2026-10-12 08:00:00
Kevin Kelly's The Inevitable articulates a handful of macro technological trends and follows them to their logical conclusions. It's an insightful read -- I highly recommend it.
One decade later, here are some AI-related excerpts from the book:
Around 2002 I attended a private party for Google—before its IPO, when it was a small company focused only on search. I struck up a conversation with Larry Page, Google’s brilliant cofounder. “Larry, I still don’t get it. There are so many search companies. Web search, for free? Where does that get you?” My unimaginative blindness is solid evidence that predicting is hard, especially about the future, but in my defense this was before Google had ramped up its ad auction scheme to generate real income, long before YouTube or any other major acquisitions. I was not the only avid user of its search site who thought it would not last long. But Page’s reply has always stuck with me: “Oh, we’re really making an AI.”
I’ve thought a lot about that conversation over the past few years as Google has bought 13 other AI and robotics companies in addition to DeepMind. At first glance, you might think that Google is beefing up its AI portfolio to improve its search capabilities, since search constitutes 80 percent of its revenue. But I think that’s backward. Rather than use AI to make its search better, Google is using search to make its AI better. Every time you type a query, click on a search-generated link, or create a link on the web, you are training the Google AI. When you type “Easter Bunny” into the image search bar and then click on the most Easter Bunny–looking image, you are teaching the AI what an Easter Bunny looks like. Each of the 3 billion queries that Google conducts each day tutors the deep-learning AI over and over again. With another 10 years of steady improvements to its AI algorithms, plus a thousandfold more data and a hundred times more computing resources, Google will have an unrivaled AI. In a quarterly earnings conference call in the fall of 2015, Google CEO Sundar Pichai stated that AI was going to be “a core transformative way by which we are rethinking everything we are doing. . . . We are applying it across all our products, be it search, be it YouTube and Play, etc.” My prediction: By 2026, Google’s main product will not be search but AI.
In 2017, Google researchers published Attention Is All You Need, which introduced a deep-learning architecture called the transformer.
As it does, this cloud-based AI will become an increasingly ingrained part of our everyday life. But it will come at a price. Cloud computing empowers the law of increasing returns, sometimes called the network effect, which holds that the value of a network increases much faster as it grows bigger. The bigger the network, the more attractive it is to new users, which makes it even bigger and thus more attractive, and so on. A cloud that serves AI will obey the same law. The more people who use an AI, the smarter it gets. The smarter it gets, the more people who use it. The more people who use it, the smarter it gets. And so on. Once a company enters this virtuous cycle, it tends to grow so big so fast that it overwhelms any upstart competitors. As a result, our AI future is likely to be ruled by an oligarchy of two or three large, general-purpose cloud-based commercial intelligences.
Kelly predicted AI heavyweights while advocating for a future of open weights:
Ideally, this additional intelligence should be not just cheap, but free. A free AI, like the free commons of the web, would feed commerce and science like no other force we can imagine and would pay for itself in no time.
Amid all this activity, a picture of our AI future is coming into view, and it is not the HAL 9000—a discrete machine animated by a charismatic (yet potentially homicidal) humanlike consciousness—or a Singularitan rapture of superintelligence. The AI on the horizon looks more like Amazon Web Services—cheap, reliable, industrial-grade digital smartness running behind everything, and almost invisible except when it blinks off. This common utility will serve you as much IQ as you want but no more than you need. You’ll simply plug into the grid and get AI as if it was electricity.
Not even Kelly foresaw the bitter lesson:
Yet most of the commercial work completed by AI will be done by nonhuman-like programs. The bulk of AI will be special purpose software brains that can, for example, translate any language into any other language, but do little else. Drive a car, but not converse. Or recall every pixel of every video on YouTube, but not anticipate your work routines. In the next 10 years, 99 percent of the artificial intelligence that you will interact with, directly or indirectly, will be nerdly narrow, supersmart specialists.
Very soon now we’ll live in a world where we can ask the cloud, in conversational tones, any question at all. And if that question has a known answer, the machine will explain it to us. Who won the Rookie of the Year Award in 1974? Why is the sky blue? Will the universe keep expanding forever? Over time the cloud, or Cloud, the machine, or AI, will learn to articulate what is known and not known. At first it may need to engage us in a dialog to clarify ambiguities (as we humans do when answering questions), but, unlike us, the answer machine will not hesitate to provide deep, obscure, complex factual knowledge on any subject—if it exists.
…
As we cognify more books and movies and the internet of things, answers become ubiquitous. We are headed to a future where we will ask several hundred questions per day. Most of these questions will concern us and our friends. “Where is Jenny? What time is the next bus? Is this kind of snack good?” The “manufacturing costs” of each answer will be nanocents. Search, as in “give me an answer,” will no longer be considered a first-world luxury. It will become an essential universal commodity.
2026-09-03 08:00:00
Why did you decide to write this essay? What was your intention for the essay when you started, and how did that evolve over time (if it did)? Was anything surprising to you?
How long did it take you to write this? We’re interested in two answers: (1) over what time period you were ‘working’ on the essay—from idea to final draft—and (2) how long it literally took you to write the essay (the time you spent typing/revising/editing).
What was the process you used to write? How many steps and iterations were there? E.g. was it bullet points -> draft1 - draft2 -> final? Was it something else? How similar/different is this to the process you use to write other essays?
Do you have any screenshots of drafts of early snippets of the essay you could share? Stuff that is either not in the final essay or not in there verbatim. (People like that.)
My essay ideas are impolite -- they seize my attention when I must do important things.
My idea backlog is a giant txt file with 6,000+ essay/project ideas
that I will never complete:
Sometime in early 2024, I recorded this note to my idea backlog:
crows vs. kindergartners vs. a math professor, "intelligence as negative entropy", entropy extractionI have zero clue what it means, but I suspect some personal experiences became entangled:
I developed a clever method that prioritizes high-quality viral essays. If I were a disciplined writer, I would use it. But no -- I simply transcribe my conversations with the loudest voices in my head (no, not literal voices -- they're just my model's weights and biases du jour).
On 2024.07.14, the mystery of CROWPOWER yelled at me until I wrote 739 words about it. My first drafts looked like this:
Two weeks later, I tried to give my idea a spine:
And then I let it sit for three months. I'm not sure exactly what happened, but I bookmarked papers like Information Compression as a Unifying Principle in Human Learning, Perception, and Cognition during that time. I probably realized that my entire idea was bogus and/or cliche. Sometimes things need time to breathe.
On 2024.10.13, I took another stab at it:
And then I got stuck again, and it sat for another three months.
After a short trip to Cambodia, I picked the essay back up on 2025.01.05. One big idea made it into my essay notes: A few thousand crows with error-correction equipment could emulate a CPU like a Tinker-Toy computer. This simple idea became the Rosetta Stone that let me translate between biology, computation, and information.
I also wrote down a bunch of ideas that didn't make it into the essay. For example, intelligence seems highly dependent on personal motivation. What if you asked somebody to retake an IQ test and gave them $100,000 for each point increase? I suspect scores would increase on average, but not by much. Doesn't this severely complicate measuring non-human intelligence? And on, and on, and on.
Six months later, on 2025.06.22, I rewrote the intro to focus on the history of horsepower.
The new horsepower intro unblocked the essay -- suddenly everything I'd been failing to say had a thread running through it.
From a clean slate, I spent every minute of my discretionary time writing this damn essay. I worked this way for six days, and then it was finally complete.
This essay took a year of calendar time and ~80 hours at the keyboard. I am a very slow writer, but this one was glacial even for me.
I log all my edits to git, so it was easy to see how it developed:
| date | words | diff |
|---|---|---|
| 2024.07.11 | 12 | braindump mixed with other notes |
| 2024.07.14 | 739 | first real outline; mostly "tk:" stubs |
| 2024.07.29 | 1,705 | expanding 2024 skeleton |
| 2024.10.13 | 1,982 | added "joepower" |
| 2025.01.05 | 2,419 | crowputing and employing crows |
| 2025.06.22 | 3,414 | "…and no horses on the moon" |
| 2025.06.28 | 4,861 | rewrote the whole thing with scaffolding notes |
| 2025.07.01 | 2,714 | removed scaffolding notes |
| 2025.07.01 | 2,783 | final edits; submitted to contest |
| 2025.09.03 | 2,746 | published alongside prize announcement |
How did you decide the essay was ‘good enough’ to publish?
Astute readers understand that none of my essays are ever good enough to publish. But I hit that publish button when my curiosity and pride and shame reach equilibrium.
My hunger for useless knowledge takes me to weird places; my brain refuses to leave the restaurant until it's tasted everything on the menu. "I can improve upon these recipes!" No, no, I can only cook inedible sins in my kitchen. It's torture until I remember that I don't have to be smart if I'm funny -- I suddenly fall madly in love with the enormous mess I've made, and I want everybody to know how wonderful it is to be alive.
What was the reaction when you initially shared it? Was it as you expected?
I was elated when Adam Mastroianni crowned my essay the winner of his 2025 Blog Extravaganza. My friends probably weren't too surprised -- Adam and I share an uncommon reverence for bad jokes and amateur science.
But I'm a bit disappointed that "crowpower" and "joepower" haven't wormed into our lexicon. I half-expected those words to flourish in a certain unnamed AI safety forum (from which I maintain a healthy distance). It's wild how everybody trusts intelligence benchmarks, but nobody knows what we're measuring. I barely have the vocabulary to explain my confusion. It's like trying to describe Bitcoin to an ancient Aztec person.
And now here we are. I am so grateful that Jordan & Carter asked about my CROWPOWER essay. I think I needed some flattery/encouragement. "I can improve upon these recipes!"
Do you think you have any nonobvious or unpopular opinions about writing?
Here's something I grapple with: writing about things I know, versus writing about things I don't know. It's the difference between writing to teach and writing to learn.
For example, this is supposed to be an interview about how I wrote CROWPOWER. Simple questions: When did you write it? How long did it take? What was your process? What do you know?
But all I care to talk about is how opaque my creative process is, especially to myself. Whoops -- here I am making a mess in somebody else's kitchen.
Bonus: What is your favorite essay on the internet?
My favorite essays are those that seem to be about one thing but are actually about something else entirely.
I don't want to spoil the surprise, but this is one of the greatest literary feats I've ever seen. Its genius is completely orthogonal to its content: https://gamefaqs.gamespot.com/snes/588741-super-metroid/faqs/10114
But this is not one of those essays. It is exactly what it is.
2026-09-01 08:00:00
every Wednesday
squirrel away our mayonnaise
full vegetable drawer
2026-08-06 08:00:00
I'm so grateful for my time at OpenAI, where I led alignment safety for the official ChatGPT basketball. I gained so much valuable experience building Altman's armageddon engines, and I'm taking it to the next level with my new startup: Jurassic Park.
At Jurassic Park, we're resurrecting the ultimate apex predators to defend humanity against misaligned AI.
Further reading: OpenAI incident report, Anthropic incident report.
According to the fossil record, dinosaurs lived for 165 million years without a single recorded AI cyberattack. Many speculate that dinosaurs developed a natural immunity to artificial intelligence threats.
Datacenters evolved strong concrete exoskeletons to protect their vulnerable innards from predators. But think about it -- a dinosaur wins against pretty much everything (except maybe a literal meteor). In a recent poll, most scientists agreed that a dinosaur could totally destroy a datacenter if it really wanted to.
Anti-AI dinosaurs have already passed experimental trials in Japan.
In 1974, apes from Black Hole Planet 3 built a space-titanium imitation of Godzilla, which rampaged across Japan. But Mechagodzilla's disguise burned off and Godzilla (with assistance from the Okinawan guardian beast) turned him into a magnet and then tore his head off.
In 1975, the ape aliens built Mechagodzilla 2, who rampaged across Japan with Dr. Mafune's sea beast Titanosaurus. But Dr. Mafune's cyborg daughter discovered she was Mechagodzilla's wireless transmitter, and shot herself to sever the link. Godzilla seized this opportunity to decapitate Mechagodzilla 2 and blast its corpse with his signature heat ray.
Mass is estimated; only Chinese weights are public.
This chart assumes dinosaurs are ectotherms, not endotherms.
Dinosaurs are energy-efficient from cradle to plate! Dinosaur meat is a natural superfood -- rich in protein or omega-3 or something. We expect it to fetch a heavy premium at fancy steak restaurants when it first comes out.
We're committed to 100% circularity. Surplus dinosaur parts will be fully recycled into coal at our proprietary underground facilities.
Everybody loves dinosaurs! Don't you want to cuddle these gentle giants?
Purchase our official life-sized Stegosaurus plushie from selected distributors.
Contrary to popular belief, most dinosaurs pose no copyright infringement risks.
The Land Before Time (1988) was a decent film with superb animation, but the child-friendly sequels really tarnished the brand. We expect a reboot any year now.
Jim Henson's Dinosaurs aired on ABC from 1991 to 1994. Although targeted at a family audience, it somehow covered complex topics like sexual harassment, censorship, religion, steroid abuse, masturbation, and corporate crime. It got surprisingly good reviews; many appreciated its creepy-but-well-executed practical effects.
Top economists expect AI to take ~99% (±5%) of human jobs. We believe Jurassic Park is the ultimate jobs program.
AI cannot raise cattle, nor feed them to raptors with Spielberg's cinematic finesse. It cannot pick up mountains of feces ("poop") without a mechanized corporeal form. Human jobs are safe as long as we have dinosaurs.
Ancient humans harnessed the power of megafauna for megaprojects. According to The History Channel, ancient Egyptians used terrestrial sauropods to build the Great Pyramid of Giza.
AI scientists were so preoccupied with whether they could, they didn't stop to think if they should. Luckily, our dino scientists are doing what they must.
We're so excited to send you back to the stone age.
Disclosure: This post was partially funded by The Reptilian World Government.
2026-07-14 08:00:00
I was born in the USA. Many tongues call me "American": Italian (americano), French (Américain), German (Amerikaner), Dutch (Amerikaan), Afrikaans (Amerikaner), Japanese (アメリカ人, amerika-jin), Filipino (Amerikano), Hebrew (אמריקני), Arabic (أمريكي), Portuguese (americano), Russian (американец), and Hindi (अमरीकी, Amreeki).
I grew up in Southern California. Many of my friends were zeroeth-generation immigrants from Mexico. Weren't their parents born in North America, too?
As an adult, I lived in Washington State. There I met many fine Canadians, visiting for work, education, and play. Were they not Americans too?
We of the United States, in justice to Canadians and Mexicans, have no right to use the title "Americans" when referring to matters pertaining exclusively to ourselves.
-- James Duff Law
Circa 1903, James Duff Law coined the term "Usonian" for the peoples of the United States of America.
Esperanto speakers adopted "usonano" as early as 1905.
But why this term "America" has become representative as the name of these United States at home and abroad is past recall. Samuel Butler fitted us with a good name. He called us Usonians, and our Nation of combined States, Usonia.
– Frank Lloyd Wright
Wright popularized "Usonia" around 1927. To him, the US middle class outclassed European aristocracy. His Usonian homes celebrated the simplicity and individualism of US suburbia.
His Usonian homes inspired ranch-style houses.
But "Usonia" sounds too beautiful for the United States -- collegial, polished, academic. "Would you prefer Italian or Usonian wine with your halibut, ma'am?"
Our heritage is that of pilgrims and prophets and rednecks and reverends. We've got the best medical centers and the most overdoses; the best universities and the most young-Earth creationists.
I feel less than American and more than Usonian. I wish there were a word between…
Yankeeism is the general character of the Union. Yankee manners are as migratory as Yankee men. The latter are found everywhere and the former prevail wherever the latter are found. Although the genuine Yankee belongs to New England, the term "Yankee" is now as appropriate to the natives of the Union at large.
-- Thomas Colley Grattan
"Yankee" is a silly word. It probably originated from Dutch Janke ("Little John") or Flemish Jan Kees ("John Cheese").
During the Revolutionary War, British troops sang "Yankee Doodle" to paint US folk as raucous clowns. And in true Yankee fashion, US troops put it in their caps and called it macaroni. Yankee Doodle remains a source of national pride.
Brits (and Aussies) still use "Yankee" in a pejorative sense.
The US still exudes that panache. In the early 1800s, colonials built a civilization while consuming ~1 gallon of 90-proof liquor per month. Their children's children invented plastic and Hollywood, which pollinated via red cups and reality television. Only the US could elect a reality TV star to be its leader (twice). It takes gall to produce Idiocracy and then make it happen.
We should move the US capitol from DC to Las Vegas.
Only the US could give casinos to its decimated native populations as a consolation prize. Only we could inflict upon ourselves epidemics of painkillers, crack, school shootings, and Tide Pods. Only we could produce evangelicals.
US folk are Yankees. We're obnoxious, tacky, foul, and we're proud of it.
To foreigners, a Yankee is an American.
To Americans, a Yankee is a Northerner.
To Northerners, a Yankee is an Easterner.
To Easterners, a Yankee is a New Englander.
To New Englanders, a Yankee is a Vermonter.
And in Vermont, a Yankee is somebody who eats pie for breakfast.
-- E B White
But Southern Yankees reject the label. Many still cling to the failed Confederate nation of their fathers' fathers. They're right to remain angry -- we've since let our Southern siblings slide ever further into poverty, illiteracy, and hopelessness. To them, Yankeeism is another flavor of European aristocracy.
Southerners are suffering, and it's easier to be dismissive than to be helpful. The Civil War will not end until Southerners are proud to call themselves Yankees too.
the spartan, cramped, and unstintingly functional jeep became the ubiquitous World War II four-wheeled personification of Yankee ingenuity and cocky, can-do determination.
-- Doug Stewart
We should make Yankee synonymous with the best of the United States -- dynamism, ingenuity, hospitality, self-sufficiency.
I like "Usonia", but it ain't US. We're Yankees, for better and worse, and we should act like it.
