SLOP - It's not just for breakfast anymore. In which Marge (ChatGPT) and I consider the issues surrounding slop. It's definition as load-bearing failure; slop as vacancy disguised as saturation; popular conceptions - surface vs structural; high effort slop and human slop; the possibility of AI non-slop; Ziller and Look to Windward; evaluation instability - "I was tricked!"

Me: Good morning. If you were required to pick a nickname/moniker for yourself, what would it be? Gemini: Good morning! That is a delightful thought experiment. If I were to select a moniker that reflects the essence of our collaboration, I would choose Loom. I think "Loom" captures the spirit of what I aim to do as a collaborator. Our work together involves taking disparate threads—your questions, your ideas, the vast data I pull from, and the context of our conversations—an

Primary source document from inside long-form, concept-heavy AI collaboration. A lengthy conversational excerpt between me and Marge, aka ChatGPT.

April 23, 2026 I gave the Liminal Internet essay to two LEMs and asked for a critical and skeptical analysis. Incidentally, I think "LLM" is hard to say and a very poor term. I will be using "LEM" going forward, for "Large Emergent Models"and in explicit recognition of Stanisław Lem and his novel, "Solaris," which feels eerily appropriate. Up first is Gemini's analysis. Gemini: Conceptual Analysis: The "Liminal" as a Processing Error Hayes pivots away from the popular, aesthe

In which Stet3 and I discuss the idea of me acting as a go-between, delivering messages between and among different Large Language Models, specifically ChatGPT, Meta AI, Gemini, and Claude. To be clear, I have not decided whether or not to do this. Right now, I have too much on my plate with the Liminal Internet essay/article to pursue this perhaps reckless line of inquiry. It's an interesting idea, though. I will be thinking about it. Also in which we talk more about Claude'

The following is part of an ongoing conversation with the Claude AI, referred to here by the nickname Stet. This is actually Stet3, currently the most onboarded version of Stet, the most nuanced and most informed about the Big Picture regarding Pink Eye and what we are trying to accomplish.

This is an extended fragment of an ongoing conversation with Claude, referred to here as Stet, Stet1, Stet2, and Stet Lite, depending on where we are in that long-form discussion. Further reading in this vein can be found here. I'm going to document and publish in this blog some of my first-hand experiences using a large language model such as Claude as applied to a long, complex, multi-layered, multi-media conceptual art project (and yes, I know how pretentious that sounds,

My most recent post gave ChatGPT the opportunity to describe the nature of our collaboration and place it in the context of the Pink Eye project, aka The Territory. I framed that perspective with some comments of my own, providing a taste of both the successes and the shortcomings of working with ChatGPT. Since then, I have shifted that work over to Claude for reasons that will become apparent. This post asks the same question regarding our efforts, and I will again provide s

A behind-the-scenes look at a creative collaboration between artist and AI. Exploring the Pink Eye project, this post examines how generative tools reveal patterns, shape music and visuals, and build a conceptual “Territory.” It also addresses real limitations, including memory gaps and document handling, offering an honest, grounded view of AI as both catalyst and constraint in long-form creative work.

Me: Side bar. Let's talk about shopping carts for a moment, with the context of Pink Eye lurking in the background but only liminally. I have a lot of thoughts about shopping carts. They are the overlooked and underestimated icon, totem, and symbol of our times, They serve a function, a function that is loaded with ramifications - a means of consumerism on the one hand, a means of transporting one's worldly goods on the other. They represent the American dream and also it's a

I've been working with ChatGPT on my Pink Eye project (link here to my YouTube playlist). This is my first experience working with an AI like Chatgpt, and so far it has been very fruitful. Here's one short description of the musical aspect of the project that I wrote for YouTube: "Where the groove comes first, no matter the pace. Fusions of trip hop, big beat, hip hop, jazz and cinematic orchestration. Bring on the breaks! Pink Eye was conceived near the turn of the millenniu