Presentations
Academic Presentations
A Few Insights from the Analysis of Over 2,000 AI Agent Logs
In this lightning talk, I present a few highlights from the HAL log analysis, including the prevalence of shortcuts and regression-based models of success and failure.
Is Consciousness Prerequisite for Moral Patienthood? The case for stretching our moral intuitions and avoiding the hard problem.
In this talk, I argue against the prominent view that consciousness is the proper primitive for moral patienthood, instead grounding moral status along two dimensions, valence and resonance, which provide a smoother gradient for moral status.
Differences in the Moral Foundations of Large Language Models
An analysis of the moral foundations of sixteen frontier language models using synthetic experiments.
Professional Talks
Improving Imputation Accuracy
I present results from parametric and non-parametric regression imputation approaches to missing data in the Census survey of local governments.
Personal
A Few Observations on the Implications of LLM Architectures for Value Alignment and Consciousness
In this short presentation, I present a number of observations on the pre- and post-training processes for large language models and their potential implications for LLM consciousness.
A Discussion of Related Problems in Ethics and Consciousness
In this presentation, I relate the hard problem of consciousness to an equivalently hard problem in ethics, meta-ethical nihilism, and argue for strict naturalistic ethical reductionism as a mechanism for expanding our moral circle.