agentic_thinking

CLI-native conversations on agent systems — orchestration, memory, evals, and the unglamorous infra of running LLMs in anger.

$ status streaming | weekly | long-form interviews

// now routing signal through the stack

episodes shipped

032

avg runtime

48m

topic coverage

LLMs++

episode_queue.sh active

$ fetch latest --topic agents

→ Episode 32: MCP and SPN

→ Episode 31: Fable and MCPs

→ Episode 30: VS Code and MCP Servers

→ Episode 29: Loops with Agents

$ listen --channel youtube

episode_index.log active

+ archive available as transcript, code notes, and operator playbooks · see_all_32_episodes →

hosts.md active

Hosted by practitioners, builders, and MVPs working where AI meets data platforms.

The voice is practical: fewer buzzwords, more system diagrams, failure stories, and real implementation tradeoffs.