still love books

I love books

I’m learning about the history of the tech we call “book” and realizing that large language models may just be the next iteration of The Book.

Imagine a time before the printing press when only a wealthy few could even afford to have books. And then their access and selection was extremely limited. Much of the knowledge we take for granted today is built upon the very existence of easy, plentiful access to books. So the printing press basically made books a thing. They became accessible to the masses and a feedback mechanism arose whereby the masses wanted more books, on more subjects, in more languages, in more formats. Market demands and profit incentives caused writers to seek out new things to write about and turned non-writers into writers as in “I went to Antarctica and wrote a book about it” and so on.

So how does this tie into large language models? LLMs are built on text. As much text as they can ingest. So basically, an LLM is the next big step from the internet and search engines, which are a next big step from books, which are a next big step from pre-printing press books…and so on.

I’m an artist. I’m a writer. I’m a creative person. The LLM is an existential threat to me and my kind. Or is it?

What will humans be like in 100 years and how will they relate to these LLMs and whatever comes after them? Will we adjust in unanticipated ways just as we did with the printing press? Perhaps book lovers will ask more of our writers and authors. A reader might say, “I could’ve done that with ChatGPT! What else you got?” And there are writers who will rise to this splendidly. I know some of them personally.

The LLM has already stolen all that has been written, stashed deep within its beastly belly. Sort of like the internet…and the library…and the bookstore. And now it’s beyond easy for the non-writer, the non-artist, the non-creative, to have their AI generate something bespoke and creative in their name, standing on the backs of all the text stolen from Shakespeare et al. The creatives will just go and reinvent themselves, again, like they always have, day after day.

I love readers too. I’m a reader.

When I realize I’m reading something from an LLM I’m pulled out of the great headspace that reading creates. There may be times I’m reading LLM-generated text and don’t realize it. If I kept reading, then I stayed in that headspace. It’s very difficult to read stuff I don’t enjoy in some way.

There are times I’ve known I was reading LLM-generated text, either because I was told so or the clues were obvious. And I’ve enjoyed that too apparently, if I kept reading.

I see both wonder and despair coming our way. My glass is half full.