Field Notes · Closing

LLMs as a Cognitive Virus

This article examines large language models through the metaphor of a cognitive virus, exploring how LLMs might spread ideas or influence thought patterns.

OKF Agent Memory – Git-native persistent memory for AI coding agents

OKF Agent Memory is a Git-based storage system designed to provide persistent memory capabilities for AI coding agents.

LLM representations have implicit symbolic structure

This paper investigates whether large language models' internal representations contain implicit symbolic structure rather than being purely continuous embeddings.