Tag

LLM

3 posts

October 2026

The Instruction Was in the Ticket: Designing LLM Agents That Survive Prompt Injection EchoLeak, the GitHub MCP leak, and the Supabase support-ticket exfiltration all had the same shape: an agent read text an attacker wrote, and that text chose the agent's … Research 27 min read

May 2026

Build a Second Brain That Maintains Itself: A Complete Guide to Karpathy's LLM Wiki Most AI tools that answer questions about your documents rediscover everything from scratch on every query. Andrej Karpathy's LLM-wiki pattern flips that: the AI compiles … Guide 25 min read The Complete Guide to LLM Benchmarks (2026): What They Measure, How They Work, and Which One to Use If you are new to AI development and find LLM leaderboards confusing, this is the guide I wish I had. We walk through every benchmark that matters in 2026 — MMLU-Pro, … Guide 44 min read

All posts