Theology of LLMs
Build Little Worlds
Microcosms, language models, and the Logos
Finite minds learn large worlds by building little ones: syllables, tokens, equations, diagrams, primes, parables, and models small enough to hold in view.
Large language models make that ancient movement newly strange. Tiny formal units scale into systems that disclose surprising order in the large world of human language.
The frame
A little world is not an escape from reality. It is a bounded form where pattern can become visible. The small thing is limited enough to examine, but rich enough to point beyond itself.
Microcosm and scale
Theology, mathematics, and science have all learned from microcosms: seeds, words, numbers, model systems, and controlled examples. We abstract away the full complexity of the world so that a real order can be seen.
LLMs add the opposite motion. They begin with tokens and probabilities, but gain new power by scaling up: more text, more parameters, more compute, more layers. The site follows that back-and-forth between the little thing and the large world.
First exploration: Theology of LLMs
Large language models are the first major case study: small units of language and mathematics scaled into machines that reveal unexpected structure in human speech, writing, analogy, genre, and meaning.
Latest posts
All postsLoading posts...