Causal Ghosts →
On counterfactual inference, moral responsibility, and the irreducible uncertainty in what never happened.
causal-inferencecounterfactualsepistemologymoral-philosophypearl-ladder
Archive
Essays
4 entries in the archive
2026
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
Prediction Is Not Understanding →
Why our most successful systems may still know nothing — and how empirical code uncovers the structural gap
causal-inferenceepistemologyinterpretabilityshortcut-learningsuperposition