One of my first experiments at the intersection of algorithms and poetry. All is All Poetry is a model-agnostic framework for poetry generation.

The framework is a basic instruction on how to use language models like Markov chains and variational auto-encoders for poetry.

Since these early experiments, I have been viewing computational language generation as a way to discover linguistic oddities, or as Max Kreminski writes, to “lose poetry”. Much of my work is now on the critical literacy side: how can we fight back on the trends of AI intellectual labor theft using (sometimes technical) insights. Still, markov chain poetics is far from corporatized LLM’s, and generative poetics still provokes many interesting questions and unique insights about language, knowledge, and art.