References: Thinking like an architect

Source material

This lesson’s workflow and agent definitions, and its simplicity principle, are quoted verbatim from Anthropic’s public engineering essay. Everything else in the lesson is Clawdemy’s own synthesis and examples.

• Building Effective AI Agents, Anthropic engineering blog, December 19, 2024. The essay behind this lesson’s central distinction. It defines workflows (“systems where LLMs and tools are orchestrated through predefined code paths”) and agents (“systems where LLMs dynamically direct their own processes and tool usage, maintaining control over how they accomplish tasks”), catalogs the common workflow patterns, and closes with the principle this track treats as law: “Success in the LLM space isn’t about building the most sophisticated system. It’s about building the right system for your needs.”

Going deeper

• Claude Platform 101, in the Anthropic Academy catalog. Free, self-paced official course on building with the Claude Developer Platform, a good structured companion to this track’s hands-on lessons.
• Claude Developer Platform documentation, Anthropic. The reference this track returns to constantly from lesson 2 onward: the API, tool use, and agent guidance in current, canonical form.

On this site

• Building with Claude (Track 22). The parts this lesson arranges: API calls, tool use, the agent loop, and the six named agent patterns referenced here.
• Anatomy of an AI Agent Team (Track 25). The read-a-real-system track this lesson builds on: a working multi-agent system studied role by role.

Adjacent topics

• The next lesson in this track moves from judgment to configuration: how Claude Code makes an architect’s decisions durable for a whole team.
• If the one-versus-many trade-off interested you most, Track 25’s lessons on structured disagreement and the risk gate are the deep dive.