@ts_floydDaisy Hollman, an Anthropic engineer, says the goal is not to prompt Claude yourself. Build graphs and loops that let Claude prompt itself. Her hour-long talk covers how Anthropic uses Claude Code internally to create agents with memory, error correction, and steady improvement across runs. Watch the presentation and then read the loop-building guide.

@ts_floydDaisy Hollman, an Anthropic engineer, says the goal is not to prompt Claude yourself. Build graphs and loops that let Claude prompt itself. Her hour-long talk covers how Anthropic uses Claude Code internally to create agents with memory, error correction, and steady improvement across runs. Watch the presentation and then read the loop-building guide.