Sep
30
Your first AI agent in the terminal
You start having never used one and leave with OpenCode installed, connected, and your first change already made.
Wednesday, September 30, 2026
05:30 PM a 07:00 PM GMT-6 · 90 min
Online
We email you the link when you hold your seat
What you'll do
Plenty of people hear about AI agents, get the idea roughly, and never actually open one. Not for lack of interest: the first step is full of decisions nobody explains. Which tool, which model, what you have to pay for, what happens when it gets things wrong.
This session exists to take that first step with you, live. There are no slides about what artificial intelligence is. We open a terminal, install OpenCode, connect it, and ask it to do things on a real folder, so you can see what it does well, what it does badly, and where the limit sits.
You leave with something working on your machine, not with notes.
OpenCode installed and working
On your machine, not on the presenter's.
An agent connected to a model
Actually answering, ready to take jobs.
Your first change made by an agent
Asked for by you, reviewed by you, understood by you.
Judgment about what to ask it
Which jobs it handles well today and which it doesn't, so you don't lose an afternoon.
Who it's for
A genuine first time. If you already use agents daily, this isn't your workshop.
It's for you if…
- You have never used an AI agent and don't know where to start
- You've heard about this for months and want to see what it actually is
- You know some programming, even a little, and can handle files and folders
- You'd rather start with open source tools you can look inside
It's not for you if…
- You already work with agents every day
- You want to compare models, run benchmarks, or tune configurations
- You want a session on AI theory
What to bring
Deliberately short. A laptop and ninety minutes is enough.
- A laptop running Windows, Mac, or Linux. Anything from the last five years works
- Node 18 or newer. If you don't have it, we install it in the first few minutes
- A free account with the model provider we'll use. We email you the link and the steps before the session
- An empty folder to practice in. No project required
How the time is split
The topics we cover, in this order.
What an agent is, and how it differs from a chat
The difference isn't the model: this one can read and write files.
Installing OpenCode and connecting it
We do it together, step by step. If something breaks, we stop there.
Its first job: read, then explain
We hand it a folder and ask it to tell us what's inside.
Its first change: it writes, you review
This is the part that matters: how to review what comes back before accepting it.
What not to ask it yet, and where to go next
Today's limits, and the path if you want more.
Frequently asked questions
Do I need to know how to program?
Barely. If you know what a file and a folder are, and a terminal doesn't scare you, that's enough to follow the whole session.
What do I need to sign up for?
Nothing. The workshop is open and OpenCode is open source. We connect it to a model with a free tier, so you can follow the whole session from minute one.
Does it work on Windows?
Yes, on Windows, Mac, and Linux.
What if I get stuck installing?
The first few minutes are exactly for that. We move at the same pace and don't continue until the room is ready.
Is it recorded?
Yes. The recording goes out by email to everyone who held a seat.
How is this different from the Claude Code workshop?
This is the first time: install, connect, and understand what an agent does. The Claude Code one assumes you already work this way and runs on real repos. This one first, that one after.
If you've never opened one, this is the day
You bring a laptop and leave with an agent running on it.
Hold my seat