How This Gets Written

I love to write. I always have.

I wrote for my high school paper, The Rampage. I wrote for my college paper, The Heights. Since then it’s been scores of legal briefs, law review articles, op-eds, motions, memos, and speeches, plus a steady output of oddball emails to family and friends that may be my favorite form of all. I love the process. I love the thinking that writing forces on you, the way an argument you were certain of comes apart in the third paragraph and you have to go find out what you actually believe.

Artificial intelligence is one of the hardest questions of this era, and perhaps of any era. I don’t say that casually, and I don’t claim to have it worked out. What I can tell you is how I handle it here. The internet remade research and commerce and a dozen other things it was never built for. AI is doing that to writing.

I have no interest in producing AI slop. Zero. Nada. None. Maybe there is a place for it. Maybe there is not. I don’t know. I do know that it wouldn’t be worth your inbox or my time.

My first editors were, in order, Mom, Mrs. Mattix, and Mrs. Dundas. Pushovers? Wallflowers? Not so much. Four decades later I’m still grateful to all three.

Since then I’ve written for outstanding editors of every kind: senior attorneys in practice, professors, judges, and professional editors at The New York Times, Slate, The Wall Street Journal, Newsweek, Verdict, The National Law Journal, and plenty of others. A professional editor is a privilege and a luxury. The good ones tell you plainly what’s wrong with the first draft; and what still needs work in the fifth draft. They don’t soften it to spare your feelings. I try to do the same to my own work. I’m a tough editor of myself, but I also know the ceiling on that, because nobody catches themselves in their own blind spot.

I’m also one person. I teach law full time. I am devoted to my work with ABC News, and I also care deeply about my work for MS Now, CNN, and others. I still publish op-eds, edit briefs, and publish in academic journals, and I write and deliver remarks in fora ranging from the halls of government to public events.

This newsletter is free and there is no one behind it but me. It exists to reach people and to raise the level of the discourse rather than to sell anything. Every hour it takes has to come from one of those other places, and nobody is assigned to edit it.

So I use AI, on a limited basis, as an editor. Without it there would be fewer pieces here, and the ones that ran would get a thinner edit. The phrase “as an editor” can mean a lot of different things to different people, so here is my process.

I write the text. I ask for an assessment of its strengths and weaknesses. I ask what needs improvement. I identify phrasings and points that I’ve made that don’t quite seem to hit their intended marks, and ask for suggested improvements. I don’t know the exact percentages, but I’d guess that I reject 90ish percent of those suggestions, but they help me refine my thinking. I also use AI to produce the title banners, flow charts, and graphics I could never make on my own - or at least not on a remotely timely basis.

The limits are strict. I’m highly restrictive about the text I put in front of it. I set rules against going outside my sources and my words. When I’m workshopping, I narrow the scope of each question deliberately, and I make every final call.

The product is mine. The judgment is mine. The errors, when they occur, are mine too.

I cherish good writing and believe in its power. More than that, I believe the discourse in our democracy is sorely lacking in fidelity to facts, in rigor of analysis, and in substance. That’s why this newsletter exists. You are why the hours are worth it. If limited AI assistance can help better serve those ends, I’m willing to give it a try.

And suffice it to say, I’m open to feedback and suggestions.