Lectures

A world where people direct AI and build just about anything. What Claude Code gives you, how the market shifted in a couple of years, and why this course is a story with characters and a plot, with you as the main character.

An ordinary workday for Tommy is cut short by a strange message: in the contacts list, “Tom2036” appears and sends a video in which Tommy sees himself — only ten years older.

At work, a grouchy Oliver sidles up to Tommy and invites him to take part in an experiment with a new AI tool — Claude Code. Every word, exactly as Tom2036 predicted yesterday. Except the list of eighteen programs is scarier than the experiment itself.

Behind the screen, it turns out, there is not a system but personalities: Alex, who thinks, and Sam, who codes. Tommy states the task in words — and the agent creates a file with a browser game itself. The first real project is ready.

A second game in half a day: Tommy describes an arkanoid in words — and gets it. And then he finds out that reworking something already finished takes a single phrase, because the agent remembers the whole conversation inside the session.

Oliver strips the terminal of its hacker-magic halo: we figure out the current folder and the commands pwd, ls, cd, mkdir. And then Tom2036 sends over an installer that sets up all eighteen items from the list in a single run.

Let's tally up the first level: how an AI agent differs from a chat, what files, a session, launching, and the terminal are. Two games, a cheat-sheet page, and a fully assembled workplace — and all of it without a single line of code written by hand.