May 24, 2026May 24, 2026 · 3 min read · 5 viewsDay 04: Why more context can make an AI system worseWhy a larger context window does not remove the need for context discipline.3 min read · 5 views
May 23, 2026May 23, 2026 · 3 min read · 9 viewsDay 02: Why LLMs feel intelligent but still generate one step at a timeThe practical reason LLM behavior needs constraints, validation, and repeatable output design.3 min read · 9 views
