Synthetic Participants (Silicon Sampling)
Language models can stand in for survey respondents. Whether that works depends on two things at once: what the sample is asked to show, and how the model was conditioned.
Writing
On cognition, AI, organizations, policy, and whatever I can't stop thinking about. Published here, on LinkedIn, and on Substack.
Language models can stand in for survey respondents. Whether that works depends on two things at once: what the sample is asked to show, and how the model was conditioned.
Language models can label text at scale. Turning those labels into a valid estimate takes a statistical correction, and it works only under four conditions.
AI engineers and research methodologists use the same words for different checks. A researcher can pass one check and fail the other without noticing.
What Generative AI Accidentally Gets Right About Dialectics — a mapping of the GAN training process onto Hegel's dialectical idealism and Adorno's negative dialectics. GAN training is structurally Hegelian at individual stages but Adornian in its overall trajectory.
Fisher through Bourdieu — how the doxa of capitalist realism forecloses the conditions of possibility before imagination can even begin. A reading of Mark Fisher's hauntological critique through Bourdieu's field theory, orthodoxy, and the structured structure of the thinkable.
Charity is not a virtue, a transaction, or a bond. It is a mode of existence — a crossing. Drawing on Latour's AIME framework, a proposal for [CHA] as a new mode, and what happens when algorithmic infrastructure captures the legibility of need.
The myth of American self-sufficiency runs directly against human evolutionary biology. Drawing on Hrdy, Hawkes, Gould, and Eisenberger — the case that interdependence isn't weakness, it's the original human technology.
What happens to knowledge workers when AI automates the repeatable? A conversation with Google Gemini about C-suite role inversion, credentialing collapse, and what skills survive.