AI and Large Language Models

It’s easy to feel ick when considering the absolute nightmare that can befall us when we let AI and LLMs think and do for us. They are plagiarism machines. But consider this: they are also tools that can help inform your decisions, instead of making your decisions for you. I recommend using LLMs as a sort of “living thesaurus” you speak with, something that can give you a different perspective regarding your work.

If you subscribe to the idea that we don’t actually create our writing, but that we dip our cups into the primordial soup of the universe and pour it onto the page, adding seasonings here and there… well, then this is just more of that! These technological ethical considerations toward writing have been heated ever since Copy and Paste, and possibly even Gutenberg’s printing press. Will you surf the wave? That’s up to you.

I have had an incredible experience with ChatGPT and Google Gemini and NotebookLM by starting a prompt with this (feel free to edit based on your own experiences):

Then I paste in my poem, and ask it to help me work with me stanza by stanza. This is similar to what I've done in the past with Writing groups.

Not everything the system suggests is good. In fact, I'd say it has about a 33%-50% success rate at good suggestions, but what it does is help you look at your poem from another perspective, and to reduce your personal bias towards the poem.

As you adjust and rewrite your poem along with it, feel free to copy back in your edited work, or to ask questions. The best part about LLMs is that it is following along with the entire conversation, not just the latest thing you typed.

At one point, I've said things like

or

and see how it adjusts to work with me.