Schreibvergnügen is the joy of writing.

Despite all my critiques, the truth is that I am writing a lot more in the AI era and loving it.

I’m writing more than ever even though I keep hearing that writing is dead and that I should be pivoting to video (again).

Historically, someone, somewhere is always saying that writing is dead and that the youths don’t read anymore, i.e. here’s an example that Paul Fairie pulled that dates to 1907.

The current era of fear mongering includes yet another pivot to video and now, also, a huge pile of “AI will replace you” paired with a “because AI stole your writing” poke in the eye.

Given all that, I want to note that I’m writing more than ever and loving it to the point that I had a shirt made, Schreibvergnügen. This is a manufactured German word that means joy of writing, like Fahrvergnügen was for the joy of driving.

There are so many reasons to love writing, and you’re all welcome to your own. These are just the reasons I’m experiencing right now. What’s notable is that they are all related to this new AI era. I’ll say a bit at the end about how I’m navigating my own lengthy critiques of AI companies. But there is no going back to pre-AI, and so I’m finding the positives.

I’m back on morning pages.

I’ve been typing morning pages, targeting 750 words, and that takes me about 17 minutes. It turns out that a lot of what is in my head in the morning is anxious thoughts about all the things I absolutely have to get done that day. These are essentially todos. So I added an AI prompt that turns my morning pages into my daily todo list.

Make a task list of any things that I said I needed or wanted to do in the current note plus any unfinished tasks from yesterday's journal. Format it as a task list. If possible, write the tasks in terms of first step followed by the goal. It's ok to guess at the first step.

Morning pages are as great as everyone says. But as with all self-care, it’s easy to talk yourself out of doing them if you can’t justify them as useful. This little bit of AI assistance tips my morning pages into being a utility on top of already being an excellent mindfulness practice. That means I do them more often.

An aside: the official morning pages practice is meant to be free form and without goals. But I’ve never been pedantic about self improvement practices and I don’t think you should be either. I’m making this aside for an AI reason and it’s that before posting this post, I asked Claude to read it and tell me how a cynical reader might react (below). Was Claude right?

‘“Morning pages as productivity optimization is missing the point.” Julia Cameron devotees will bristle hard at turning morning pages into a todo list generator. The whole premise of morning pages is that they’re not useful that’s the point. Framing AI assistance as the thing that finally makes morning pages worth doing will land as exactly the kind of productivity-brain thinking that critics argue accelerates.’

I wrote a user guide for my coworkers, then they all wrote user guides too.

There’s an emerging usage that I call hybrid writing, where I am writing something for other humans, but knowing that it’ll also be used by AI. That writing does double duty, which means I’ve got even more reasons to take the time to write it.

In this example, a user guide (sometimes called a README) is a “here’s how to work with me document.”

They’ve, rightfully, gotten a bad rap because they push the work of adapting onto the people around you. “I’ll be less difficult to work with if you just read and apply this manual about my peculiarities.”

And yet, we’re all different, unique, difficult-in-our-own-way people. So a user guide can be a roadmap for when a motivated coworker really needs to make a point that you can hear. I appreciate having them for other people.

FWIW, Camille Fournier has more depth about user-guides/READMEs.

I’ve read and can remember most of the user guides of all of my immediate coworkers. But then I set my AI chat (Claude) to pull them in whenever I’m preparing for a 1:1 conversation. It’s a magic trick for getting to shared understanding before a meeting starts.

What makes these hybrid audiences is that they were written for humans and read by humans first, but then they are also in regular use within AI Chatbots.

This extra utility seemed to have tipped people into writing them and taking them more seriously. That’s certainly true for me.

I started writing a book.

I have the material already, having written hundreds of blog posts based on research I was doing at my last job. But what I didn’t have was a way to organize all that material.

A friend told me his theory that programmers are a year ahead of writers in terms of knowing how to use AI tools. I’m surrounded by programmers though, and they are going nuts about how productive AI is making them. (although, honestly, still with some skepticism about).

But I think I’ve soaked up one of the core patterns programmers use when they build with AI. They work in stages: plan then generate then edit.

So I applied that to turning my blog posts into a book and am about 25% of the way through writing the first draft of that book. I’m still as pedantic as I’ve always been about serious writing: I don’t want to publish any advice that I haven’t tested on myself and other people, and read and revised multiple times. So I don’t think this book draft is AI generated slop—rather it’s AI assisted organization. The planning stage blew my mind, helping me go from goal to title to outline to chapter structure in less than an hour.

Ok, what about the downside? My critiques of AI are probably pretty similar to your critiques. I think the AI companies stole my writing (wrote about that here), hate how AI generated slop has me questioning whether what I read is real, don’t want my ideas or my wordsmithing to be replaced, and I worry about the climate impact. Here’s how I’m navigating this era.

• A key life principle is to leave the world better than I found it. So I think a lot of the AI slop/hallucination ends up being a choice and I know I have the choice to only publish things that I’ve written, edited, tested, and believe in.
• As for the climate side of leaving the world better, I do know that AI usage has a higher per dollar climate impact than other cloud computing services. But I also have a lot of say at work about reducing our overall cloud costs. So we are trying to reduce overall computing costs to balance out increased AI usage.
• I’m more optimistic that AI companies will pay a price. I don’t entirely believe OpenAI is going to go out of business. But I do think AI companies will be forced to pay. Anthropic paid book authors (even for a book I wrote in 2006). And I also got to deliver a message to one of the search companies that if they don’t come up with an exchange of value then all of the information that they currently search is going to move into private spaces. I think they got the message.

Medium, where I work, has a lot to say about these topics too: Principles for using AI, managing our carbon footprint, an internet standard to make AI companies pay.

Taking this back to joy. Sure, I’ve had the experience of angsty writing. That’s usually when I’m writing under pressure.

But when I’m not under pressure, I experience writing as pure joy. It’s just thinking that has a dream behind it. The dream is that maybe I’ll have something to say that other people will like, or better yet, that will make their life better.

For me, writing is thinking with the dream of being useful.

So I’m looking at the above and seeing all the ways I’ve found to use AI to draw out more of my writing. It’s not a little bit — I’m writing a lot more. As a result, I am experiencing a lot more joy in my life.