1.4 What You Can Build¶
Abstract opportunity doesn't move anyone. This lesson gets concrete: real categories of things people build with agents today, what each takes, and how to pick your first three builds so the course has a target painted on it from day one.
What you will learn
- Map the main categories of agent-built projects, from weekend builds to products.
- Match each category to the course chapters that unlock it.
- Choose your personal build list — the projects you'll actually make during this course.
Builder principle
Pick the build before you learn the tool. Learning without a target is how courses go unfinished.
The build categories¶
Websites and landing pages¶
A common starter build: a small site for a business, portfolio, event, or community. Scope, content readiness, accessibility, review, and deployment constraints determine whether it takes hours, days, or longer.
Unlocked by: Chapters 3–6. Deployed live by: Chapter 11. Market hypothesis: a specific organization may value a new or improved site. Validate the need, alternatives, budget, ownership, and support obligation locally.
Personal automations¶
The quiet life-changers. A morning digest of news that matters to you. Files that organize themselves. A weekly report that writes its first draft before you sit down.
Unlocked by: Chapters 4–6, supercharged by Chapters 7–9. Value hypothesis: a repeated low-risk task may take less effort after automation. Measure two comparable runs and include maintenance and error time.
Tools and dashboards¶
Small software that solves one problem well: a booking tracker, an invoice generator, a content calendar, an inventory counter for a small shop.
Unlocked by: Chapters 5–8. Market hypothesis: a specific team may value a narrow tool more than its current manual process. Interview users and measure the current process before claiming demand.
Content and knowledge systems¶
A second brain that actually works: research assistants, draft generators tuned to your voice, documentation that maintains itself.
Unlocked by: Chapters 5 and 7 especially. Value hypothesis: a knowledge system may reduce repeated retrieval or drafting work. Test quality and total effort; more output is not automatically more value.
Full products¶
Apps with users, accounts, and payments. This is where the engineer lane deepens — testing, CI, code review, multi-agent builds — but no-code-lane graduates get here too, usually one project after their first.
Unlocked by: the whole course; Chapters 9–11 carry the weight.
Agentic infrastructure (engineer lane)¶
The meta-category: harnesses for real codebases, Skill libraries for your team, MCP servers for your company's internal tools, validation pipelines that make agent output trustworthy at scale. This is what the expensive cohorts teach — it's Track 3 and 4 here.
Pick your three¶
Right now, in your builder's journal, choose:
- The quick win — buildable by Chapter 6. Small enough to finish, real enough to matter. Example: a one-page site for someone you know.
- The time-saver — an automation attacking one task from your Lesson 1.1 list.
- The capstone candidate — the ambitious one. A product, a client build, a harness for your team. You'll refine it as you learn; write the current version now.
Write all three with this format:
BUILD: (what it is, one sentence)
FOR: (who uses it)
DONE MEANS: (what you can point at when it's finished)
That format is your first act of specification — the skill from Lesson 1.3. It's also the seed of every agent brief you'll ever write.
Reality check¶
Scope is a common failure source
Large or vague scope raises completion and review risk. If your quick win takes more than one sentence to describe, narrow it and test the smaller version first.
Check your understanding¶
- Which build category attacks the tasks you listed in Lesson 1.1?
- Why is "pick the build before you learn the tool" better advice than the reverse?
- What are the three parts of the build format, and which one forces the most honesty?
- What's your quick win — and can you say it in one sentence?