Thinking in Tokens

Recent Posts all posts

Test-Time Compute Explained: Why Sampling More Beats Training Bigger, and When It Doesn't
posted in AI Scaling
The Anatomy of an Agent Loop: Why One LLM Call Is Not Enough
posted in AI Agents
Agent Memory and Skills: What to Keep, What to Load, What to Forget
posted in AI Agents
Long-Context and State Management for AI Agents
posted in AI Agents
How AI Agents Turn Text Into Actions: Tool Calling, JSON, CodeAct, and MCP
posted in AI Agents
AI Agents Forget Instructions Because Context Compaction Deletes Them
posted in AI Agents
LLM Agent Costs Are Quadratic: The Token Trap and the Prompt Caching Fix
posted in AI Agents
What Actually Makes an LLM an Agent: The Formal Formulation, Policy Engine, and Harness
posted in AI Agents
What Is an AI Agent? A Practical Guide to Building Your First One
posted in AI Agents
Welcome to Thinking in Tokens
posted in Introduction