Crossposted from my newsletter.

As one of my sabbatical projects, I’m building a chicken coop. We have long flirted with the idea of getting chickens, but one thing or another stopped us: too many little kids, a lack of space to keep them, or the startup costs. But recently my father-in-law scavenged a bunch of lumber from a building site, and suddenly one of my reasons to hold off was gone.

I do a lot of DIY projects. I have built a few pieces of outdoor furniture; I can do small plumbing and electrical jobs; I have spent hours upon hours refinishing and repairing decks; and I re-roofed our house this past summer. When my father-in-law offered me the lumber, I didn’t question whether I would be able to build my own chicken coop—I just started planning the project.

Truth is, though, I am not an especially skilled DIYer. I lack the patience and precision required to master complex tasks requiring a high degree of polish. Fine furniture and major construction are beyond me without the help of someone more skilled to correct my mistakes. I have already accepted that my chicken coop will look like someone’s shack in the woods, and not just because of the reclaimed materials.

My weakness in DIY tasks isn’t really one of knowledge or ability but of patience and attention. My attention flags as I have to sand a board for the third or fourth time to get it really smooth, or I get in a hurry as I snatch fifteen minutes to work on a project before dinner. It’s the same failing I demonstrate in other areas of life, intellectual or relational. I am a hasty person. I would be a better person in all respects if I would more frequently slow down and pay attention to the task in front of me without rushing on to the next thing.

I have been thinking about my limitations as a builder since, at the prompting of my friend Matthew Lee Anderson, I read a discussion of so-called “noble uses of AI” by the philosopher Kevin Vallier. It’s one of the more nuanced and reflective “pro-AI” essays I have read, and so since I have spent a lot of words publicly burying large language models, I wanted to think through the line of argument.

Vallier’s analysis turns especially on a distinction between kinds of intellectual tasks. He asserts that AI

can handle lower-order cognitive tasks, freeing us for activities that engage our highest capacities. A researcher who once spent hours hunting for sources can now spend those hours thinking about what the sources mean. A writer who labored over formatting can focus on whether the argument is true.

Of course, people may not do this. We all know that AI can seduce us into passivity. But I would argue that it also elevates the opportunities for virtue. The mind freed from drudgery really can rise. It can shift us from the mechanical to the meaningful.

Vallier’s strength is that he acknowledges that “cognitive offloading” like that offered by AI tools does indeed cause intellectual skills to atrophy. However, he argues that we make such tradeoffs in other realms and that such offloading, advisedly done, may be worthwhile. He offers the example of a patient who uses AI to find medical research to inform her conversations with her doctor:

She did not hunt for the research herself (there’s no free cognitive lunch), but she did something harder: evaluate it. Her speed at reading dense medical literature may atrophy. But her ability to weigh trade-offs, to question authority, and to integrate information grows stronger.

I’m not convinced, however, that the distinction Vallier wants to make here can be sustained. It would be lovely to think that there’s some kind of “higher” “deliberative capacity” that involves elevated thought and no drudgery, in the same way that it would be nice to think that if I had the right tools, I could practice pure craftsmanship without the “drudgery” of sanding and staining and making precise cuts. But in fact the real craftsmanship is the drudgery, and the real thinking is the attention to detail that’s required to wade through complex texts and think carefully and precisely about grammar and formatting. The great thinkers, like the great craftsmen, are not those who rise above drudgery but those who embrace it and therefore transform it.

In what sense can we say that Vallier’s hypothetical medical patient evaluated her medical research because she read AI-generated summaries of it? Did her “ability to weigh trade-offs, to question authority, and to integrate information” improve because she skipped the step of reading the research materials herself? Granted, she may have reviewed more information, faster than had she read the original studies. Maybe she didn’t have time or expertise to read those studies—fair enough. But I find it hard to believe that Vallier seriously believes she’s thinking on a “higher” level than person who has actually read the research herself.

Vallier advocates that we should offload “mechanical cognition: the tedious, repetitive operations that don’t require judgment.” But human beings, because we are not machines, do not perform cognitive tasks mechanically. Even while conducting seeming drudgery, we are learning and exercising complex acts of judgment. Consider transcription, a task well on its way to full-blown automation. My wife had occasion to transcribe a talk for her personal use recently. She reports to me that through the so-called drudgery of transcribing, she listened to the material in a different way and learned it more intimately; moreover, she exercised judgment throughout the seemingly mechanical act of transcription, sometimes correcting the speakers’ expressions to make them comprehensible in text. If transcription isn’t a purely “mechanical” cognitive task, then, it’s unclear to me that any such thing exists at all. (Consider also Ian Leslie’s post about putting together scissors.)

Good thinking, it turns out, isn’t a matter of getting “higher,” to some realm where we are spared, like gods, the need to attend carefully to humble, repetitive tasks. Rather, good thinking is a matter of getting lower, following the nuances of an argument or sifting through page after page of tedious research. The most skilled thinkers aren’t those who dispense with the need for this drudgery, but those who have the strength to go on attending to the tedious aspects of thinking.

I want to note also that Vallier’s essay was published by an organization advertising itself as “the academy for philosopher-builders,” with “builders” defined, of course, as technologists and builders of businesses. I confess I suspect that those whose definition of “building” doesn’t involve the use of the hands may not know a lot about what’s involved in so-called “mechanical” tasks.

Not long ago, I asked a friend who’s an expert builder for advice on how to level a bathroom door that had stopped latching. He described a process of tiny adjustments: drive a long screw in the top, check the level, shim up the bottom hinge a little, check again, scrape away at the trim a bit, check again. I was just thinking how tiresome that all sounded when he said: “It’s kind of fun to do that kind of little tinkering job.” For all my advanced degrees, at that moment I felt I was the one who possesses the lesser intelligence.