Stop Asking Claude to Agree With You
The best claude skill I’ve found writes no code. It just won’t let me start until I know what I’m building.
AI & Machine Learning
Truncating LLM Context Is Not Retrieval: Why slice(0, N) Was the Wrong Fix
We capped a 2.6M-character LLM prompt with slice(0, 10000). Latency dropped, dashboards turned green, and the model started answering with the wrong context....
Building Scalable AI Agents: A Journey Through Multi-Agent Architecture
Or: How I Learned to Stop Worrying and Love the Token Budget