Research
What we're thinking about.
Start here12 min read
Teaching Machines to Understand Humans
Jonathan Politzki
Recommendation, cognitive science, and simulation are approaching the same question from different directions: can a model learn to predict human behavior? The write-up brings those fields together and lays out what it would take to build a foundation model of people.
Papers
SEP 2026Toward a Foundation Model of Human BehaviorPDFThe working paper behind the essay. A definition, a survey of three converging fields, and the data and scale required to build a foundation model of people.FEB 2026Local Drift-Adapters: Mixture-of-Expert Embedding Translation for Heterogeneous Vector DatabasesPDFRe-embedding a large corpus to adopt a better model is prohibitively expensive. Eight per-cluster adapters close 70% of the gap to oracle.DEC 2025AI Memory: A Comprehensive ReviewPDFMemory in LLM systems from first principles: the Experience, Storage, Recall pipeline, the tradeoff between retrieval precision and semantic coherence, and the architectures navigating it.
Writing
AUG 2026How Much Catalog Fits in a Semantic IDA semantic ID with three levels of 256 codes has 16.8 million addresses. On a 66,000 item catalog it collides on one item in nine. The gap between those two numbers is the whole design problem, and it is measurable from the collision rate alone.MAY 2026The Internet of Agents Needs a Model of YouProject Deal showed AI agents can transact for us. When everyone has 100 agents acting in their name, the load-bearing question is whether any of them actually represents the person who sent them.APR 2026Emotion Vectors and the Future of Human ModelsWhat Anthropic's new interpretability research means for foundation models for human behaviorAPR 2026Why General-Purpose Embeddings Fail at Modeling PeopleAnd what outcome-trained representations get rightFEB 2026Local Drift-AdaptersRe-embedding a billion-document corpus to upgrade a model is prohibitively expensive. Eight per-cluster adapters close 70% of the gap to oracle.FEB 2026Introducing the Embedding AdapterZero-downtime migration between embedding modelsJAN 2026Jean TechnologiesBuilding foundation models for human behaviorDEC 2025The State of AI MemoryYou cannot model a person you have not collected context about, so we built memory first. The model is the meta of memory, the representation that captures not just who someone was, but who they are and who they intend to become.OCT 2024General Personal EmbeddingsA trusted infrastructure for the age of AI
Open resources
ONGOINGEvals and the user-simulation leaderboardEvery measurement Jean runs on models of human behavior, with its ground truth and controls, and released simulators scored on the same held-out behavior.ONGOINGawesome-foundation-models-of-human-behaviorGitHubA public reading list of models pretrained on records of what people do. Issues and pull requests welcome.