101: The Landscape
Where are we? AI coding tools run from autocomplete to agents that carry out multi-step tasks, and in the reading of a survey taken in April 2026, most developers stay at the autocomplete end. This chapter maps that landscape for one human working with one agent. Moving along the spectrum takes a new mental model more than a new tool. The tools are plentiful and the methodology around them is scarce, so their documentation explains features and leaves the practice to you. A figure about unused capability or productivity needs its source, its date and its counting convention, or it is an estimate. And once code is cheap to write, the scarce skill is knowing what is worth building. The exercises install Claude Code and run a first task.
The chapter opens the first year, for readers new to agentic coding: one human and one agent. The second year gives the practitioner a method at project scale, and the third runs multi-agent fleets.
Materials
This Chapter's Insights
- ACI-001: The Agentic-Assistive Spectrum
- ACI-002: The Methodology Desert
- ACI-003: The Ten Percent Problem
- ACI-004: The Tool-Practice Gap
- ACI-005: The Productivity Claim
- ACI-006: The Generation Trap
Browse every insight on the course map.