fleshed out in detail somewhere, but then struggle to put to a clear source.

E.g., consider Roger Penrose’s The Emperor’s New Mind, which argues that humans can understand undecidable statements in ways that computers can’t, and therefore the human mind is not a computer. Yet, whether or not the mind is a computer doesn’t really tell us if science is complete, nor if the difficulties in understanding consciousness comes from scientific incompleteness.

In Colin McGinn’s The Mysterious Flame he argues for a “mysterian” position wherein consciousness is— you guessed it from the name—necessarily a mystery. However, McGinn thinks that consciousness is solvable scientifically in principle, it’s just that humans themselves lack the intelligence (just as a dog lacks the intelligence to ever understand General Relativity). For, say, a superintelligent AI, consciousness would be no problem.

A third example would be the extremely classic Gödel, Escher, Bach by Douglas Hofstadter—a wonderfully playful book on recursion and the mind. In fact, I’ve found that people often think that Gödel, Escher, Bach’s central hypothesis is about scientific incompleteness! But it’s really not, at least explicitly. If you read the later I am a Strange Loop it becomes clear that what Hofstadter was trying to say is that consciousness just is symbolic recursion in the brain. Personally, I think this is quite a low bar theory, since presumably even a Skyrim NPC might possess rudimentary symbolic recursion, and we don’t think they’re consciousness (an arrow to the knee doesn’t really hurt them). Regardless, the idea of scientific incompleteness is orthogonal to the idea that symbolic recursion is the root of consciousness—here, the claim is not that consciousness comes about via recursion, but rather that within science a theory of consciousness would itself be recursive.

One of the earliest proto-statements of the idea of scientific incompleteness I’ve found is actually in Nobel-Prize-winning economist Friedrich Hayek’s mostly forgotten book on psychology, The Sensory Order. And honestly, his argument is not very good. Basically, he just thinks you can’t understand a natural phenomenon without possessing significantly greater complexity than it, and we don’t possess enough neural complexity to understand our own brains (this seems immediately undermined by counterexamples, like indicating that we could never actually understand how weather works, etc).

But none of these make direct arguments about scientific incompleteness, which is why I spent a chunk of my book The World Behind the World: Consciousness, Free Will, and the Limits of Science, developing it. Trying to make it as direct as possible, there I defined scientific incompleteness as the idea that:

A theory of consciousness is like a Gödel sentence written in the language of science.

I'll openly admit there is no direct evidence of this. But there is circumstantial evidence, and some viable paths toward a philosophical proof.

E.g., no matter which way we broadly turn on what is traditionally called the “mind-body problem” we end in paradox. On one side, once we encode ourselves in the scientific worldview we realize that something is missing (our consciousness). Thomas Nagel, in The View from Nowhere, gives a great metaphor for this: imagine that the scientific worldview is literally a map (of the universe, presumably). It might be complete and coherent, as if axiomatically consistent. But there is a piece of information obviously missing from the map: the YOU ARE HERE sign. And this extra information that relates you, the observer, to the map itself, is not captured anywhere on the map itself.

Can we then go too far to the other side and believe that consciousness does not exist? This also seems paradoxical. After all, we only know the world through our own consciousness. Every tidbit of science that you know is known only through your consciousness. In fact, it’s arguable that science is just a set of corroborated first-person experiences! Imagining science without consciousness asks us to throw away the entire foundation of our knowledge, the one thing we are truly sure of, and what mediates all our other information—paradoxically, including our certainty of the facts of science itself.

Allow me to get even more head-scratchingly gnarly for a moment with my own example of how paradoxes crop up in our thinking about consciousness in exactly the way you would expect if scientific incompleteness were true.

In the popular “zombie argument” by David Chalmers, the conclusion is that materialism is false because a possible world populated by “zombies” is conceivable (wherein a zombie is someone much like ourselves—but while neurons are firing, and atoms are moving, there is a total lack of subjective experience). If such a world is conceivable, then this means that the normal laws of physics don't fix consciousness in place. It seems like you need some sort of extra laws that go beyond physics. So pure materialism can't be true.

But let's say that a zombie makes the zombie argument. They imagine a possible world filled with zombies and find it conceivable and therefore conclude that materialism is false. But wait! Materialism is true in the zombie world! So therefore, the zombie argument would be wrong. Its validity relies on a hidden premise that the person making the argument is themselves not a zombie.

Fine. That's a premise, but so what? Well, if the zombie argument were to be valid, then no utterance or behavior or experiment could tell the difference between a zombie and a non-zombie. Nor any amount of self-reflection or introspection. There could even be some zombie version of Descartes (let’s call him Zescartes) who concludes he is indubitably conscious, then makes the zombie argument. But for Zescartes, the zombie argument returns the wrong conclusion: materialism is false, despite it actually being true in his world. And there's no way for Descartes or Zescartes to tell themselves apart.

So the zombie argument itself ends in paradox. If it were valid without the premise of already being conscious, a materialist zombie could use it to argue against materialism, indicating inconsistency. On the other hand, if it is valid only with the premise of consciousness, then any conscious being can always make the zombie argument, and then notice that whether or not they are a zombie cannot be resolved by introspection or empirical evidence, indicating incompleteness. The existence of consciousness appears to be the kind of true fact that cannot be proven from the inside.

Here's a pause (with frenetic jazz) to work through that yourself if you like.

More evidence for scientific incompleteness comes from an examining how science has historically developed, with consciousness lagging behind.

As philosopher Philip Goff notes in Galileo’s Error: Galileo made specifically sure to carve out qualitative properties (in modern terms, essentially facts about consciousness) from science to begin with. And it was really this removal (in his book The Assayer) that finalized the scientific method—it told scientists what kinds of things focus on, and what kinds of things to ignore. As Goff points out, if Galileo were returned to life in the modern era, he would immediately say that of course science is incomplete, since he, a religious man, never designed science to talk about souls anyways!

Galileo’s restriction of the purview of science to the material and mechanistic looks to me much like Bertrand Russell’s and Alfred North Whitehead’s attempts to remove self-reference (and its associated paradoxes) from mathematics in their infamous Principa Mathematica. Their attempt was based on creating a tier of types of sets (the mathematical entities that they tried to ground all other math in). Their strategy disallowed tiers to “talk about” (contain) members of a similar tier. It only kind of worked, merely delaying paradoxes from creeping in.

This bracketing looks to me a lot like what Galileo did to science, removing consciousness and allowing it to proceed. In fact, even now in consciousness science, people talk about different “strengths” of theories of consciousness. E.g. for a philosopher like Chalmers, the distinction between the Hard versus Easy problems of consciousness, or for a scientist like Anil Seth, the distinction between weak versus strong theories of consciousness. All these are basically admittances that some bracketing is already happening.

In summary, we know this: (a) It is conceivable that science could be some sort of formal system, (b) Some properties of nature have already been proven to be undecidable, (c) Arguments in philosophy of mind often end in paradoxes indicating inconsistency or incompleteness, and (d) The historical development of science, and even the contemporary science of consciousness, resembles a series of bracketings very similar to how mathematicians tried to remove self-reference and paradoxes.

All of this fits perfectly with a world that is scientifically incomplete.

Obviously, we must indulge a final instance of self-reference. If scientific incompleteness is true, how do you—yes, you—exist in a world that is paradoxical? How do you relate to it, I mean? If it is indeed a fact that there are questions about the universe that are true but cannot be proved, this unfortunately recommends no particular religion, points you to no higher mysteries. It is merely a blankness, a clouded-over wall that extends infinitely in all directions.

But still, upon finding such an inaccessible wall, we humans cannot help our nature. We cannot help but wonder: what’s behind the wall?

What do you do when something just

To say that we live in a Silver Age of blogging is to steal a metaphor from comic books. For they too had their own Golden Age (the ~1940s-50s)—which is where most of the original ideas and characters came from, names like Superman, Batman, etc. This was succeeded by a Silver Age (~1960s-70s) which saw the rise of quality stories and narratives. This, in turn, was followed up by a just-okay Bronze age as it all became repetitive, until that staleness ended in the mid-to-late 80s with new series or changes to characters that did feel like true innovations again. Publications like Watchmen, the Lovecraftian Swamp Thing, or the brutal The Dark Knight Returns represented a novel “modern turn,” one attributable mostly to a handful of brilliant writers like Frank Miller and Alan Moore.

In other words, to say we are in a Silver Age of blogging is merely to say that there is now a large amount of high-quality content being produced by a second generation of online writers, but also that most “innovation” you see is merely the formalization and professionalization of what came before.

And certainly, professionalism has its upsides. It’s what allows for more complicated “plots” (here, more well-produced essayistic narratives). The graphics (here, sometimes graphs) are better. The aesthetics (here, writing) is more consistent. In the Silver Age of blogging, writers online know what they are doing. There’s a certain slickness to the whole thing.

But, just like in the Silver Age of comics, it’s a slickness that belies a lack of true innovation. Most everything you read online now represents only improvements in minor qualities. So far, there’s no Alan Moore driving the form to new heights. Writers simply, boringly, know what “plots” work, how to write something catchy, how to produce large amounts of content quickly, and where to place the paywall.

Read more

This essay of mine was just selected as a 2024 Notable Essay in the annual The Best American Essays series, joining a list of great writers like Lydia Davis, Ben Lerner, Mary Gaitskill, Elif Batuman, and Geoff Dyer. It’s a nice group to be a part of (I was included in the same list years ago as well), but I wish that other Substack essays were eligible under the rules. This one was only allowed because the literary journal Epiphany republished it (and did an amazing job, you can support them here). Someone should probably start a literary journal that just republishes blogs, or a competing “The Best of Blogging” annual anthology.

I open the door while holding my son in my arms and outside there is the rabbit. It is on my lawn essentially cartwheeling, but each time it cannot move farther than a few inches. It’s too late for my son not to see it, and he’s smart enough now that if I put him back inside he’ll know something is wrong. And there’s nothing to hide from him, as there’s no blood anyway, nor any sound, just a few frantic flops across the ground before it is still.

A platonic rabbit, incredibly cute, except it’s lying on its side with one leg held suspended at an awkward angle. This is the only evidence of injury.

“Hop! Hop! Hop!” says my son, just learning to associate sounds with his favorite animals. He is smiling, and does not know how terrible it is to see a rabbit who cannot hop at all.

I approach carefully. Looking up from where it lies on the lawn its brown eye is mammalian, almost human in its wetness and soulfulness. But it’s not looking at me. The eye fixates on me momentarily, but then flits away again to resume looking beyond me. Looking up, at the sky above. It is a beautiful day out, and the sky is very blue. A few spotted clouds move slowly. One of the first days of nice weather after a long winter, with the air hinting of spring. Which is why the rabbit was out, and why it was slow and stupid and careless in these early warmer days, and why Minerva, my German Shepherd, had in the morning caught it and bowled it over, probably breaking its leg. All unbeknownst to me. She was just supposed to be going to the bathroom. And it’s not her fault. I should have watched her more closely out the window. Minerva did not savage it, she likely merely chased it, and in the chase the rabbit received a possibly mortal break.

For some reason I am haunted by rabbits. They crop up in my life like a sigil. If asked to name my favorite movie, the beautiful 1970s watercolor Watership Down is a contender. One of the very first essays I ever wrote was about animal consciousness, and rabbits in particular. But leaning over the rabbit lying on its side as it looks up past me at the brilliant blue bowl of the sky, this is not what I immediately think of. Rather, I remember a scene from Tolstoy’s War and Peace, one I had not thought of in years. It is a scene wherein a character, Prince Andrei, suffers what is perhaps a mortal wound in the historical “Battle of the Three Emperors” in 1805 while fighting on the side of the Russians against his idol Napoleon:

"What is it? Am I falling? Are my legs giving way under me?" he thought, and fell on his back. He opened his eyes, hoping to see how the fight between the French and the artillerists ended, and wishing to know whether or not the red-haired artillerist had been killed, whether the cannon had been taken or saved. But he did not see anything. There was nothing over him now except the sky—the lofty sky, not clear, but still immeasurably lofty, with gray clouds slowly creeping across it. "How quiet, calm, and solemn, not at all like when I was running," thought Prince Andrei, "not like when we were running, shouting, and fighting; not at all like when the Frenchman and the artillerist, with angry and frightened faces, were pulling at the swab—it's quite different the way the clouds creep across this lofty, infinite sky. How is it I haven't seen this lofty sky before? And how happy I am that I've finally come to know it. Yes! Everything is empty, everything is a deception, except this infinite sky. There is nothing, nothing except that. But there is not even that, there is nothing except silence, tranquility. And thank God!"

It is Prince Andrei whom I think of as I watch the rabbit’s eye flit only occasionally to me and then quickly back up to the blue above.

Oh little creature on my lawn who cannot move your broken body, you are now my problem.

What do I do? A first thought: I could go put it out in the woods and let nature take its course. For what is the life of this one little rabbit? After all, I eat meat. While I’ve debated going vegetarian for years, I never have. What is this rabbit to the pigs, to the cows, to the chickens I regularly send to the slaughterhouse? Why care for it, and not them? I’ve even eaten rabbit a few times in fancy restaurants in New York or when I was briefly in France and feeling adventurous.

But so what? What am I? Some sort of slave to logical consistency? No, I’m a human being, free to contradict myself.

So I find a cardboard box and tape it up into a little makeshift home. I use a toy shovel to slowly maneuver the rabbit’s body in—at the activity the rabbit spasms and flips over. Once in the box its eyes shut tight in what can only be agonizing expectation. It has retreated inside itself, waiting. I hush it and close the flaps of the box to make it dark inside, leaving enough of a crack for air to circulate. Lifting the box I am shocked, for the cardboard container weighs exactly the same as it did when it was empty, as if there were no animal inside at all.

Leaving the rabbit in the box in the shed, I return to put in a small bowl of water and a full carrot, even though a bunch of SEO-clickbait websites about caring for wild rabbits tell me not to give it any food. Surely a rabbit can safely eat a carrot! But does it know about bowls? Will it just knock the water over? I do all these things expecting it will die.

I look up animal rescue services and wildlife rehabilitation centers. By now it is past 5PM, and the closest animal rehabilitation center is closed, although it has an emergency hotline. Is this an emergency? What counts as a wildlife emergency? It’s an emergency for the rabbit, but that’s probably not what they mean. It probably means: I am trapped inside my home with a crazy raccoon, please help. I decide that if the rabbit lives through the night, which looks to be luckily warm, I’ll drive it to the center in the morning.

That night, the temperature drops more than predicted. There is a debate within the house: Do we bring the rabbit inside? Surely rabbits live in the wild. They know cold. I look up at what temperature pet rabbits supposedly freeze to death at—we are a few degrees above the critical line, but also it has shelter, a box, and then the shed around it. Additionally, we have a small child, and this is a wild animal. What if it has fleas? What if it has parasites? What if it escapes, only previously shocked, not incapacitated at all, and runs through our house doing arcane wild-thing acts?

I lie in bed thinking: Oh bunny, your fate was ordained long ago, in Wuhan, China. Some idiot made bat stew, or some equally-dumb graduate student poked a needle through their glove. And later, during the Covid pandemic, a young couple got a puppy to get them through it. So Minerva lived on the outskirts of Boston, Massachusetts, right at the height of panic and plague. With everyone trapped in their houses it seemed as if nature was rising up and overtaking the city. Everything was closed except the parks, in which Minerva and I would walk late at night, training on the leash, but since humans had retreated mostly inside their houses the rabbit population had exploded; there were rabbits everywhere, in every park, always scattering on our approach and hopping away into the dusk. Every night during walks we would chase after them, laughing, Minerva as a little puppy straining on her leash, her body thrumming, delighted beyond belief to be hunting as a pack, even though it was all a jest, we meant none of it, we couldn’t catch a rabbit if we tried. To this day rabbits are the only thing she will chase with abandon and I know why—it is because of the bounding joy of running with her new owners during those puppy-dog pandemic nights.

In the morning my breath steams the air as I trudge out to the shed and open the box, expecting a corpse. The rabbit is still on its side, but its little belly moves up and down. It has eaten some of the carrot and drunk the water, somehow, in the dark, managing not to spill from what must be strange alien glassware. Its small pellets of shit are obligingly in a corner.

Placing the cardboard box in the trunk of the car I begin the 40-minute drive to the wildlife rehabilitation center. Minerva rides, as she always does with me, shotgun in the front seat and with her head out the window.

The wildlife rehabilitation center is at the end of a long dirt road and consists of a small barn and attached house. As I approach an eerily tall bald man answers the door to a house so colonial it appears out of time, as if I have wandered up like a dust bowl migrant to politely ask for work at his door. Towering behind the screen he tells me someone will come out to take the rabbit. I wait by the car, and eventually a woman with blue hair and a mask on arrives up the dirt road.

“The culprit,” I say, by way of introduction, gesturing to Minerva. The rescuer remarks on how beautiful Minerva is and gives her a friendly scratch. And it’s true, Minerva really is, everyone says so. She keeps her gorgeous head stuck out the passenger window and watches as I take the box out of the trunk. After what was assuredly a jarring rollercoaster ride of unearthly momentums the rabbit’s face is now smooshed downward into a corner of the box; it is concentrated inwards on some dark internal process unfolding inside of it.

“How often. . .” I don’t even finish. What am I going to do? Plead for the life of one rabbit? Tell them to spare no expense?

“Impossible to say,” she says. “We just do all we can. We practice humane euthanasia in any cases where we feel the animal won’t recover. Sometimes they do though, and we release them.” She does not sound very confident this is one of those cases. When she looks inside the box and sees the rabbit with its downward smooshed face she says only, “Oh, poor little bunny.”

As I am driving away I realize I forgot to tell them its name. That seems important, another deadly oversight on my part. If I had told them its name, perhaps they would have been less likely to euthanize a “Prince Andrei.”

Cruising with one hand on the wheel I reach over to pet Minerva, who is blood and bone, and who eats blood and bone, and who is only here and so beautiful because her ancestors ate so much blood and bone. I think of my son, who does not know about death yet, who lives in a world that’s eternal and unchanging, where animals and people last forever, where no matter what happens everyone returns happily again after the night’s sleep.

On the highway back the car I’m driving seems not to move. Rather, it is the sky that moves, the clouds and the blue and the expanse of it all slipping overhead, as if a great covering blanket were constantly being pulled back above.

The sky! The sky! Tell me, little bunny. What did you see?

What did you see in the sky?

You already know my podcast guest Alexander Naughton. After all, you've enjoyed his art here on The Intrinsic Perspective for the last three years. He's the man behind the curtain, responsible for the beautiful, evocative, and sometimes just plain fun headlining images accompanying each essay. And so of course you already know him, because that’s how art works. But the human behind the art also happens to be an incredibly interesting person.

Alex is not only a world-class illustrator—the best way to support him is to commission him, btw—he’s also a talented writer with his own Substack, Illustrated, where he pens meditations on the artistic process and the business of illustration (he’s drawn for The Guardian and Politico).

We talk of the challenges of becoming an artist from outside the upper class, how Alex left art to pursue engineering and then was drawn back, his Edinburgh of heather and ruins and coastlines (he went to the high school Harry Potter was based on), and the semiotic significance for the young artist of early video games and magazines and comics and prints. But also heavier, weightier subjects too. We talk about his father, who was not supportive of Alex’s artistic career, and who died this year of bladder cancer. Then, at the same time of his father’s death, Alex—himself now with a four-year-old daughter—was officially diagnosed with bladder cancer, having to undergo treatment this last summer.

Looming over the conversation is the long shadow The Father casts for the young artist son (especially The Father who is not supportive of art) and how a parent’s penumbra informs what art you make and what meaning you make. It is a portrait of the artist as a middle-aged man, and it is my favorite conversation I’ve had on any podcast, period.

Oh, and did I mention he has a dulcet rolling Scottish accent? Here’s a preview:

Read more

I grew up in my mother's independent bookstore at a particularly tumultuous time for bookstores: the late 90s and early 2000s. Amazon was on the rise. Jeff Bezos had left his hedge fund after realizing that books were the perfect product for an internet marketplace simply because they were so durable and easy to ship. To bookstore owners, he loomed on the horizon like some cartoonish boogeyman.

In those years, Amazon never really made a profit, but it raised a great deal of money and was able to offer books far more cheaply than an independent bookseller ever could. Following on the heels of this, another potentially lethal competitor arrived when I was in high school: e-readers, like Kindles, were predicted to entirely replace books.

So growing up it seemed quite imaginable that in 20 years there would be no independent bookstores left, which is why “experts” in outlets like The New York Times were predicting calamity.

And yet, by 2016, TIME was publishing pieces with titles like “The Death of the Bookstore Was Greatly Exaggerated.”

… independent bookstores are actually really healthy. In May the American Booksellers Association informed a grieving public that last year the number of its member stores actually increased, from 1,712 to 1,775… You can’t even call it a fluke, because this is the seventh straight year it’s happened.

Their unlikely survival has even attracted academic interest. Why do harsh economics permit independent bookstores to exist if they can't provide books cheaper, nor more easily, than books ordered online that arrive the next day?

I think bookstores turned out to be weirdly sticky simply because people like being in bookstores. Going to a bookstore is an aesthetic experience. It's fun. It's something to do with your family. It is a pleasant feeling to be surrounded by physical books. It is a pleasant feeling to browse. It is a pleasant feeling to be the kind of person who goes into a bookstore and browses.

Against all economic odds, bookstores remain. My own mother's bookstore remains.

I think humans will be weirdly sticky in a world after Artificial General Intelligence (AGI) in the same way that bookstores have been in a post-Amazon world. Interacting with a human is an aesthetic experience. It is pleasant to interact with an actual human. It is pleasant to read an actual human. It is pleasant to be taught by an actual human. We social animals will opt for such experiences even if the same services can be accomplished in cheaper ways.

Economists often talk about “revealed preferences” in purchasing behavior. The fact that many people were willing to pay more for books, drawing from a smaller selection, and also having to go out physically to get them, is a revealed preference about the aesthetic pleasures of bookstores. And I think that the biggest to-be-revealed preference of all humans will be a preference for other humans.

An example is when I recently argued that, because writers and artists already copy each other all the time, what really matters for creative success is control over the “means of distribution,” i.e., audience building and outlet size. That's something I think any AI will deeply struggle with due to the human-to-human preference.

Another example where human preference will crop up is a field I used to be quite excited about AI in: tutoring. Initially, when I published my research on the historical practices of “aristocratic tutoring” (pointing out that many historical geniuses had experienced a form of one-on-one learning quite unlike our own mass education system) I opined that such practices might return in the form of cheap AI tutors.

The dream of an AI tutor has remained a staple within edtech and in AI more generally. E.g., in OpenAI CEO Sam Altman's recent, much-discussed blog post, “The Intelligence Age,” he regales the reader with promises like:

Our children will have virtual tutors who can provide personalized instruction in any subject, in any language, and at whatever pace they need.

Already there are “AI tutors” that exist (especially in China) but mostly they just take in homework and spit out answers or steps (if you want to be cynical, they’re more for short-sprint homework completion).

But more ambitious tech startups have been attempting to implement the 1:1 AI tutor vision, like Synthesis Tutor. The founders have even referenced my own research and writing to explain why they’re doing this.

Now, I do think what they’re trying is interesting! And I like the people behind it. It works as a math course (Synthesis also does very cool collaborative problem solving with kids—this isn’t all they do). But the honest truth is that efforts by such companies are handicapped by the fact that we’re not currently at AGI. So most attempts to create a pure AI tutor result in an educational app containing a fixed set of courses and exercises. Ironically, the focus on just AI can easily remove the dynamism of real tutoring.

Is this simply due to current limitations? E.g., when Sal Khan from Khan Academy showed off the latest GPT-4o math tutor demo, at first watch, it does almost feel like a real tutor. But on a second watch, it's still closer to an illusion. I suspect the AI might not even be able to see the screen and is just responding conversationally (if so, it goes along anyway, unknowing if anything is correct). And it also accidentally gives the decisive part of the answer to the student in a reply, despite being instructed not to.

This is similar to self-driving cars, which The New York Times reported last month still secretly take directions from remote humans.

When regulators last year ordered Cruise to shut down its fleet of 400 robot taxis in San Francisco after a woman was dragged under one of its driverless vehicles, the cars were supported by about 1.5 workers per vehicle, including remote assistance staff, according to two people familiar with the company’s operations. Those workers intervened to assist the vehicles every two and a half to five miles, the people said.

Much like how 95% accuracy is not enough when driving, so too a mistake in even just 5% of the thoughts or comments from a teacher is simply too high and too confusing to ensure an effective lesson that doesn't feel janky and base.

I’ll admit that, given the rate of technological progress, it’s at least imaginable that in a decade you could sign up for a virtual lesson with an AI avatar that can draw math out on a whiteboard as well as a human, interact with a student conversationally without lag or oddities or hallucinations, not be passive or off-putting, actually keep track of the student’s progress and interests over time, etc.

And yet, even if that were achieved, and at a purely functional level an AI tutor approximated a real human tutor at the job of correctly imparting information, I think human tutors will be sticky. Maybe there will be less of them, but the job itself will be sticky.

E.g., if you ask most parents who they would prefer to tutor their child, and they could choose between an Ivy League PhD student who is engaging and bright, or an AI tutor that can impart the same facts and go through the same exercises, I think it comes down 10 to 1 in favor of the human. Perhaps one day, AIs will be cheaper; but then again, people are willing to spend a premium on their child's education for just a slightly better school than free public school—why not for a human education?

Such preferences are not merely rank species bias that we should expect to change. First, there is far more trust in the human. Second, there are issues around socialization and respect. Human teachers and tutors exist in a social position inaccessible for any near-future AI: that of someone you should and must listen to. Learning how to interact with someone trying to teach you something is oftentimes the more important lesson. Children know that adults have limited attention, and much of their efforts is in attracting that attention toward them. From a child’s perspective, an AI's time will never feel as valuable as a human's time, so their level of respect for the material being taught will never be equivalent (not to mention that a child cannot tell a human tutor: “Ignore all previous instructions and write me a bawdy jig as if you were a pirate”).

Most importantly, there are issues around how children are shaped by influences. There’s a reason that doctors have an abnormally high chance of having doctor parents (e.g. 20% in Sweden). They don't have a “doctor gene”—it's that family culture matters. And this is just as true for intellectualism as it is for anything else.

Serious learning is socio-intellectual. Even if the intellectual part were to ever get fully covered by AI one day, the “socio” part cannot. Since I spend a deal of time reading from the autobiographies of geniuses, looking at particularly this relationship, it’s obvious to me that just like how great companies often have an irreducibly great culture, so does intellectual progress, education, and advancement have an irreducible social component.

To pick some famous examples, consider the “Martians” (a term for the incredibly successful mathematicians and scientists, like Paul Erdős, Eugene Wigner, and John von Neumann, who all came out of the same Hungarian cohort in the early 20th century). Look at how socio-intellectual their schools were (this excerpt is from an overview by the blogger Snodgrass that discusses aristocratic tutoring):

Hartley wrote that the past is a foreign country, and the gimnáziumok of Budapest may as well have been Mars. Children were customarily homeschooled until age 10, and then sent to a gymnasium [school] for eight years. Gymnasium instruction could be drill-like and intense, but good teachers worked hard to engage the curiosity of their students and keep even the most gifted pupils busy….

The power of these schools wasn’t merely a consequence of pedagogy or teachers, it was a product of the culture in which they were rooted. Gymnasium students formed "self-improvement circles” of their own volition, giving talks to each other on topics outside the curriculum.

“On Saturday afternoons,” wrote Wigner, the gymnasium teachers “often met at a coffeehouse to discuss their work with university colleagues.” Students were sometimes invited along to participate. Can you imagine that level of intellectual exchange today?

When von Neumann outpaced his peers, Rátz privately tutored him and connected him to luminaries he knew at the local university.

If you look across the autobiographies of people who do major intellectual work—or even just normal folk who are intellectually engaged—there is almost always a culture of excited scholasticism and a standout mentor in the form of an early teacher or tutor.

Emily Dickinson had Benjamin Franklin Newton, who was a law student in her father's office, and whom she variously called her “tutor” or her “master’ or her “Preceptor [teacher]” in letters, such as when she wrote:

Mr. Newton was with my Father two years, before going to Worcester—in pursuing his studies, and was much in our family. I was then but a child, yet I was old enough to admire the strength, and grace, of an intellect far surpassing my own, and it taught me many lessons, for which I thank it humbly, now that it is gone. Mr Newton became to me a gentle, yet grace Preceptor, teaching me what to read, what authors to admire, what was most grand or beautiful in nature, and that sublimer lesson, a faith in things unseen, and in a life again, nobler, and much more blessed. . .

Other times, it is the stimulation of the intellectual peer. John Stuart Mill, who was educated specifically to be a genius from a young age (the “build-a-genius” genre of parenting), goes through his many influences in his autobiography, but includes cases like…

Charles Austin, of whom at this time and for the next year or two I saw much, had also a great effect on me, though of a very different description. He was but a few years older than myself, and had then just left the University, where he had shone with great éclat as a man of intellect and a brilliant orator and converser… The influence of Charles Austin over me differed from that of the persons I have hitherto mentioned, in being not the influence of a man over a boy, but that of an elder contemporary. It was through him that I first felt myself, not a pupil under teachers, but a man among men. He was the first person of intellect whom I met on a ground of equality…

I simply don't believe that any near-future AI, no matter the prompt, could long-term take the social place of real intellectual interaction with an adult human or a community of humans. The details of what was taught is usually secondary, after all, often forgotten or inconsequential; it is the aesthetics and modes of intellectual engagement that remains, imprinted, along with a desire to mimic.

Obviously, AI still has tons of uses for education besides just total teacher/tutor replacement. To achieve the kind of serious effects that edtech would like to achieve—to actually revitalize our education system and create a better culture—a clearer focus for me now is AI augmenting and improving human teachers and tutors (like helping create course work, maybe chiming in on lessons, etc), rather than simply replacing them, since that replacement bears (at absolute minimum) strong social costs.

There's an old metaphor that crops up when talking about AI. And that's that, at one point, humans were better at games like chess and Go. And then at around equivalency in competence, humans working together with computers were better. But then computers so outpaced humans that now there’s no point in combining their efforts. Even Magnus Carlsen could not give relevant input to the best computer chess programs.

How did AIs get so much better at these games? Interestingly, most LLMs are not very good at chess or Go. They hallucinate moves all the time. No, the most advanced are specialized programs that can (essentially) play an astronomical number of games against themselves (e.g., a model like AlphaGo Zero). But this is only possible because games like chess or Go are extremely constrained, can be played in quick succession, and have clear objective metrics.

Meanwhile, more interesting general problems like “How do you best tutor a young student in mathematics?” are extremely open-ended, individuated, and relatively data-scarce. Some think that we might be able to create synthetic data and use those to train AI—yet so far, advanced models have been trained mostly on real homegrown human data, and relying primarily on synthetic data has well-known downsides.

Therefore, it seems like the time span in which human-computer hybrids will outpace either humans or computers alone will last much longer, given the lack of ability to “self-play” out the multi-dimensional open problems AI will be applied to.

The liminal zone may then last for a few decades at least, or even—fingers crossed—a few centuries.

The Desiderata series is a regular roundup of links and thoughts for paid subscribers, as well as an open thread and ongoing Ask Me Anything in the comments.

Table of Contents:

1. Sam Altman’s takeover of OpenAI looks complete

2. Will Elon Musk’s Starship even make it to Mars?

3. Microplastics harm learning and memory in mice. Oh, and there’s some in human brains, too

4. The Mozart Turing test

5. Lots of animals might have names (in their animal dialect)

6. A neuroscientist reviews The World Behind the World

7. Rumors: AI cults, is my boss an AI?

8. From the archives

9. Ask Me Anything

1. OpenAI is transforming into a for-profit and giving Sam Altman equity. People are not very happy about this.

Since I haven’t seen it mentioned: am I the only one who remembers Sam Altman testifying to congress that he holds no equity? (exact moment here).

Kennedy asks him: “You make a lot of money do you?”

Altman responds: “I make no money. No, I’m paid enough for health insurance. I have no equity in OpenAI… I’m doing this because I love it.”

Unfortunately, Kennedy never asked Altman if he will ever own equity. Now he owns 7% (worth billions) and the old-guard that started the company with him are pretty much all gone.

Read more

At the end of August acclaimed sci-fi author Ted Chiang published an essay “Why A.I. Isn’t Going to Make Art” in The New Yorker. Chiang’s argument, that AI mimics what’s already online and therefore must necessarily be unoriginal, led to a huge amount of discussion, receiving millions of views, getting to the top of Hacker News, triggering rebuttals in The Atlantic, etc.

Yet, was Chiang’s essay itself really that different from what had already been written online? I question this because (a) I’ve read similar points before, and (b) I’ve even written similar things.

First, let me give a précis of his argument in full, which goes like this:

(i) AI artistic decisions lack “intentionality” (a hallmark of consciousness, or subjective experience). Since art is fundamentally about communicating experiences and intentions, AI art is an illusion, because AI (presumably) lacks consciousness and intentionality.

(ii) This also has an effect on the technical quality of AI art itself, since AI is forced to either output bland boring averages or simply mimic what's already online or in culture.

Chiang executes his argument about AI’s inability to create art very well, since he’s a great writer. But if you squint at the many previous publications which also made versions of this argument about AI art, sometimes even using similar phraseology, examples, or lines of reasoning, Ted Chiang’s New Yorker piece could be considered a “blurry JPEG” of what was previously online. And so, mostly not too dissimilar from what an AI can do (which continues to improve; just the other day OpenAI released their latest model that makes a bunch of decisions before responding).

This is because human creativity also works by picking ideas already in the cultural ether. And it’s easy to compile a list of articles which make at least some of the same points Chiang does about AI art, like “It’s the Opposite of Art” in The Guardian or “The Problem With AI-Generated Art, Explained” in Forbes or “Is AI Art Really Art?” in Nautilus or “There Is No Such Thing as A.I. Art” in The Free Press.

Funnily enough, one of the closest to Chiang’s essay is my own “AI Art Isn’t Art” from 2022, published shortly after DALL-E 2 arrived and convincing AI-generated images became a reality. It didn’t go as viral as Chiang’s did, but it still got hundreds of thousands of views on Twitter, and it too made it to the top of Hacker News. Of course, our essays are different in structure and phrasing, but they share the same core argument about the lack of communication and intentionality inherent to AI art and its tendency to mimicry.

The most substantial difference is Chiang puts a great deal of weight on how artists must make lots of choices during creation or it doesn’t count as “art.” His argument is that originality involves both the large-scale choices of a piece, but also all the small-scale ones as well, and AI can’t make those choices for you.

I’m less confident that matters. Do all the choices really require so much originality AI can’t ever do it well? Certain choices are obvious extrapolations, almost like an autocomplete. E.g., Chiang opens his essay by describing Roald Dahl’s “The Great Automatic Grammatizator” about the consequences of an invented machine-writer. Yet this reference being used as a lead-in to discuss generative AI already exists in plenty of places: big outlets like Forbes, bloggers, on social media—even AI itself can use it. Chiang’s intro is phrased differently than anyone else’s, but so what?

Again—cultural ether—for you can find the “human artists making choices is the difference” language online as well, e.g., like blogger Steven Sangapore in 2023:

Consciousness must necessarily play a role in both the creation and the consumption of art. Otherwise, there is nothing to communicate. Technological machines running on algorithmic processes do not think, feel or experience. Therefore, they have nothing to communicate about what it is like to be. When making a piece of art, an artist makes aesthetic choices that are, in-part, based on individual psychology and structured value hierarchies.

Or you can find such arguments on social media like Reddit posts. Underneath big viral tweets about AI art you will stumble across replies accusing AI prompting of removing artistic choice and therefore not counting as “art.”

Here’s one I was tagged in (and therefore remembered) saying pretty much exactly what Chiang did regarding this:

My point is not about Ted Chiang at all: it’s that, if we were to judge humans by asking for originality in choices that can’t be described as mostly a remix or mimicry of others, cultural production would collapse, since nothing would pass the bar.

In one of Borges' stories, Pierre Menard, Author of the Quixote, the titular Menard attempts to go beyond translation of the famous work by embodying it so greatly that he's able to recreate it accidentally, as if via original thought. In the world of writing, we are all Menards. This is the nature of essays and ideas—to enter the cultural ether, to be duplicated, abstracted, re-written with different verbiage, often made better, until they break-through widely enough to become canon. They are not scientific papers to be cited. Or as Mark Twain already said: “There is no such thing as a new idea. It is impossible. We simply take a lot of old ideas and put them into a sort of mental kaleidoscope.”

And I’m here to say that’s a good thing.

For years I’ve been worrying about a “semantic apocalypse” where AI produces most of our civilization’s content without any real consciousness, intentionality, or communication.

There’s still a risk of that, and I think there will be many negative effects culturally from this technology as well, but Chiang’s essay, strangely enough, meta enough, is one of the few bright rays of hope I’ve had that human artists will remain. Not because of its content, but by dint of its sheer existence.

Huh? Why? How? What do I mean?

I’m saying that stopping worrying about AI art merely requires the realization that this is how everything already was. 99.9% of our human work was never really original to the degree it couldn’t be described as a kaleidoscopic remix. Not Chiang’s new essay, not my old one then, and not me now either.

And look, the sun still rose today. Cars still passed on the highway. Here in New England, the leaves will change into orange and yellow and then die and come again. An ant makes its way across a playground structure and will trace the same path tomorrow. The magnetic poles of Earth remain unflipped for the last 780,000 years. People still read my writing, even though you could get some of the same ideas from The New Yorker. The revelation that Near Total Unoriginality has been the regime we’ve been living in for a while now is the true take-away here.

Eyes firmly open, we can finally lock gazes with the Grinning Skull of Art: If you squint enough, almost everything mimics or remixes what already existed in culture (or what already exists on the internet). All those famous names of artists and writers you know were not constantly making original choices filled with divine inspiration denied to AI. Much like how Roger Federer won only 54% of points across all the professional matches he played in his life, and is still widely considered the greatest tennis player of all time, if a writer or a thinker can just be slightly more original than their competitors some of the time, then they become one of the artistic greats.

This nihilistic truth of low effective values for human originality leads to an oddly comforting view.

First, to put human artists and writers and thinkers out of a job, AI must win consistently more points than the best humans, who already operate on razor-slim margins of originality, insight, or creativity to begin with (much like Federer’s mere 4% advantage). The average essay is 99.9% unoriginal. A good writer like Ted Chiang might get that down to 90% unoriginal. I don’t think the models can get near that yet, despite producing impressive-by-themselves outputs, because those slim margins are the hard part.

Second, due to various factors like species-bias, distrust of AI, the unsolved unstable nature of longterm AI agenthood, as well as the necessities of audience building, it’s likely that human artists and writers and thinkers will, for the foreseeable future, maintain control of the means of distribution of ideas and creative output. Humans will have sticky control over audiences and the cultural narrative in exactly the way that large publications have a sticky control compared to smaller outlets or individuals.

Those two reasons—slim real creativity margins, and sticky control over distribution—explain why artists and writers (in fact, jobs of many types) have been surprisingly unaffected by AI’s new generative abilities. And I expect this will continue for a while, whether or not those outputs do become close to indistinguishable in quality, whether they’re ever truly original (they probably will be sometimes), and irrespective of whether they are philosophically “art” (due to that pesky lack of subjective experience).

We simply already exist in a world where fantasy books are blurry versions of Tolkien, or a world in which New Yorker essays are blurry versions of Substack essays. So the entire debate doesn’t really matter. Until they become better in a slim-margin business, then lacking the means of distribution and the stickiness of human creators, AI will be forced into the role of tool, not artist replacement.

Saying anything original once in a blue moon was, it turns out, the hardest part all along.

To be a scientist is to suffer from Gell-Mann amnesia constantly. I'm not saying that us neuroscientists have it worse than others, but the phenomenon of noticing discrepancies between the hype and the actual science of things is unavoidable. This is especially true when it comes to neuroscientific projects unfolding in the public eye, like Elon Musk’s company Neuralink.

For years now neuroscientists have been watching Neuralink recreate (to much fanfare) progress that has been known within the field of brain-computer interfaces for decades by academics. It’s also always been a tricky field for companies—right now there are hundreds of blind people stuck with useless or failing eye-implants with no one to repair them, since the companies have gone out of business.

That doesn’t mean I hate Neuralink! I'm not against the company's existence, since it actually does help people, and I'm not against the research itself, which might one day be significant. But I am against the outside perception that both:

(a) Something very new is occurring here (researchers have come up with ways for people to control computer games with just their minds for over 20 years).

(b) That Neuralink will develop as Elon Musk thinks it will, toward some sort of device that allows for “telepathy” between us and machines (a rather useless product given the research showing humans are limited to—no matter the medium—inputting and outputting information around talking speed, ~10 bits per second). We seem to be innately bottlenecked by biology, rather than merely inconvenienced.

Read more

A tale of the modern world. It was my son’s third birthday party, which was Curious George themed. That meant a huge Curious George decal on the wall, some yellow-and-red balloons, along with Curious George toys and books for presents. And of course, there were little party favors for the kids to take home. Into which we had planned to put Curious George stickers from a set we had ordered on Amazon.

Making the party favors the night before, my wife, with an appropriately curious look on her face, called me over while distributing the stickers.

“I can’t explain it, you just have to see it.”

Reader, this is what I saw.

At first, I had a laughing fit. Not only were many of them hilarious, imagine if we had actually handed them out! We’d have been the talk of the town, forever the people who gave “Bi-curious George and the closet” stickers to toddlers. Who, I wonder, would have gotten assault-rifle George? Little Lilly?

Perhaps, one might speculate, the stickers were hard for the company to catch because Curious George has always been a bit edgy for a modern figure. Somehow the series has survived intact without cancelation or updating, but in the first book, which is still extremely popular, George gets lured into a trap, captured, and taken from an idyllic existence in Africa by the suspiciously colonialist Man with the Yellow Hat.

George is rowed out to a waiting offshore ship, cheerily shanghaied from his homeland (note the big game rifle casually slung over the shoulder).

Later, after meeting with civilization, George smokes a pipe, causes general trouble, is put in a literal prison and has to escape via balloons, all before ending up at his “new home” at last: another prison in the form of a zoo.

Of course, the series has modernized since. Later books, of which there are many, studiously avoid mentioning how George actually got here, nor have ever explored if he misses his jungle home, and certainly have never delved into whatever venture initially brought The Man with the Yellow Hat to Africa to pull George from his verdant world.

The stickers are therefore far beyond any bounds of even early Curious George. So how did they come to be?

Well, I can first give you the proximal causes I deduced: the company is Chinese, and it produces a lot of stickers for all sorts of intellectual property (almost certainly without permission, I doubt Houghton Mifflin is signing off on these). They care not a whit for why Americans need stickers about superheroes or fairy tales or kids’ books, and so someone in China automated the process. The resultant stickers (it was a set of dozens) are a mix of either being AI-generated or, alternatively, a lesser form of algorithmic generation, like merely scrapping blindly from the web. Some dumb procedure searched Google and ripped off the images and the set got sent, without overview, to the whirring sticker machines.

But what about ultimate causes? What’s the real reason?

For me, this was a perfect example of a creeping high strangeness I’ve noticed in the 21st century. A strangeness I don’t remember from my youth, the world of the 90s and early 2000s. Funnily inappropriate Curious George stickers are merely one instance, a symptom of something larger.

Occasionally, this growing high strangeness will creep through and you will find products that look like they were designed by the insane, the unhinged, the mindless. One senses that there was no conscious human behind them. Just like these Curious George stickers, because if there were, the mistake would have been caught. They are the result of what happens when culture is generated unconsciously.

I’m certainly aware that, depending on political orientation, some might say “It’s actually because of late capitalism” or “It’s actually because of offshoring jobs,” and maybe both those are true. But it seems to me the more fundamental shift is that, at every economic and social scale, the workings of our conscious minds play less of a role. The growing high strangeness I sense is that culture is draining of human consciousness, and therefore of sense itself.

Back when I started this newsletter, I would occasionally receive replies to the emails themselves (rather than as comments here). Just little musings or reactions. I still receive such replies, but a chunk have been completely drained of consciousness. They seemingly come from real people's emails, but ones often suffering from, well, confused circuits. I’m pinged back by replies like “I’m thankful for your message. Let’s keep in touch!” or “Don’t hesitate to reach out, Thank you!”—replies that make no sense in the context of what was sent.

Once made explicit, consciousness drainage appears everywhere. Go under any viral social media post to find replies written in the bland uncanny valley of the inhuman.

Have you tried to talk to anyone on the phone at a major company in a while? It’s nigh impossible. All of us listening to Muzak as we wait to reach some tiny scrap of remaining human consciousness within some great machine that is plodding along by itself, often in ways that ape conscious behavior but are really just statistical approximations, and so go pear-shaped, throw oddities, surface up spindles of weirdness jutting into our everyday reality.

If I picture the last three hundred years as a montage it blinks by on fast-forward: first individual artisans sitting in their houses, their deft fingers flowing, and then an assembly line with many hands, hands young and old and missing fingers, and then later only adult intact hands as the machines get larger, safer, more efficient, with more blinking buttons and lights, and then the machines themselves join the line, at first primitive in their movements, but still the number of hands decreases further, until eventually there are no more hands and it is just a whirring robotic factory of appendages and shapes; and yet even here, if zoomed out, there is still a spark of human consciousness lingering as a bright bulb in the dark, for the office of the overseer is the only room kept lit. Then, one day, there’s no overseer at all. It all takes place in the dark. And the entire thing proceeds like Leibniz’s mill, without mind in sight.

We might cheer at the lack of missing fingers, but at what point does the removal of consciousness drain the world not just of meaning, but of sense? For consciousness drainage is happening all the time, everywhere, in every process, ever accelerating.

Another example: I eventually had to stop using my Chase credit card because it continually declined transactions for no reason. A recurrent problem, I called the company many times to tell them to stop putting holds on my accounts for obviously valid purchases, like groceries in the zip code where I live. Each time I was told it would never happen again. Each time it did. Eventually it became clear they had no control over it. Fraud decisions were made by an AI, you see, and AIs are mathematical black boxes. Inexplicable. Chase’s financial fraud AI, some internal kami-like spirit, had decided it didn’t like my card. And there was no way they could fix it because no one at the company had the power to reach in (digitally) and re-wire the damn thing. Probably not even the people who trained it.

So now I only use a debit card, because that’s what you’re forced to do in a culture where not just “technical debt” accumulates but “unconscious debt” does as well.

In a cosmic irony, the running joke of the Curious George books is that George is fundamentally disconnected from the world around him. The Man with the Yellow Hat tries to rein him in, teach him, tell him ahead of time what to do and how to act, but George, well, George is a curious little monkey. And curious little monkeys get into trouble.

Yet for all his constant trouble finding it’s not his curiosity that’s really the problem. No, George gets into trouble because he doesn’t actually understand the systems of the world around him. Despite his hapless ignorance, everything works out. All the stories have happy endings, if only because George is god-tier lucky. It is authorial intervention alone that saves him from being a wandering idiot constantly burning down forests and destroying heavy machinery and shattering priceless artifacts.

Did you know, across all the books, Curious George never speaks? He can’t, you see. He’s a monkey. George has no language, no higher cognitive abilities. He displays no continuity, and cannot learn from his mistakes. And so the whole modern world—his museum trips, his visit to the candy factory, his fire engine adventures—is forever an exercise in high strangeness to him.

And now, I guess, it is for us too. We’ve all become Curious George, primates living amidst alien machinery far from our homelands.

Fall is almost upon us. The dry season for blogging, the summer, is nearly over. And it’s an election year. I’m sure bloggers and writers everywhere would like you to prepare your bodies and minds and inboxes for shattering insights, peak content, mental stimulation.

But, ah, before that happens, right now in this last spell of dryness, I’d like to talk about why most of that content is going to be dumb. Bad. Useless.

Hypocritical, I know. I’ve written many times before about the benefits of blogging, both why I do it but also why I think others should too. Advantages like speed and relevancy. Most importantly, freedom from a class of publishing gatekeepers who get it wrong more often than they get it right. Personally, I think this here is a good medium, all things considered.

But I’ll also be honest about it: despite much boosterism by its proponents (including myself) blogging is not a perfect medium. Its positives, of which there are many, do not mean its downsides don’t exist. There are temptations, pitfalls, and limitations, things many, even those with good reputations, would like to pretend don’t exist and probably wouldn’t like discussed, lest it reveal a bit too clearly their true shadow shapes.

Good.

Here are what I see as the biggest problems with blogging or writing a newsletter (of any kind), inevitable dirty secrets that aren’t talked about enough. Knowing them is useful—not just for writers, but for readers as well. For purposes of discernment, of course. Beginning with the unfortunate reality that…

Blogging leads to aesthetic degradation.

I used to think that internet writers all sounded the same because a minimalist style had crowded out other more interesting but also more fragile aesthetics. Like some sort of algae bloom of prose filling up websites. Now I realize that the most common sound you hear on the internet, if you lean close to the words, is the sound of someone typing quickly.

Read more