Research

Authors: Terry Chen, Kaiwen Che, Matthew Song Abstract Despite advances in large language model (LLM) capability, their fundamental limitation of not being able to store context over long-lived interactions persists. In this paper, a novel human-inspired three-tiered memory architecture is presented that addresses these limitations through biomimetic design principles rooted in cognitive science. Our approach aligns the human working memory with the LLM context window, episodic memory with vector stores of experience-based knowledge, and semantic memory with structured knowledge triplets. ...

AI-powered study group assistant that analyzes real-time conversations, detects misconceptions, and facilitates deeper learning through Socratic questioning and contextual knowledge retrieval.

Observing and understanding the strange quirks of individuals and crowds

Exploring the evolution of technology and its impact on society.

What becomes possible when language models can hold context, make decisions, and execute across long horizons?

Different places run on different assumptions — about infrastructure, trust, money, and daily life. Traveling is the fastest way to see them.