Who it’s for
Graduate-level learners, including those without a computer science or technical background. If you come from the humanities, social sciences, policy, or other fields, you can succeed here: the curriculum is designed so that an AI Tutor supports you on demand, teaching programming, math, and tooling when you need them. Technical readers get the full agent-building path; everyone follows the same arc.
What makes it different
- Agent-first architecture. Every deliverable is an agent tool — a server your AI invokes. The orchestrator discovers and calls tools via the tool protocol (e.g. MCP).
- Metabolism of knowledge. AI as a catalyst in the circulation of ideas. Representation, retrieval, and reasoning as transformations within epistemic networks.
- Postmodern lens. Knowledge produced in networks, representation that excludes as well as includes, agency distributed across humans and systems. Optional readings from Borges, Dick, Blade Runner, and others.
- RAG, memory, and retrieval as first-class modules (Chapter 7).
- Tool schemas, ReAct, and orchestration (Chapters 9–10). Students build the tool client that discovers and routes to all tool servers.
- Labs as tools. Each chapter has three labs; every lab produces a tool. The capstone integrates them.
Structure of the book
Four parts, twelve chapters. Supports a three-semester MS degree or self-paced study.
Part I — Foundations Intelligence as Process, AI in Practice, Search and Planning
Part II — Learning Systems Learning from Data, Neural Systems and Representation, Language and Models
Part III — Memory, Reasoning, Action Memory Systems, Reasoning and Inference, Acting in the World
Part IV — System Integration Architectures of Intelligence, AI in Institutions, The Student’s Artificial Intelligence (Capstone)