AI Is Not That Hard

AI Primer

The foundation season of AI Is Not That Hard. One question per episode, ten to fifteen minutes, no prior knowledge assumed.

Listen in order if you are new to all of this. If you are not, skip to the one a main episode pointed you at. Episodes 1 to 16 are released together on launch day; the rest follow weekly.

If you're ever stuck, ask your agent.

Why now, and what you are actually dealing with

  1. AI Primer: how to use this seasonWhat this is, who it is for, and in what order to listen. The bottleneck-moved idea in five minutes.
  2. A model is not a programWhy the AI gives different answers each time, makes things up, and forgets.
  3. Chatbots, copilots and agentsThe difference between asking a chatbot a question and giving an agent a job with tools and permissions.
  4. Vibe coding versus building a systemWhat separates a demo that works once from something you can run for a year. The five basics.

The tools you will touch

  1. What a terminal is, and why AI made it matter againThat black window, why agents live there, and why it is less scary than it looks.
  2. Using a terminal: the ten commands you needWhere am I, what is here, how do I move, make, look inside, and get out.
  3. What a CLI isWhy so many tools have a command-line version, and why agents prefer them to clicking.
  4. Coding agents: Claude Code, Codex and friendsWhat they are, and which one a beginner should start with.
  5. Your first session with a coding agentWhat to say, what it will ask you, what permissions are, and when to say no.
  6. Standing instructions: CLAUDE.md and AGENTS.mdHow to stop repeating yourself, and why every serious project has a file addressed to the agent.
  7. Git and GitHub: the undo buttonWhat version control is, what a commit is, and why it lets you be brave.
  8. APIs and API keysHow programs talk to each other, what a key is, and why it must never go in the code.
  9. Models, tokens and what it costsWhy there are small, medium and large models, what a million tokens costs, and how to cap the bill.
  10. MCP servers: giving the agent handsHow an agent reads your email, drives a browser or queries your analytics.
  11. Spec first: writing a brief an agent can build fromWhat to write down before any code exists, and why this is the product manager's home turf.
  12. Talk, don't type: voice input for briefing agentsWhy dictating to an agent gets more out of you than a tidy typed sentence.

Where things live and how they keep running · weekly after launch

  1. Hosting options: from laptop to cloudWhere a thing can run, and what each rung of the ladder costs.
  2. Databases in plain EnglishWhere the data lives, SQLite versus Postgres, and when it matters.
  3. Scheduling: making it run while you sleepWhat cron and launchd are, and how "every 30 minutes" actually happens.
  4. Domains, DNS and getting it on the webHow a name points at a thing, and the minimum needed to put something live.
  5. Logs and knowing it worksHow to know the system did what you think, and how to find out when it did not.
  6. Permissions and guardrailsWhat the agent may touch, how to fence it, and why it must never widen its own fence.
  7. Backups, undo and blast radiusWhat happens when it goes wrong, and how to make wrong recoverable.
  8. Pausing and killingHow to turn the bill off without losing the work, and how to decide to.

Wrap and routing · weekly after launch

  1. Twenty-five terms in twenty minutesA spoken glossary in alphabetical order.
  2. AI Primer: where to go nextGiven what you want to build, which main episodes to listen to.