I’m obsessed with complex adaptive systems—their chokepoints, feedback loops, and the signals that reveal their inner workings. This obsession drives me to create knowledge that compounds over time and systems that grow stronger under stress. Not fragile artifacts that crumble at the first shock, but antifragile ones that thrive on volatility.
What follows are artifacts of that wandering—writings, projects, research, companies, and tools born from the pursuit of lasting value.
Writings
- 12 Factors of Knowledge Capitalism - A framework for understanding how knowledge creates, compounds, and captures value in networked systems. Twelve foundational principles describing the emerging architecture of an economy where knowledge itself becomes the primary form of capital.
Projects
GitHub: github.com/remyjkim
- cc-architectures - Reference architectures for building claude-code like apps and architectures. Covers patterns for multi-agents, to orchestrator-worker models, skills, and more.
Research
-
Survey on Exploration in Reinforcement Learning - A comprehensive survey of exploration methods in RL, covering intrinsic motivation, curiosity-driven learning, and count-based approaches.
-
Mutual Information & State Intrinsic Control with Tsallis Entropy - Experiments on using Tsallis entropy for intrinsic motivation in state-space exploration.
-
SeekerGym: Benchmarking Agentic Information Seeking under Uncertainty - An evaluation framework for assessing how well LLM agents autonomously search for and integrate relevant information, featuring belief-structuring and meta-reflection techniques.
Companies
- Curation Labs - Building 21st century’s knowledge capital infrastructure.
Tools
-
12-factor-knowledge-capitalism - The framework repository for Knowledge Capitalism—twelve principles for understanding how knowledge creates and captures value.
-
local-mlx-server - A lightweight local inference server built on Apple’s MLX framework. Enables running LLMs locally on Apple Silicon with minimal setup.