i  <esau

"Oh, glory," she whispered, and went to see what lay below on that side. Passing a dip in the rock, she halted. A pool of some eerie substance was cupped there. It shimmered with green, yellow, gray, and blue lights, much like the colours that she'd seen in the sky the night before. They moved over its surface in globes, waves, or strips. Watching the pool made her giddy. She swayed, and put out a hand.

This is an excerpt from The Realms of the Gods by Tamora Pierce, a beloved fantasy novel from my childhood that I re-read recently.

Having previously been in the psychedelic world of Blood Meridian, Pierce's prose is comparatively direct. Scenes are painted economically as the romp careens along, but ocassionally, they'll give me pause to wonder: what would that actually look like?

Unfortunately, on the following aphantasia scale, I think I'm around a 2.

It's interesting watching Samurai Jack and trying to figure out what the cartoon violence rules were. A film studies version of petals around the rose.

Famously, most of Jack's battles were with robots so the audience rarely sees blood, just oil.

But even still: which ways are you allowed to depict fighting?

Looking at the fight between Jack and The Guardian, a few rules can be inferred.

• Weapon contact with Jack, when it happens, must be brief.
• Ripping Jack's clothes superficially is okay (this is kind of like his health bar.)
• If no one is harmed by the attack (e.g. because it's dodged,) it's okay.
• You can imply that Jack has been hit off camera.
• Avoid wide shots of violence that show the wind up, contact, and aftermath.

I been noticing more smarter AI in customer service these days.

A recent example: I docked my BIXI but the thing glitched and the ride didn't terminate. I only realized when I tried to get my next bike out.

In the past when this happened to me, I just called up BIXI support, went through a short phone menu, and explained the situation to someone on the phone. They verified my identity and manually terminated the ride so I could go on my way.

Basically, the ideal phone-based customer support experience. The wait wasn't long, the procedure was efficient, each step made sense, and it solved my problem.

This time, the app prompted me to ask for support via "chat" first. Historically, this could have meant all sorts of things: a naive keyword recognition program that would recommend FAQ entries to me; a tech support worker from a country with a low median wage pasting in responses from a script, 5 tickets at a time; occasionally, the type of person I'd been talking to on the phone, but over text.

This time, it was some sort of LLM. It dynamically and patiently walked me through the identification and triage process. It handled my typos and information I sent bit-by-bit.

Eventually, it concluded the issue needed human involvement. Presumably at the end of its instructions it has a "If asking the customer to jiggle the bike doesn't fix the problem, request that they phone a human." part of its prompt.

This was fine. The trouble was that I then had to do all the same stuff with the human. The same identification process, the same identification of the nearest street intersection, the same bike jiggling. Finally, the human terminated my ride.

It would clearly be better if I could call the human and have them say "I see you've already gone through all the steps with our AI. I'm going to unlock the bike now."

Better yet (though not something BIXI alone can implement), identifying myself is done via a two-factor prompt on my phone (get on this, governments!) and I can use text-to-speech to compose my messages to the AI. Maybe we'll even eventually find them reliable enough to let LLMs terminate rides themselves!^🤖

As it stands, the experience feels like costs are being cut. It doesn't need to. Automated customer support could be faster, cheaper, and more convenient than the best phone support I've had.

I'm still skeptical of this future. The companies that have already automated as much of their customer support as humanly possible are miserable to deal with, even though in a sense I'm grateful for the downward pressure they put on prices. These are typically companies of enormous scale. The BIXIs of the world might be able to take a different path.

In Montreal, there’s a number of small, enclosed, wooden shelves on street corners where you can “take a book, leave a book.”^📚

In the past two days I got some fucken scores:

That’s Mathematics For Elementary School Teachers (1987 Ed.) and Webster’s New Universal Encyclopedia (1997 Ed.)

I love them. It’s so nice to just read a big reference book.

In the maths book, the typesetting and layout is excellent. It feels so carefully crafted. Plus^➕ elementary level maths is very approachable. Reading it feels like playing Breath of the Wild, where every new page is something interesting and joy-bringing to think about.

The encyclopedia also rules. I love Wikipedia rabbit holes as much as the next guy, but there’s something so arbitrary and serendipitous in reading an alphabetically arranged list of 25,000 important things. It also has human errors that are charming. They misspelled Vostok here:

Both books are also dated in their ways. The maths one was written for USAmerican teachers to learn how to teach maths, yet it only uses the metric system - an artifact of a more hopeful time for mathematics educators, I suppose. The encyclopedia is missing almost everything that has happened during my lifetime. Here’s what it says on AI:

Maybe it’s nostalgia or novelty, but learning from a book feels so much better than reading the equivalent text on a website. Even the highest quality, most beautiful maths website I know of doesn’t excite me the way it does to read a physical object that had thousands of person-hours poured into its production.

To whoever put these books in the local little free library, thank you ♥

A lot of people on Mastodon are there because they believe decentralized and open-source techologies are essential tools for resisting coercive power. The fediverse is innately resilient to censorship, advertising, and hostile takeovers.

For these reasons, I made the switch from twitter sometime in the early 2020s, but I think it's important to remind ourselves that these virtuous properties are not a panacea to all the ills of social media.

It doesn't matter whose database the posts exist in: we can all still be ragebaited. It's a fundamental vulnerability of our psychology, with or without Elon Musk intentionally exploiting it.

I've maintained for a long time that the good part of Mastodon is that it's kind of shit by default. You have to put quite a bit of effort into finding people to follow and tweaking your settings to get it to surface a lot of interesting content. I haven't done this. I follow a few randos and some bots that post clipart.

I like the fact that when I check the fridge¹, there's just the same head of cabbage that was there last I checked. Paradoxically, the better Mastodon gets, the more I'll check it, the less good for me it will be.

Recently, someone I follow started boosting posts I found to be confidently wrong. I started getting emotionally invested in the thoughts of complete strangers. Getting provoked. I wanted to write replies to correct their misconceptions.

I've mostly been able to resist: it would not change anyone's mind. It would take a lot of time. I would get even more emotionally preoccupied by a hamster wheel.

Instead, I think I'll employ another cool thing about open-source technology: you can usually adapt it in ways to suit your needs². I can filter this person's boosts and hopefully that will be sufficient.

But there's a tension here. For the structural factors outlined in the introduction, I want the fediverse to grow. For it to grow, it needs to attract users that aren't already sold on the structural selling points - it needs to be better. For it to be better, if we optimize for engagement, it could easily become worse.

A ragebaiting machine that people can tinker with is still better than one they're forced to use, but I will maintain my wary distance in either case.

An honest answer is “all kinds,” but that’s also an annoying answer because then you are stuck wondering

“So is this guy a music fan or is he just undiscerning?”

I think a better answer is to say something like

“When I was a teenager I became obsessed with music and that’s never really changed. I liked Queens of the Stone Age and a math rock band called Hella. Now I listen to all sorts of things. I’ve been listening a lot to a folk musician named Hannah Frances recently.”

There. That gives lots of threads to whichever son of a bitch that’s trying to get to know me.

But the honest answer is also interesting to me, because as a teenager, I wasn’t as omnivorous as I am now. I once had a stronger sense that there was music I didn’t like, and it was important to signal as much.

It also felt important to recite genres in the way I’d learned to monastically catalogue them.

“Oh, you know, avant-garde hip hop, breakcore, tuvan throat singing.”

Reflecting on this, I think I was genuinely hopeful that people would continue to want to talk about music with me after responding in such a way. I just didn’t realize how pretentious it sounded (or maybe I did and I was simply a snobbish scoundrel.)

What’s changed?

For one, I don’t need my taste in music to make me feel unique among my peers. I moved to a country far away so now I’m unique for free!

I’m also just a lot more eager to find things to appreciate in any particular piece of music. Not because of some striving-patrician insecurity, but a curiosity about what it’s doing and how. Being a musician is a gift in that way. I love listening to music with someone and pointing out little details in the composition, arrangement, or production that make the music special.

That’s not to say I’ve obtained aesthetic enlightenment and love all music equally. Plenty of tunes I encounter make me go “Hm, kinda boring,” but I take that as a skill and/or effort issue instead of a reason to look down on it.

Is this all just some subtler flex? Some more impressive sensitivity? I don’t fucken know, man. I’ve been listening to Hannah Frances a lot lately. All kinds. There’s so much music to love.

There's nothing new to say in any of these "Americans Can't Read" pieces, yet still I read them with morbid fascination.

To zoom through the typical argument: if you count subtitled TikToks and Temu product descriptions, everyone still reads. It's just the novels and newspapers that are suffering. A literary elite reads more than ever thanks to e-readers and substack, but the global median trends downward. This is bad for our brains and our university professors' sanity, and portends a post-literate, thoughtless society.

If we've passed peak literacy, we're still literate enough to consider this a universal problem. We still hold in our minds the idea that everyone ought to read, and ought to enjoy novels, and that this grants capabilities valuable for the flourishing of the society we want to live in.

I note that there's a blindspot here. The counterfactual world where we felt this same way about maths, or computer programming, or herbalism, or animal tracking. I suspect there are few New Yorker writers who know how to program, so there aren't as many eulogies for this atrophying human capability, even though I have no doubts they would be accompanied by their own alarming neuroscience and test score crashes.

We don't often mourn the loss of a capability we never developed. Sometimes people regret not learning an instrument as a kid, but it rarely feels salient day-to-day that I'm lacking in appreciation for the abstract mathematical brilliance undergirding reality.

This is probably true of any intellectual specialization. Statisticians, musicians, physicists, mathematicians, historians, doctors - if I had to guess, they're all concerned by the second-order effects of thought delegation - it's just that reading & writing were once universal enough to be considered foundational for liberal society in a way that understanding physics never has been.

I'm not yet sure what to make of this. It's just something I try to keep in mind as we hem and haw about the future: there are selection effects at play to the crises we're concerned about, and loss aversion is a powerful cognitive bias.

Did you know that Fenix has started pretending that he speaks with a British accent because people said RUBBERZ sounded like a song by The Smiths? At one point in the Channel 5 interview, AI music booster Timbaland calls him up to strike a deal and he responds with an I'm Br-ish innit accent like he's Ali G from the mirror universe.^🇬🇧

This shameless opportunist is unwittingly creating Brass Eye in real life, a kind of performance art so deep that the creator isn't even conscious of it.

I'm fascinated.

AI and creativity

A week ago, I saw an experiment conducted with LLMs and generative video models. Each LLM was given a budget and told to make a music video for Bruno Mars' Uptown Funk. Cue four rounds of uninspired slop: hyper-saturated and clean, oneiric details that collapse upon inspection, never showing the same thing twice. To the extent that the directorial philosophy of these models shines through, it's "interpret the lyrics as literally as possible, every single shot."^📈

Fenix Flexin is the inversion of this experiment. A human-made music video for an AI-made song - the WAY better direction. For one, it's less exhausting to watch a video that has object permanence. For two, an actual human then has to make the stupidest lie on camera over and over.

Why can't he admit it? Because eating crow would be embarrassing. His biggest song ever was written entirely by AI. He struck gold by accident and has no talent. What he does have is some special blend of unconsciousness and hustle steering him down this path that goes nowhere good. Either he makes more AI music, which will expose him and not be as good as RUBBERZ. Or he makes his own music, which will expose him and not be as good as RUBBERZ.

In the meantime, man, so funny.

It's clear to me that we're on the cusp of a new genre of aesthetic achievements. I continue to support and admire artists who avoid touching AI in any shape or form, but they will in some sense be doing more of the same. Those that use it, or are in conversation with it, or are caught cheating with it will be doing the thing that I think will get more attention in the proverbial history books^📚 decades from now.

And so in 2026 came Fenix Flexin. In an era of bullshitters and brainrot, he won a race of sorts by crossing the lines of cheating, grift, microcelebrity, propaganda and hyperreality all into one enormous knot beyond his understanding. The marketplace of ideas had completed its transformation into a casino.

There's a viral tiktok song called RUBBERZ that is clearly generated by some low-rent generative audio model, presumably via the prompt "drake + pina colada song", where the artist behind it gave the most obvious lipsync ever to prove that the song isn't AI. He then did an interview with Andrew Callaghan where he atonally butchered his own lyrics.

It's like the whole world is getting Boratted except everyone can see through it and his brazen refusal to admit that he's lying only intensifies the spectacle.

This is the unholy dynamic powering much of the attention economy. Louis Theroux also studied it in his Manosphere documentary: the most shameless attention-maximizers can succeed because the algorithm lacks any ethical filter. As long as it's not illegal (or too close to kink), whatever captures people's attention is fair game.^🧊

I'm not plugged in enough to know if this is a tiny blip in the ocean of otherwise forgiveable content, or one bubble in a frothy wave of AI slop that is destroying our ability to understand reality. One data point is that comments left prior to @medasin's exposé seemed to earnestly enjoy Fenix Flexin's masterpiece:

So unless youtube comments are more of a dead internet ant nest than I want to believe, I worry for our illiterate and bamboozled teenagers.

This shit is straight out of the 80s and 70s 😂

As I always say, maybe it's fine. My generation had Friday by Rebecca Black and MLG memes and we're only somewhat fried and responsible for things like the current US administration.

There's something to the song: the fucked up metre, the vaporous clichés, the out-of-genre adlibs that Fenix actually ad'd himself.

Plastic smiles on Sunset Boulevard

Livin' fast and playin' every card

Broken hearts and bitter bills

Swallowin' these jagged little pills

Your messages are floodin' in (Floodin' in)

Thinkin' that you're gonna win (Gon' win)

I just let the rubber band snap

The Ouija board jelly-bean-guessing-game that is The Algorithm discovered something novel. Something ineffably noteworthy. It's given this rando his zero point five seconds of fame for it and will spit him out again shortly, giving all of us one more broken ceramic shard to make our mosaics of the 2020s with.

If it's not already happening, I won't be surprised when people start making art that intentionally emulates AI slop to argue that actually it's significant and interesting that a large fraction of people now think that the main thing about the 70s and 80s — two decades of complex, vivid, horrible and joyous human existence — is that they had particular cameras and instruments.

I think about the miracle of libraries often.

And I think something similar as I hear blips of Esther Perel tragically slinging furniture or clothes at me as I mash my +30s button: “oh my god, it fucking rules that we can skip ads.”

“Wherever you get your podcasts is a radical statement”

Libraries are an institution and podcasts a protocol, but they're both a gift to us, a path dependency first trod by our idealistic forebears.