When we look at a representation of reality, we can choose to either see it as descriptive, meaning it tells us what the world is currently like, or as prescriptive, meaning it tells us how the world should be. Descriptions teach us, but they also give us room to innovate. Prescriptions can get us stuck. One place this tension shows up is in language.
Description not evaluation - The Cynefin Co
One of the things I am enjoying at the moment is the way a lot of things are coming together conceptually. It happens like this when you develop or repurpose knowledge from different sources. Individual practices and ideas make sense in their own right. Then you read more, practice more, and the various origins […]
The Art of Writing One-Sentence Product Descriptions
A look back at early interviews with Facebook and Uber CEOs illustrates an ingenious way to communicate hard-to-describe products.
Tracing the thoughts of a large language model
Anthropic's latest interpretability research: a new microscope to understand Claude's internal mechanisms
Misarticulation: Why We Sometimes Feel Our Words Don’t Match Our Thoughts
People do not always say what they mean. In everyday conversations, people regularly sense that they have not fully communicated what they had in mind—a subject
Job Descriptions Should Be Better
Most job descriptions are terrible. We need to make them better.
The Triadic Mind: How Language Reveals the Limits of Human Cognition
Languages are the most complex symbolic systems humans have ever created. Yet children acquire them effortlessly, without formal…
How to counter "everything is a construct" worldview?
Many of the people in humanities I encounter seem very wedded to mindset I refer to as "construct = everything". For example, they say bazar things such as gender is 100% a construct. They say...
Say precisely what you mean.
The words designers use when they know what they’re looking at.
Computational hermeneutics: evaluating generative AI as a cultural technology
Generative AI (GenAI) systems are increasingly recognized as cultural technologies, yet current evaluation frameworks often treat culture as a variable to be measured rather than fundamental to the system's operation. Drawing on hermeneutic theory from the humanities, we argue that GenAI systems function as "context machines" that must inherently address three interpretive challenges: situatedness (meaning only emerges in context), plurality (multiple valid interpretations coexist), and ambiguity (interpretations naturally conflict). We present computational hermeneutics as an emerging framework offering an interpretive account of what GenAI systems do, and how they might do it better. We offer three principles for hermeneutic evaluation—that benchmarks should be iterative, not one-off; include people, not just machines; and measure cultural context, not just model output. This perspective offers a nascent paradigm for designing and evaluating contemporary AI systems: shifting from standardized questions about accuracy to contextual ones about meaning.
Metaphor and the Political War
Metaphor and the Political War
Language Access in Healthcare — Discourse Graph
An open, AI-assisted evidence synthesis of how language concordance — matching patients with providers or interpreters who share their language — affects healthcare outcomes. Every question, claim, evidence item, caveat, and source is its own addressable node.
Development Beyond Information: Human Becoming in the Age of AI
Abstract. Discussions of artificial intelligence in education increasingly equate development with the acquisition and deployment of information. This article argues that such accounts rest on false assumptions. Development is better understood not as an informational state but as a relational, historical process of becoming through which persons emerge and continue to change. Drawing on Vygotsky, on Ilyenkov's distinction between living and petrified ideality, and on the enactive approach to cognition, we characterise development through four features: its historicity, its dependence on reciprocal sense-making, its mediation by others who orient the learner, and its cultivation within relations of care. Large language models operate over the sedimented products of these processes without participating in them. As objectifications of human purposes, such systems may mediate human development or foreclose it. We conclude by considering what this account implies for education.
God Help Us, Let’s Try To Learn About Mechanistic Interpretability Techniques
...
