Agentic AI with Pete

Agentic AI is different. Be productive with agentic AI for knowledge work now.

Small group, six weeks from a calm and compassionate instructor. Get set up, get grounded, get productive.

12 sessions, 75 min each Two tracks, 10 seats each Starts week of March 10, 2026
Pete Kaminski

Agentic AI is a new way of working with AI. Not chatting. Not prompting. Working together, with an AI that can read your files, manage projects, and take real action on your computer. It's a genuine leap in what's possible.

It's also a minefield. Too many tools, too many choices, too many opinions. Most people who try to figure it out on their own get lost, or exhausted, or give up.

I've been working with agentic AI full-time for a year, building real products, developing work habits that work, and mentoring people through the transition. I've found a clear path through the minefield, and I've walked it with people at a variety of skill levels. It works. This course is me walking that path with you.

We'll use Claude Code, installed via plugin using Obsidian or VS Code as the file manager. This is straightforward to use but not a starter kit. It's full-strength, the same tool seasoned professionals use. We'll also compare other agentic AIs and models.

You need to be comfortable working with files and folders on your computer. The AI handles the fancy stuff.

Who is This Course For?

Knowledge workers ready for the bleeding edge

You write, research, manage projects, build documents, or coordinate teams. You've queried ChatGPT, but you haven't found a way to make AI a real part of how you work. You're ready to go beyond chatbots. You're okay with some rough edges while the technology is still new.

People who learn by doing, with a guide

This isn't a lecture series or a polished online course. Every session involves hands-on work with your own AI on your own computer. You'll publish your work, automate your repetitive processes, and build a personal knowledge system, all with me in the room, making sure to smooth the path when you need it.

Early adopters who value judgment over hype

There are a lot of AI courses selling quick tricks. This one teaches craft. You'll learn what agentic AI actually does, how to direct it, and how to develop your own judgment about tools that change every few months. The specific buttons will evolve. Your judgment won't go stale.

A note on privacy, data, and costs

Agentic AI means an AI that can read and write files on your computer. That's powerful, and it raises fair questions. What data gets sent where? How private is your work? How much does it cost beyond the course itself?

We'll cover all of this in the first session, with honest answers. The short version: you'll need a Claude subscription (about $20/month) on top of the course fee. Your data stays between you and Anthropic under their privacy policy. You control what the AI can access on your machine. No surprises.

What We'll Cover

This is the Founders Cohort — the curriculum is real but the packaging is raw. We'll adjust as we go, because AI changes fast and so should we.

Each session is 75 minutes: roughly 20 minutes of Pete explaining, 20 minutes of hands-on doing-it-together, 20 minutes of questions, and 15 minutes for getting started, winding up, and overflow. The pace is fast and fun. You'll capture the gist of each topic in the session, then go back to the recording and notes to make it your own.

  1. Setting up YOUR agentic AI Tue Mar 10 (A) / Tue Mar 10 (B)

    We'll install Claude Code on your machine through Obsidian or VS Code and get you working with it before the session ends. Catch-up sessions available if you need extra help.

  2. Easy version control with agentic AI Fri Mar 13 (A) / Thu Mar 12 (B)

    Version control isn't just for programmers. Learn how to track changes, undo mistakes, and keep a clean history of your work, with your AI doing the heavy lifting.

  3. Publish - practice putting work out into the world Tue Mar 17 (A) / Tue Mar 17 (B)

    Your first real project: a sample "business card" website, built collaboratively with your AI. You'll see how to direct it, review its work, and iterate.

  4. Building your personal knowledge base Fri Mar 20 (A) / Thu Mar 19 (B)

    Organize what you know into a structured, searchable system. This is the foundation of everything else — a knowledge base that grows with you and makes your AI progressively smarter about your work.

  5. Project management for agentic AI Tue Mar 24 (A) / Tue Mar 24 (B)

    How to break work into pieces your AI can handle, track progress, and manage the collaboration. The meta-skill that makes everything else work better.

  6. Evaluating AI writing and coherence at scale Fri Mar 27 (A) / Thu Mar 26 (B)

    What happens when a project gets big? Could an AI write something large but still coherent? You'll see how to work with consistency, quality, and your own voice across extensive documents, whether that's a book, a series of articles, or a complex proposal.

  7. Building rich records from group meetings Tue Mar 31 (A) / Tue Mar 31 (B)

    Instead of a simple AI summary, build a densely-linked, super-informative knowledge base recapping a group meeting. A practical skill you can use at work immediately.

  8. Choosing harnesses, models, and managing your data Fri Apr 3 (A) / Thu Apr 2 (B)

    Under the hood: how to pick the right model for the job, manage what data goes in and comes out, and make informed decisions about the tools you use.

  9. Comparing Claude Code with other tools Tue Apr 7 (A) / Tue Apr 7 (B)

    An honest comparison of the tools available now. Not advocacy — a judgment exercise. What criteria should you use for evaluating tools that change every few months?

  10. Creating a web app Fri Apr 10 (A) / Thu Apr 9 (B)

    How you can ship real software, and why you might want to. Plus a brief overview of product management, the way to build the right thing.

  11. Tune and package specific “agents” aka “skills” Tue Apr 14 (A) / Tue Apr 14 (B)

    Your AI gets better when you teach it how to do specific work you need. Learn how to build, share, and acquire packaged custom instructions, workflows, and knowledge for you and your AI.

  12. Living with AI superpowers Fri Apr 17 (A) / Thu Apr 16 (B)

    The meta-session. How to sustain a productive partnership with agentic AI. You're the bottleneck on judgment and big-picture decisions — and that's by design. We'll cover pacing, keeping up with the state of the art, and knowing when to step away and let your brain do the work.

Bonus: Graduation day (week 7) Tue Apr 21 (A) / Tue Apr 21 (B)

An optional session. We look back at six weeks of working together, take stock of what stuck, and talk about where to go next.

Why is This Course Worth It?

You'll be productive, not just informed

Most AI courses leave you with concepts. This one leaves you with a working setup, real projects you built, and the skills to keep going on your own. After six weeks, agentic AI will be part of how you work.

Small group means real attention

Ten seats per track. One instructor. Personal 15-minute consults whenever you need them. This isn't a webinar — it's a working group where Pete is paying attention to whether you're keeping up.

You'll build judgment, not just habits

AI tools change fast. The specific buttons and commands will evolve. What won't change is your ability to evaluate tools, direct AI work, and know when to trust it and when to review. That's what we're really building here.

A year of hard-won experience in 15 hours

Pete has been working with agentic AI full-time for a year — building products, developing workflows, mentoring people through the transition. This course is the condensed version so you can skip the false starts and dead ends.

Deeper access to Pete's knowledge and expertise

Course participants get 50% off personal coaching sessions with Pete for two months.

Schedule & Details

Two tracks, same curriculum. Pick the time that works for you. Attend either session any given week if you need to shift your usual track.

US/Europe Track (A)

Tuesdays & Fridays
8:00 am PDT / 11:00 am EDT
London: 3:00 pm GMT → 4:00 pm BST from Mar 29
Central Europe: 4:00 pm CET → 5:00 pm CEST from Mar 29

US/Australia Track (B)

Tuesdays & Thursdays
3:00 pm PDT / 6:00 pm EDT
Sydney/Melbourne: 9:00 am AEDT (next day)
→ 8:00 am AEST from Apr 5 (clocks fall back)

See session times around the world → (opens in new window)

Both tracks share the same Discord community. All sessions are recorded. You can go back and watch the recording from both tracks.

Sessions
12 sessions, 75 min each
Twice a week for 6 weeks
Starts
Week of March 10, 2026
Cohort Size
10 seats per track
Community
PKAI Discord
Support between sessions
$1,995*
$795
Founders Cohort pricing — first run, limited seats
Future cohorts will be priced at the full rate
Both tracks are sold out for the Founders Cohort!
Email Pete to be notified of future courses

*Estimated market value. This course has not yet been offered at that price.

Prefer one-on-one? Private sessions are available at $200 $100/hour for course participants — get in touch.

Who is Pete?

Pete has spent 35+ years building things with emerging technology — and then helping other people do the same. He has a pattern: he spots transformative tech early, builds real products with it, and teaches others how to use it well.

He started in the early 1980s writing video games at Activision and Cinemaware. In 1993, he led development of one of the first GUI internet navigators — years before mainstream web browsers. In 2003, he co-founded Socialtext, one of the first enterprise wiki and collaboration platforms. In 2024, he co-founded YouBots.ai, an AI SaaS platform with multi-agent orchestration. He started using Claude Code the day it launched in February 2025 and has used it daily since as his primary development tool.

In between, he's been a fractional CTO, a consultant on Agile practices for the Cutter Consortium, and the founder of a five-year public-benefit project providing collaboration tools to distributed communities. He studied Computer Science and Literature at Caltech.

Pete is known for bridging technical depth with plain-language clarity — for helping non-technical people do sophisticated things with technology. That's what this course is. He's not here to impress you with jargon. He's here to walk you through it.

“What makes Pete exceptional isn't just technical depth. It's how he shows up. He has a teacher's heart. He's calm, measured, and incredibly clear. When I don't understand something that lives in his world, he explains it in terms that fit in mine. No ego. No overwhelm. Just steady forward motion.

Learning directly from him has fundamentally changed how I approach my work. My productivity, leverage, and confidence working with AI are on an entirely different level because of it. If you get the opportunity to learn from Pete, take it.”

— Nate St. Pierre, co-founder of YouBots.ai

Want to Talk Before Signing Up?

Not sure if this is right for you? Let's find out. Book a free 15-minute call and we'll figure out together whether this course fits where you are right now. No pressure. Sometimes the honest answer is "not yet," and that's fine too.

Book a Free Call