Good Morning. Hello. How are you? #1524
Olivia and Bob, Billy Eichner's Parks & Rec Mantra, starting writing my AI book and the different feelings I have since I last started a book.
Good morning, good morning, welcome to my chronicles of insomnia. Eh, actually, last night wasn’t too bad. Took me an hour and a half to fall asleep. That’s normal for a lot of people. I suppose this is a good sort of bonding experience with my wife. Walk a mile in their shoes sort of thing. So long as it’s only a mile. So long as it goes away eventually.
Man. For like six months I was complaining about having to wake up at 6:20 in the morning, and I was all excited for summer. Now summer is here, and I am complaining. I should probably just get over it.
Of course this would not be a problem if I wrote you later in the day but, you know. You’re the first thing I do, friends. The first thing I do. My sleep is still on my mind.
We are listening to a band called Pennikurvers this morning, an EP called Dying to Tell. My friend Bill alerted me to them. They are old school. Broke up in 1996. But they are getting some new love these days, including the relase of this EP by the fantastic label Numero Group. Bill said it sounded like “a lo fi Velocity Girl” and he is exactly right. I am enjoying it.
Oh cool. Just got my first recruiting text spam, but not to my phone number. Straight to my iCloud account. Soon there will be recruiting text spam stuck on my monitor with a Post-It when I get to the basement office in the morning.
We are watching season 7 of Parks and Recreation and Billy Eichner is on it the last two seasons and he’s great and there’s this subplot where whenever he gets upset he needs to list three good things about the world and my god don’t we all need that today, these days. Fuckin Lisa Murkowski. Fuckin Senate! The bill got worse! The Senate’s supposed to moderate things!
Italian Ice, Total Boat, Cassettes.
Join the GMHHAY slack! Reply to this email and ask for an invite if you’re a human who likes chatting with other humans about topics such as these within!
Yesterday I finally got off my ass and started writing this AI book. I had a 3,000 word outline already, composed in the Apple Notes.app from all of my various brainstorms and thought flashes for the last, eh, month or two. In a few of the recent bouts of insomnia, I had re-organized the outline into something pretty coherent.
This was handy because the first steps in “starting” this new book had already been taken, and the first steps are, of course, the hardest. That and working a bit every day. Even an hour a day and you are good to go. But you gotta do it every day. Well, you can take the weekends off. I give you permission to take the weekends off.
I started by taking the outline and migrating it over to Scrivener, my book-writing software of choice. I don’t know what I’d do without Scrivener, it rules. In fairness, MS Word has gotten better about multi-document, long-form writing since I first started to use it to try and write a book in 1994 or so. But I am committed to Scrivener. For one reason, some of my books will take five, ten, fifteen years to write. I have been working on this ad book for well over a decade now, and I am writing this book to my daughter that will, by the very nature of the project, take eighteen years.
I have written completed and published four books. So I am not new to this. But the last one, the first volume of Good Morning. Hello. How Are You? (volume 2 coming soon!) isn’t really a book. Or, rather, it wasn’t written like a book. The last actual book I wrote was, gawd, eight years ago now. Having a kid, you know? Cuts down on your extracurricular hobbies for a while.
And, reader. I am out of practice.
On the one hand, it is going just great. I have been tossing and turning and stirring and insomniaing for weeks now, thinking about AI and society. I’ve been forced to think about it for years now. I’ve had conversations, fleshing out my opinions, with dozens of people. So the actual message of the book is pretty easy. I just started writing at the beginning, and away I went. Five thousand words the first day. Bob’s your uncle.
So, too, is the gardening of the book: the editing, the copying and pasting and moving of paragraphs and sections to sharpen the message. The re-reads and tweaks and massaging of the text. Love that shit. Deeply relaxing and satisfying. I had missed it, since it’s a skill I do not much employ on GMHHAY. Well. Occasionally I move whole topics around, but.. you know. Light touch. But even when you go deep on the gardening, I find it satisfying. It is joyous to me.
But the… hrm… well… what shall I call it? The zoom level, the magnification, the reading level of the book. That is proving exceptionally hard for me this time around.
I want this book to be a polemic. A tight, emotional burst of passion decrying against something. Polemics do not educate, per se. They rant.
But I also want it to be universally accessible. Which a polemic, even about something as technically complex as AI, can do. In theory.
But I am also accutely aware of how much misunderstanding there is about the technology behind AI amongst “mainstream” America. Many people, most people, are unaware of how, exactly an LLM works, how they are not actually thinking, how fallible they are. And while “AIs make mistakes” is not central to my polemic, it’s in there, as the Prego guy says.
So that means I probably gotta explain exactly LLMs work. Same with the cloud. The cloud is definitely a part of my argument, and as much as I just want to rant about cloud architecture, water and power, I know I need to explain these things.
This didn’t use to be a problem for me! I was very happy to patiently explain the fundamental concepts behind my arguments in Agency and the Trek book.
But for some reason, this time, it doesn’t feel quite right.
And I am starting to realize that the whole thing doesn’t quite feel right because of LLMs and AI. The whole feelings behind writing a book have changed. For two reasons:
First, I feel like I am writing this thing to feed a fucking machine, and I hate that. It does all sorts of weird things to me: will footnotes make this more annoying to an AI or no footnotes? Will a tight text be more digestible than a discursive one with lots of asides and parentheticals? He says, hoping the answer is yes, because he is lousy with paentheticals since all he writes anymore is GMHHAY. I find myself thinking about the AI machine reader not quite as much as the human reader, but quite a lot. I blame Robin Sloan.
Second, I am writing into an environment fueled by misinformation, including AI-driven misinformation, and I just… don’t want to deal. Like take electricity. AIs use a lot of electricity. Just tons. And it is getting worse. But most people, even good people, even smart people, don’t get this. They have only done some cursory research, they view AIs as a marginal extension to the pre-existing cloud infrastructure of the world. But it is becoming harder and harder to actually educate someone about something in this world. The normal approach — here is a summary of the facts, here are some sources for you to read to learn more — is only sort-of effective when there are so many incorrect sources out there, willfully incorrect. I don’t want to do it! I want to have a nice adult conversation between people who are up to speed on the topic who are debating the meaning of agreed-upon facts. To write a book that spends gobs of time educating people on those facts is not the goal here.
I found myself in a similar situation with my unpublished magnum opus about advertising economics. And at the time, the misinformation ecosystem was radically different, and the topic itself was far less contentious of a battleground for the hearts and minds of humanity than AI is. There wasn’t a lot of FUD. But there was still an educational need.
And so I ended up writing, gawd, I think that book is, like 500 pages now, maybe 600 pages, when only maybe 200-300 of it is the actual book I wanted to write. Because of all the basic education I needed to do, for only a subset audience. But when you want a book to have a wide audience, it’s necessary. But it’s so tedious.
And, arguably, it’s half the reason I never finished that book. I don’t want to make that mistake again.
I want this one to be light and fast, like Trek or Man Nup.
I guess Agency struck a decent balance. I mean, I will get there. And editors will help: I am hoping to use Lisa as my editor again, and she will be great for telling me when things make no sense to the layperson.
One other thing that I am realizing is that I am writing this book from a “post-classical” point of view. By that I mean, for example, reply-guy trolls love to use phrases like “ad hominem,” and think about Rome all the time. They just love it. But the very concept of ad hominem is basically trash in modern society and a) no one cares about old Greeks who first thought of something and you don’t need to mention them, just the ideas, which are timeless and b) no one really believes in the concept of ad hominem anymore anyway, even when they say the do. So you might as well accept this reality. But a long section about me me me is boring AF and detracts from a solid polemic. So I am relegating it to a different section.
Or, for example, when I get to talking about humanity, it’s not gonna be a litany of Wittgenstein and Hume and shit. Not even good old Swedenborg. No one needs that anymore. It’s just gonna be the ideas. If anything, the references will be far more pop culture: The Good Place instead of Kant. A plea for universality and accessibility that will alienate me from the intelligentsia and kill all my chances for a pulitzer, for which I would otherwise, of course, be a shoo-in, amirite?
But anyhow. These things will work themselves out. And I have you guys to pour my asides into, hopefully keeping the text lean. Wanna keep this thing like 100, 150 pages. I want it to be a quick read. I want it to be read far and wide. Even if I don’t plan on, you know, getting a publisher or doing any promotion.
Jane was lovely yesterday, real return to form. We went to the newish sit-down mexican restaurant for the fourth time and she is still absolutely enamored with the free tortilla chips and Pace Picante-style free salsa. Just thinks it is the greatest thing in the world, and girl doesn’t even pay for her food. It is a joy. Also the place has these giant chairs outside and she just loves them. Loves them as much as she loved the wavy blow—up men at the Car Wash in Fairbanks.
Free Salsa. Wavy blow-up men. Big novelty chairs.
Several hours after I wrote to you, the BBC Music channel, official youtube partner of Glastonbury, did, in fact, post the professionally shot video of Olivia Rodrigo and Robert Smith doing “Just Like Heaven” and it is pretty f’n great, just gloriously absurd. It is kind of hilarious how weirdly uncomfortable Robert looks up there, even though the dude is perfectly used to playing to giant headline crowds.
At the risk of being like those “release-the-Snyder-cut” choads, I now call on BBC Music to release the professionally shot video of the other song Bob and Olivia did, “Friday I’m in Love,” even though in my goth heart it ranks as one of the worst Cure songs ever written. I am getting older and mellower and coming to accept that maybe Robert’s pop predilections are actually pretty good. But I can only say this because we are currently wrapped in the loving goth cocoon of his current Songs of a Lost World.
We were talking about Cure covers specifically and covers in general over in the GMHHAY Slack and Jes alerted me to this glorious, hilarious, but insanely ear-wormy cover of Just Like Heaven by these synth doomers. So I am passing it on to you. It is like a virus. We’re all gonna ~~die~~ deadsy.
Have a lovely day talk to you tomorrow.
—
Thanks for reading.
And hey! Maybe buy one of my books!
Good Morning, Hello, How Are You vol 1.