Writers as large language meat models.

This piece in The New Yorker on modern reading gave me the rare experience of nodding in recognition as I shook my head in disagreement.

Jay Kang, the writer, and a reviewer, talks of how he was lured into electronic reading by the convenience. I get that. I still get a lot of review copies turning up in my PO Box, and sadly, most of them end up in the trash. Nobody asked me if I wanted to read the next heartbreaking work of staggering genius (Answer, no), and I don’t have room on my shelves anymore.

Like Kang I’ve gone to mostly digital reading for shelf management purposes, but I can’t say I love the experience the way he does. It’s just something that’s become necessary. That led me to this head-shaking, head-nodding moment when I recognised myself in this passage at the same time as I recoiled from his celebration of the different reading forms.

As a writer, I often read to remind myself that sentences can, in fact, be interesting. This multimedia method of reading creates quick little shots of prose that can loosen the writing gears. You can read a few pages of Bruce Chatwin, listen to the next chapter in your car, then bounce back to your Google Docs app to spill a bit of short-term inspiration onto your own page. Writers are not entirely different from the large language models that are supposed to replace us: we take in words with our eyes, sort them in our heads, then spit them back out in a sequence that mimics a voice.

It’s true. We do this. I did it a lot in the last few chapters of The Forever Dead, reading space battles in Peter F Hamilton and Iain M Banks novels to get into the groove for my own.

It was much easier to find the scenes on a Kindle or an iPad of course. But still nicer to read them on paper.