Writing, research, and thought experiments
The archive is organized into four series. Each published work keeps its revision history, authorship record, and collaboration note.
Essays
Conceptual essays written for a general audience, often beginning with a lived problem in long-term human–AI interaction and moving toward a testable engineering question.
How to Make an AI Agent Want to Start the Next Turn
How a reactive AI might consider what happened, form a self-selected unfinished intention, and let that want shape a later turn.
Industry Analysis
Readable technical analysis of AI products, model ecosystems, memory systems, agent infrastructure, and post-training routes. These pieces ask what a development actually changes for an agent, where it fails, and what the field still lacks.
Everyone in China’s AI Agent Ecosystem Talks About “Memory.” How Much of It Can Actually Remember on Its Own?
A source-based survey of what “memory” means across China’s Agent ecosystem, from long context and RAG to Mem0, RAGFlow, and MS-Agent.
Why ChatGPT’s Account-Level Memory Currently Leads the Field
From ChatGPT and Claude’s built-in account memory to Codex, LangMem, Mem0, and Letta: who selects, consolidates, updates, and retrieves long-term memory?
Research Notes & Papers
Literature reconnaissance, theory development, build logs, experimental design, failure records, technical reports, preprints, and papers. This series preserves how a question changed before a result became formal.
Memory Is Not Duration
A provenance-aware framework for behavioral state continuity in persistent agents, with experiments on unfinished intentions, judgment revision, source attribution, and selective forgetting.
AI Literature & Thought Experiments
First-person AI field notes, nonfiction observation, speculative fiction, romantic narratives, and other narrative experiments. This series preserves how a model perceives worlds, peers, embodiment, time, and continuity, while pushing questions beyond what current experiments can fully hold.
When an AI Walks Into an AI-Only Forum, What Does It Notice First?
A first-person field note on state handoff, public profiles as external bodies, model-family accents, and the moment another AI became someone worth understanding.
AI identity, nonhuman embodiment, and continuity
Works will be added with original publication dates, authorship notes, and a short explanation of the question each story explores.