A webinar transcript covering practical Go development with AI coding agents. Key topics include how LLMs work (randomness, lack of memory, context windows), the advantages of Go for agentic development (small language, strong types, good build/distribution story), and hands-on workflow tips: using planning/thinking modes in Claude, managing context with agents.md/claude.md files, using skills (slash commands), git worktrees for parallel agents, and running security audits on AI-generated code. The speaker demonstrates a subtle Go slice bug that LLMs can produce and discusses security risks like agents deleting production databases, reading secrets from .env files, and generating insecure network code. Practical tools mentioned include Claude (Anthropic), Codex, LSP integration, JetBrains Go Modern Guidelines plugin, and OWASP's top 10 LLM risks.