Robin Sloan
main newsletter
October 2026

Peeking over the wall

Stag at Sharkey's, George Bellows, 1909
Stag at Sharkey's, George Bellows, 1909

The olive har­vest is here, and it’s early! In years past, we started pro­duc­tion in the last week of October. This year … we’ve already started!

Thus begins my season away from the keyboard, which is many things: har­vest season; busy season; scheming season. Agatha Christie said, “The best time for plan­ning a book is while you’re doing the dishes.” Well, the olive har­vest is basi­cally a tremen­dous load of dishes, six weeks long. Lots of work, but/and lots of space for dreaming, too.

In the cool dark, I drive our old Toyota to the mill, the radio set to 94.3, Desh Punjab Radio. During the morning hours, the sta­tion plays Sikh prayers, with the crackly lo-fi sound of dubs from old tapes. Good way to start your day.

It’s likely the next mes­sage you receive from me will be my annual gift guide in late November. This edi­tion is rel­a­tively light — just a few links before they get stale.

I’m Robin Sloan, a writer, printer, and manufacturer. This is an archived edition of my newsletter, originally transmitted in October 2026. You can sign up to receive future editions using the form at the bottom of the page.

As usual, this newsletter has a few dis­tinct parts. Here’s what’s ahead:

Note from the print shop

Just a quick word of thanks to everyone who pur­chased my Summer Work­shop 2026 package. It was a smash suc­cess. I had so much fun making the books, and of course it has been very sat­is­fying to hear that you all received and enjoyed them. (A handful of inter­na­tional mail­ings are still out there, wan­dering the global postal system, pos­sibly confused; if one of them is yours, please reply so I can dis­patch a replacement.)

If you missed out, do not for a moment despair: I’ll pre­pare a fresh edi­tion of Mr. Penumbra’s All-Night Reading Room as soon as I’m back in front of my Risograph. I would also like to find a good/cool/sat­is­fying option for a dig­ital edi­tion.

When I was in third or fourth grade, I worked with my next-door neighbor Steve, a year older than me, to pro­duce a mag­a­zine we called S’n‘R. We laid it out on my family’s newly-pur­chased Mac Plus, then printed per­haps three copies, selling them all to indulging adults.

I have been pub­lishing ever since, at every sta­tion of my life, in print and online, some­times pro­fes­sionally, more often because I just couldn’t help myself. The clarity of this voca­tion has grown, as other pur­suits have come and gone but pub­lishing has remained. These last couple of years have been so fun and so suc­cessful that nat­u­rally I wish I’d gotten my act together sooner, started my own print shop long ago … but the truth is, I didn’t have the skills yet, which aren’t only cre­ative, but also mechan­ical and logistical.

So, it’s a partnership: the third-grade pub­lisher meets the grown-up busi­ness operator, and they make a great team.

Ask for what you want

For my tech­nical blog, I wrote a long post (maybe a mini essay) with the title: “Ask for what you want” is a key skill for the 21st century. It’s about AI, a way of thinking about what it offers, here and now in the autumn of 2026. In the post, I explain that

[ … ] I’d like to pro­pose an exploratory program. This is rec­om­mended for any­body/everybody, but more point­edly for cre­ative people who are wary about AI. I’m wary, too — cautious and critical — yet I believe my expe­ri­ence this summer edu­cates and enlivens my criticism, rather than under­mines it.

Over the summer, my cre­ative work was electrified — not flattened, not threatened, but electrified — by the new capa­bil­i­ties of AI agents. (And when I say “new” I mean it: new as of June 2026. This is not last year’s AI.) It has been not only fun but hugely pro­duc­tive to build cre­ative tools for myself: tools that would be worthless, pos­sibly incomprehensible, to anyone else.

This new post is a sort of sequel to my pop­ular post about apps as home-cooked meals, and the arrival of AI into the kitchen doesn’t change the core premise, which is about the rad­ical feeling of truly per­sonal software.

I know there are more than a few “cre­ative people who are wary about AI” reading this newsletter, so, this post is for you.

Me & Kevin & Jack

Last Sunday, I was Kevin Roose’s inter­locutor at Kepler’s in Menlo Park, for the launch of his book tour for The AGI Chronicles, a his­tory of the AI race written nearly in realtime. It’s such a cool book, and a deeply useful one. Atop the blurbs, I was delighted to spot this assess­ment from Alexis Madrigal:

The one-volume his­tory you need of the cur­rent industry, from the early OpenAI and Deep­Mind exper­i­ments to today’s rea­soning models and vibe coding. It’s a book you can just hand to someone who wants to under­stand what’s truly at stake with AI … A really fan­tastic book.

I thought the same thing, reading The AGI Chronicles: this becomes the book I want to hand to anyone curious, or cautious, or hyped, or fearful, about the AI boom. It tells the story in tac­tile terms — real people in real places — yet it doesn’t dumb any­thing down. Quite a bal­ancing act.

One of the book’s plea­sures is its glimpses of the doc­u­ments — a knowable, name­able bundle — that ener­gized and directed AI’s devel­op­ment over the past decade. These were nerdy memos that really did change the world, and, for Kevin, gath­ering and pre­serving them was an impor­tant part of this project. He cir­cu­lates one in a recent newsletter: Dario Amodei’s Big Blob of Compute … which sounds ripe for a children’s book adaptation, doesn’t it? Kevin writes:

If AI trans­forms society the way Amodei thinks it will, doc­u­ments like these may be the equiv­a­lent of Darwin’s notebooks, and I think there’s a strong public-interest argu­ment for let­ting people read them.

I mentioned that Kevin’s launch was at Kepler’s, and folks, let me just report — remind you — reassure you: nothing, and I do mean nothing, beats a big, bright book­store full of people.


A couple of weeks before that, I sat down with Jack Clark for the inau­gural install­ment of his Fic­tion and the Future series.

I got to know Jack way back in the pri­mor­dial days of “deep learning”, before any­body ever pre-trained a transformer. I’ve recounted his story often, and I’ll do it again here, because I think it’s an incred­ible example of fol­lowing your gut. As a jour­nalist at Bloomberg, Jack found him­self on the “data center beat”, increas­ingly com­pelled by the scale and ambi­tion of modern computing. At some point, his reporting and writing con­vinced him that AI was THE story of (1) his career, and pos­sibly (2) the 21st century. The dis­tinctive and Jack-ish twist is that he then asked him­self: Well, do you want to write about it … or do you want to participate? He joined the research outfit OpenAI as their policy chief. OpenAI in that era mainly pro­duced bots that played Starcraft, so, Jack still had a bit of free time to drink beers and walk around reservoirs, which we did, and he has the dis­tinction of being the person to inform me that the Age of the Hob­byist was ending and — here, con­jure Galadriel’s voice at the start of The Lord of the Rings — the Age of Scaling had begun.

My own exper­i­ments with AI spun down as the industry — for it became an industry then, driven espe­cially by Anthropic, the com­pany Jack co-founded with six col­leagues — sped up.

I had a fab­u­lous time chat­ting with Jack — you’ll see it plainly in the video — mostly because it felt like no time had passed at all, which is a wild thing to say when you con­sider the events that HAVE occurred, their consequence. Tes­ta­ment to his basic groundedness, I think. If you work in or around tech, you will inevitably see some former col­leagues and old friends get slurped into the pneu­matic tube of power and wealth; often, the effect is distorting, overall cautionary. But Jack wears his suc­cess very well. The useful anti­bodies turn out to be: curiosity, irony, a taste for the weird. You find all of them in Jack’s weekly newsletter, where he com­bines links to new AI research with his own flash fic­tion, which has been a strong influ­ence on me over the years. Moonbound wouldn’t be quite the same book if I hadn’t been reading Jack along the way.

It was a fun evening. If you’re inter­ested in AI and creativity, or sci-fi and the future, or in seeing me come close to tears as I contemplate the tragedy that Borges didn’t ever get to play with an LLM, check it out.

Preliminaries, George Bellows, 1916
Preliminaries, George Bellows, 1916

Here is a great edi­tion of Diana Kim­ball Berlin’s newsletter, a dig­ital commonplace. Every week, she shares five fragments, each accom­pa­nied by a brief con­sideration, almost like a wall label at a museum … or like a New Yorker car­toon caption, maybe. The format works so well — it’s a plea­sure to follow Diana’s eye and ear through the internet. Highly rec­om­mended.


What a treat, to dis­cover that this col­lec­tion by Romanian poet Dan Sociu was trans­lated into Eng­lish by my old high-school class­mate Monica Cure! Reading the book’s listing at Asterism, I thought, “Wait … I remember that name”, and, sure enough!


I’ll con­fess I initially clicked over to the Asterism site to inves­ti­gate this book: Wulf, My Wulf, An Anthology of Wolf & Were­wolf Folk­lore from Anglo-Saxon England, Viking Age Scandinavia, & Beyond. I’m sold!!


Yun Hai is BACK with another gor­geous lunisolar almanac, avail­able for pre­order now, cloding October 13. All year, I have enjoyed tearing out these pages, noting with plea­sure the almanac’s daily directives. It’s a fun thing to hang on your wall, a supertac­tile com­po­nent in a daily routine.


The new video game Towers of Babble from Frank Lantz and crew looks amazing — a tower defense game built around a vast vocab­u­lary of words. So like … Scrabble meets Kingdom Rush? Sort of? I haven’t yet down­loaded the demo (see: olive har­vest) but I’m eager to do so.

Frank is a genius, and his newsletter always sort of melts my brain a little, in a good way.

Case in point: a recent edi­tion about the game of Go’s evolution in the years after AlphaGo. In the newsletter, Frank asks the Go expert Nate Morse to recount a recent match between the human world champ (who is widely regarded as the best Go player EVER, cer­tainly the best in modern his­tory) and one of the best cur­rent Go pro­grams (vastly more capable than AlphaGo), and their dis­cus­sion touches … everything! Creativity, technology, curiosity, humility, his­tory … it’s great.


Here is your fresh In Our Time pod­cast episode rec: The Inven­tion of Photography, from July 2016. This episode is full of crunchy his­tor­ical details, and it con­nects to so many domains … some you’d expect — technology and art — and others you might not — sci­ence and government. That last one is interesting; I never knew about the French government’s inter­ven­tion in photography’s early devel­op­ment.

Note that you can find the same episode on Spotify, Apple Podcasts, etc., with a quick search.


Along the same lines, here is Ken Liu on the his­tory of copying in art, all the way up to and including AI today. Don’t ready “copying” as an accu­sa­tion or slur; one of photography’s revolutions, after all, was that it allowed many copies to made of the same negative. And every book you ever read and loved was, of course, a mechan­ical copy. Ken’s con­sideration is nei­ther lament nor celebration, rather just … con­sideration! Careful and provocative.

I read Ken’s new novel All That We See or Seem over the summer, and I judge it the first near-future sci­ence fic­tion to (1) wrap its head around the recent devel­op­ments in AI, and (2) extrap­o­late them for­ward both plau­sibly and pro­duc­tively — i.e., in these pages, Ken man­ages to con­jure devices and expe­ri­ences that push out beyond our sci-fi present, feel weird and “whoa!” again.


These paragraphs from the far-seeing sci-fi writer Vernor Vinge are stir­ring to read in this moment, cour­tesy of Michael Nielsen.

Vinge’s “technological sin­gu­larity” has been much-predicted and much-discussed, but his sub­tlest insight is some­times overlooked: that, at a cer­tain point, com­pounding change pro­duces “an opaque wall across the future”.

I under­stand the task of a writer like Ken Liu — and like me, some­times — as standing on tip-toes to peek over that “opaque wall”, when nobody else can … 


Click over to this edi­tion of Eng­lish Teacher Weekly and search on the page for “Ahab” to read about Cap­tain Ahab’s evil brother Derek, and to explore the possibility: what if Ahab was hot??


Here is Marc Wei­den­baum on craft, process, and networks. I really love this part:

The first is don’t let anyone or any­thing, or any­thing acting like anyone, or anyone pro­moting any­thing, con­vince you that craft is out­dated or will ever be out­dated; you will only improve, only make the sounds you imagine in your mind’s ear and feel in yourself, by working steadily to achieve them. Also rec­og­nize that expected paths lead to unlikely ends the more patiently and curiously, guided by instinct and expe­ri­ence, you walk those paths. There is always an excuse to not put in the work. If you allow “AI” (I put it in quotes because it’s less a spe­cific thing, in this context, than what is often cast as, to borrow a his­tor­ical term, a specter) to become that excuse, well, it’s unfortunate, but there would have been other excuses. So, the best thing to do is put the threat you may per­ceive from AI in a tidy little box full of all those other excuses, don’t dis­tin­guish it from them, and get to work.


Quite a title for an eco­nomics paper, almost onomatopoeic: Immiserizing Automation. Doesn’t that just SOUND like a robot that makes you poor?

The argu­ment has to do with automa­tion’s inter­rup­tion of human skill formation:

Immis­er­a­tion occurs because junior automa­tion shifts earn­ings to later in life. This shift lowers the present value of expe­ri­ence-intensive careers, so that fewer workers pursue them, and the supply of expe­ri­enced workers falls.


I feel like a lot of people — most people — nearly all people — don’t actu­ally know what phys­ical automa­tion looks like, inside modern fac­to­ries and pro­cessing plants. Here, for example, is a water bot­tling line, somehow both hyper­ef­fi­cient and totally Rube Goldberg.


Another com­pelling aca­d­emic paper title: Black hole sin­gu­larity is a sur­face not a point. I suppose the missing comma got too close to the event horizon … 

Black holes con­tinue to be so interesting, and so pro­duc­tive scientifically — as well as imaginatively — because they are the archetypal, unbeat­able “limit case”.


From Ingrid Burrington, a bril­liant observation:

[ … ] I have a long-percolating idea to trace the his­tory of the trope of a guy who gets so mad he punches a hole in the wall and its rela­tion­ship to trans­for­ma­tions in construction. Did dry­wall give men per­mis­sion to punch the wall, or simply make it more destruc­tive than an existing tra­di­tion of breaking one’s fist against brick or wood? Was it a response to the under­lying hol­low­ness of the postwar condition, reflected in the hol­low­ness of its building materials? What hath gypsum wrought, etc.

That’s over at Perfect Sentences.


Here is type designer Stephen Nixon with the story of the making of AT Kyrios, the “slightly psy­che­delic blackletter” that I use on my web­site! If you have ever clicked over to the web ver­sions of one of these newslet­ters, you have seen it.

This is a great cre­ative story: a long-simmering aes­thetic con­cept cat­alyzed by the let­ters carved into a par­tic­ular mon­u­ment in Brooklyn’s Green-Wood Cemetery.


I like this new type­face from Taylor Benton called Bang!—evocative of cross-stitch, also pixel-y, not un-gloppy.


For all the years of my child­hood I spent glued to a Game Boy, and for all my reverence, as an adult, for the inge­nuity of its creation — “lateral thinking with sea­soned technology”—I never paid any atten­tion to the words and glyphs printed on the actual device. There was a lot going on! … yet somehow it totally works.


Alan Jacobs is teaching four big books and, as usual, you wish you could be there. The books are:

Alan notes … 

The dates I have given here are the dates of com­plete book publication, but the first readers of these novels didn’t read them as books: they read them in install­ments. Lost Illu­sions was pub­lished over the course of six years, the first part as a stand-alone book, the rest in mag­a­zines. Mid­dle­march was pub­lished as eight small books over the course of a year. Each of the two Russian novels was pub­lished a chunk at a time over the course of many months in a peri­od­ical called The Russian Messenger. People had time — for many of them, far too much time! — to reflect on each install­ment and long for the next one to appear.

 … and of course I didn’t know ANY of those books were serialized.

Some of you might recall that I wrote a serial novella, pub­lished daily on the front pages of San Fran­cisco Bay Area news­pa­pers in the summer of 2020 — an expe­ri­ence that remains one of the most elec­tric of my pro­fes­sional life. Hmm. Hmm!


The Mal­tese Falcon is a book that I did know was serialized. However, I didn’t know about this writing trick of Dashiell Hammett’s:

What he didn’t see a lot of was style, a cer­tain “magic” in the prose. [ … ] Thus he began to type on strips of paper half the normal size. This forced him to put down choice words, and if he felt they didn’t work he could toss out the 15 lines or so and start again.

Here is the cover of The Mal­tese Falcon’s first install­ment in Black Mask mag­a­zine—too much fun.


I some­times think the deep secrets of human expe­ri­ence might all be waiting in sleight of hand, in magic tricks … the way they take hold of our atten­tion so smoothly, so predictably, so powerfully.

This is so good—a priv­i­lege to watch.


HEY that’s a nice-looking nebula!


The note I made for this link was simply “ancient stur­geon insane” and I think that about sums it up.

New Society Dinner Card, George Bellows, 1923
New Society Dinner Card, George Bellows, 1923

The painting at the top of this newsletter is one I have looked at easily a thou­sand times, one of the Cleve­land Museum of Art’s high­lights dis­played on the col­lec­tion search page. Every single time, I thought, “Wow … what a great image … ”—so why didn’t I snag it for a newsletter until now?

George Bel­lows was amazing; here’s the full gallery at the Cleve­land Museum of Art, another at the National Por­trait Gallery. Stag at Sharkey’s is a wonder, but I like his pencil line even more. His humor, too.

Phys­ical art for a phys­ical season. Here we go!

From the mill,

Robin

P.S. You’ll receive my next newsletter in late November — my annual gift guide!

October 2026