Causal Ghosts →
On counterfactual inference, moral responsibility, and the irreducible uncertainty in what never happened.
causal-inferencecounterfactualsepistemologymoral-philosophypearl-ladder
Gradient and Reason
On machine knowledge & philosophical loss.
Machine Learning · Data Analysis · Philosophy
I. The Project
At the intersection of empirical rigor and philosophical inquiry.
Gradient and Reason is an archival effort to bridge the gap between technical execution and deep understanding. We explore the structural causal models that define our systems, the latent manifolds that hold our memories, and the ethical foundations that guide our progress. Here, machine learning meets the history of thought.
II. Featured Essays
The Hallucination and the Lie →
Why language models cannot lie, why they cannot help but bullshit, and how the mathematics of next-token prediction decouples testimony from the witness.
language-modelsepistemologyphilosophy-of-mindhallucinationcross-entropy-loss
The Topography of Loss →
Navigating high-dimensional latent spaces as a metaphor for human memory, grief, and machine unlearning.
representation-learningunlearningconcept-driftgriefmemory
III. Reading Notes
"Civilization advances by extending the number of important operations which we can perform without thinking about them."