A Change in Cadence (And a Few Other Things Worth Saying)

One of my goals for 2025 was to get back to writing consistently. I had a plan, an outline, a sense of purpose—and then, as the productivity data from last year will confirm, I did not do that. The intention was sincere; the follow-through was not.

So I reset the intention for 2026, this time with more structure behind it: a specific goal of original content every week, an actual outline of topics, a rough arc of where the series would go. They were all supposed to be short-form—the kind of thing you read in five minutes and carry into your day. Instead, each post has grown beyond what I intended, and each one has generated two or three ideas for posts that didn’t exist in the original outline. The plan is not getting shorter as I work through it. It is getting longer.

I forgot, until I was doing it again, how much I genuinely love this. Writing has always been something closer to a creative compulsion than a professional obligation for me—not unlike drawing or painting, where simply making the space and time to be creative feels good in a way that is difficult to explain to someone who hasn’t felt it. The problem with rediscovering that you love something is that you stop treating it like a task and start following it where it goes. Where it has been going, consistently, is long. These posts have been running longer than I intended, longer than is probably reasonable to ask of a reader on a weekly cadence, and—if I’m honest—longer than I can sustain producing every seven days without it becoming a grind rather than a pleasure (More than one professor, across college and graduate school, handed back my writing with the same note in the margin: Too terse. Expand. They would, I think, find the current situation quietly vindicating. It helps to be writing about something I love.).

So starting in June, I’m shifting to every two weeks. The longer form deserves a little more room: more time for you to sit with what you’ve read, more time for me to do the work well. I’d rather give you something worth your attention on a longer schedule than something rushed on a shorter one. And I’m having fun!

June also brings a surprise I’ve been sitting on for weeks and genuinely cannot wait to share. Stay tuned.

There’s something else I want to address directly, because I think it deserves a clear explanation rather than a footnote.

I have started using generative AI as part of my writing process, and I’ve been including disclosure statements in posts where I’ve done so. I want to say something about why, because the that without the why can look like a lot of things it isn’t.

Some context is useful here. My mother read. Her parents read. Growing up, our town didn’t have a public library, so she took us every two weeks to the county bookmobile—the maximum number of books allowed, every visit, plus whatever I could pull from the school library. By middle school I was moving through a novel a day on top of schoolwork. I read every Louis L’Amour title the bookmobile carried, at which point the librarian started flagging me when new stock came in so I could branch out and come back when there was something new to find. That two-week rhythm shaped how I think about time and content in ways I’m only now noticing.

I say this not as a credential but as context, because it is the origin of a problem. I now maintain a reference database of over 2,100 journal articles—a sample of those I’ve read, not a complete record—and a personal library of more than 2,000 book titles, physical and digital, to say nothing of the untold thousands read across a lifetime. When I’m writing and I know I’ve encountered exactly the argument I need—a specific study, a quotation I can almost hear, a title sitting just out of reach in memory—finding it can take an afternoon. That is the retrieval problem AI actually solves for me. I can describe what I half-remember and arrive at the source in minutes rather than hours, which means I can redirect my attention toward the parts of this work I find meaningful: the writing itself, the argument, the voice.

My son Porter had a version of this problem in his freshman year of university. He is, predictably, a voracious reader—during his campus tour, he found his way into the library three times before deciding, among several offers, that this was where he would go. In his first semester, he discovered a book he needed in a stack, and on a whim set a goal to read every book in that stack before the year ended. He did. He is also someone who, because he reads extensively and thinks carefully about language, uses words like delve—not because an AI suggested it, but because it is the right word and he has the vocabulary to reach for it. His university’s AI detection software flagged him for it. He was able to produce writing from high school demonstrating he had used the word before he had any reason to know what a large language model was. He now weighs, when he writes, whether he should introduce errors to read as more human.

Let that land for a moment.

I have spent my career studying how context shapes behavior, and this is a clarifying example. We have built detection systems premised on the idea that a certain kind of fluency is suspicious, and then we have aimed them at the students most likely to have developed that fluency honestly. AI detectors are, as the research literature has demonstrated repeatedly, unreliable. Academics deploying them with confidence are not exercising rigor; they are performing it.

I am, I’ll admit, engaged in something like malicious compliance. I fed everything I could find into Claude—blog posts, letters, essays, class papers going back as far as I have copies—and asked it to describe my style, then to identify phrases I use that commonly trigger AI detection flags. Because the stylistic choices that flag me—passive constructions, parallel structure, three-beat arguments, complex subordinate clauses, the em-dash—are mine. They are the sediment of a reading life. Large language models and I were trained on the same corpus. The difference is that I’ve been doing it longer.

There is also something in how I use it that I want to name, because it is the opposite of what people fear. In school, peer review days were among the most useful things that happened in an English class: you swapped drafts with a classmate, they read what you wrote, and they told you what was landing and what wasn’t. Does the argument make sense? Does the flow work? What are you missing? Sometimes you agreed with the feedback and changed something. Sometimes you didn’t, and the disagreement helped you understand what you were actually trying to do. Sometimes a comment sent you somewhere you hadn’t planned to go, and that was the best outcome of all. The fear about AI—and about peer review, for that matter—is that it reinforces your existing tendencies, flattens your thinking, pulls you toward the mean. That’s what mindless use does. Intentional use does the opposite. I use Claude the way I used those classmates: look at what I’m doing, tell me if it’s working, push back if something isn’t landing. I don’t always take the note. But the conversation gets me somewhere I wouldn’t have reached alone.

As a training director and clinical supervisor, I have a professional responsibility to develop genuine competence in tools that are entering clinical and professional settings—not familiarity, competence. I cannot supervise trainees in the thoughtful use of AI if I don’t understand what thoughtful use looks like from the inside. This isn’t a theoretical debate at a comfortable remove, either. In my own department, people I respect and work alongside every day have strong feelings about AI, and they land in very different places. That range of comfort and skepticism spans the full breadth of what “AI” actually encompasses—not just large language models, but machine learning, risk algorithms, the autofocus system in a camera, the spell-checker that has been quietly correcting prose for decades. The anxiety tends to cluster around the most visible and recent tools, but the technology is considerably older and more embedded in daily practice than the current conversation suggests.

Writers have always had tools, and the tools have never determined the work. Some writers organize their research in DEVONthink, manage citations in Bookends or Zotero or Endnote, draft in Scrivener. Scrivener is worth pausing on for a moment: one of its most useful features is the research folder, a space inside the project where a novelist can keep character notes, timelines, and reference material—organized, at hand, informing every page, but never appearing in the manuscript itself. That is a reasonable description of how I use AI. Others work on typewriters, or on vintage machines running WordStar, because that environment is where their thinking moves best. All of them are writers. The question has never been which instrument sat between the idea and the page.

The citation workflow is a useful concrete example. This blog uses ZotPress shortcodes to generate in-text citations and bibliographies; my writing process uses Zotero field codes. Converting between them manually is the kind of task that is tedious without being difficult—which is precisely the kind of task where errors accumulate quietly and reliably. AI handles the conversion accurately, which is not categorically different from what reference managers have always done: Zotero, Bookends, Mendeley, Endnote all exist because hand-typed in-text citations and bibliographies were slower and more error-prone than letting software do it. The tool changes; the principle doesn’t. And practically speaking, removing that conversion bottleneck is part of what has allowed me to maintain a weekly posting cadence—and what I expect will help me maintain a biweekly one going forward.

The Museum of Glass in Tacoma sits at the end of the Chihuly Bridge of Glass, and I find myself thinking about its artists when I think about this question. Lino Tagliapietra, widely regarded as the greatest living maestro of Venetian glassblowing, began his career doing everything himself—gathering, shaping, finishing. As he became a master, the work changed. He began to guide teams, to call out adjustments, to shape outcomes through vision and judgment rather than through every physical gesture. The glass that came out of those collaborations was no less his.

Preston Singletary is, without qualification, my favorite working glass artist. His Raven and the Box of Daylight—the Tlingit creation narrative rendered in blown and cast glass—is the kind of work that stops you where you stand. It is simultaneously ancient and immediate, decorative and serious, technically extraordinary and emotionally direct in a way that technical skill alone cannot produce. Singletary has faced a version of the tools question from critics who have argued that his work isn’t truly Tlingit because he works in glass rather than in wood or other traditional materials. His response has been direct: he is transforming the culture and forging new paths, and that should be allowed. The medium is not the ancestry. The vision is. Singletary, incidentally, learned some of what he knows about glass from Tagliapietra.

Greg Owen, who managed the Hot Shop Heroes program at MOG—teaching glassblowing to soldiers and veterans—and who built his career working alongside Chihuly and at Pilchuck Glass School before his death in 2020, was early in that trajectory when I first encountered his work. No one watching him at the bench would have asked whether his tools made him less of an artist.

The question, in every case, is what the artist is doing with what they have. That’s the question I ask myself.

I started writing this as a Facebook post.

I’ll wait.

The irony is not lost on me, and I’m choosing to let it stand rather than edit it away—because it is, honestly, a more accurate illustration of the whole problem than anything I could have constructed deliberately. If you need me, I’ll be over here, not writing short things.

See you in two weeks (-ish).

This post was written by the author. Claude (Anthropic) assisted with drafting, organization, and editorial refinement throughout. All views, opinions, and personal reflections are the author’s own.