Derek Thompson:

“These language models enable the automation of certain tasks that we’ve historically considered part of the creative process,” Olson told me. I couldn’t help but agree. Writing is less than half of my job; most of my work is reading and deciding what’s important enough for me to put in a paragraph. If I could train an AI to read as I do, and to determine significance as I do, I’d be essentially building a second mind for myself.

I hope I’m not the only one, but just for the record: I like researching, and I like writing. I like the hard work of making my prose more clear and vivid. I like overcoming my ignorance. I like synthesizing the disparate things I read and then trying to present that synthesis to my readers. I like it all.

UPDATE: As I was walking this morning I suddenly understood the most fundamental thing that’s wrong with the way Smith and Thompson think about these matters: Smith assumes that at the outset of a writing project he already knows what he wants to say and just has to get it said; Thompson assumes at the outset of a writing project that he understands what he needs to know and just has to find a way to know it. But for me writing isn’t anything like that. For me writing is discovery, discovering what I need to say — which often is something I had no intention of saying when I set out. And some of the most important research I have ever done has been serendipitous: I have been looking for one thing and instead (or in addition) found something quite different, something I didn’t know I needed but, it turns out, is essential to me.