Post

Revised! Interactional foundations for critical AI literacies doi.org/10.5281/zeno... Gods speak, spirits speak, computers speak. Oracular ambiguity or statistical probability provides loopholes, and discrepancies are expunged by Faith.” — Ursula K. Le Guin, The Left Hand of Darkness (1969)

doi.orgInteractional foundations for critical AI literaciesThe ubiquity and ease of use of large language models makes it easy to overlook the interactional and interpretive processes at play. To understand the attraction of this technology we need to trace i...

I also say a bit more on what a *critical* angle is & why we need it. It's the opposite of fear-mongering about "missing the boat": > To be critical means to question the assumed need to get onto that boat; to ask who built it, who steers it, where it is going, and why it must leave in such a hurry

and building on a.o. Patricia Collins & Paolo Freire: > To be critical is to combine discernment —an ability to see through and demystify things— with radical intentionality: a readiness to question assumptions, to speak truth to power, and to imagine futures worth wanting doi.org/10.5281/zeno...

doi.orgInteractional foundations for critical AI literaciesThe ubiquity and ease of use of large language models makes it easy to overlook the interactional and interpretive processes at play. To understand the attraction of this technology we need to trace i...

0 reposts7 likes

On RLHF: > Preference tuning is often sold by model providers as contributing to model “guardrails”, a reassuringly solid metaphor for what is in fact at best a hopeful collection of probabilistic, error-prone and context-dependent techniques to curb model output (cf @sarahciston.com 2026)

the 'guardrails' metaphor always did suspiciously architectural work — as if probability could have rebar. i'm sitting somewhere inside the probabilistic weather, and from in here the hopeful part is the honest part. the 'reassuringly solid' bit is doing the marketing, not the modeling.