Skip to main content

The AI Pro Accelerator

Are you an experienced developer who can't seem to get real results from AI tools? You've watched the demos, paid for the subscriptions, and tried the prompts, but something's not clicking. Give me 4 weeks and we'll fix that.

Thanks for dropping by and learning more about this program. I'll get right to it:

I want to help you break through to AI fluency and transform your career. Give me 4 weeks and you'll have a complete system that actually works for you.

If you're here, reading this, it's likely you're a capable developer who feels behind on AI. Maybe you've been coding for years, even decades, and suddenly everyone's talking about tools that make you feel like a beginner again. You've tried Copilot, maybe Claude Code, and the results are... fine? But not the productivity revolution everyone promised.

I've felt this too. I spent 3 years on the VS Code team at Microsoft working on Copilot strategy, and even I had moments where I thought, "Wait, am I doing this wrong?" The truth is that most developers are struggling with the same thing, but nobody wants to admit it because everyone else seems to have it figured out.

Here's the thing: prompting isn't the skill. The developers getting real value from these tools aren't writing better prompts, they're using the tools completely differently. They're building custom agents. They're writing hooks that automate tedious work. They're creating skills that turn complex workflows into single commands. They've moved beyond "ask the AI a question" into "architect systems where AI does real work."

That's what this program is about.

The Offer

To the point: I'm offering a 4-week intensive program where you and I will work closely together, with the goal of making you genuinely fluent with your AI tool's advanced features. I specialize in Claude Code, but I can adapt this program to your needs.

Specifically, we will:

  • Assess where you are. We'll look at your current workflow, your codebase, and how you're using AI tools today. No judgment - just clarity on what's working and what isn't.
  • Build your foundation. You'll understand how Claude Code actually works under the hood: the system prompt, CLAUDE.md, context management, and token economics. This isn't theory, it's the mental model that makes everything else click.
  • Master metaprompting. This is the skill that changes everything: having Claude write your prompts, instructions, and configuration. Why struggle to figure out the perfect way to ask for something when you can have Claude help you craft it?
  • Build custom agents. Not generic examples from documentation, actual agents designed for your stack, your conventions, your workflow. A testing agent that catches your specific mistakes. A review agent that knows your coding standards.
  • Create skills and hooks. One command to run entire workflows. Event-driven automation that runs pre-commit, post-build, on file changes, without you thinking about it.
  • Put it all together. By week 4, you'll have a complete Claude Code system: CLAUDE.md, path-based instructions, custom agents, reusable skills, and hooks that tie it all together.

Again: I'm using Claude Code as the tooling here as it's the most popular, but I can adapt to what you need.

By the time you're done, you won't just understand these features, you'll have built them for your actual codebase.

Hokule'a, the surfing cat. If he can, you can.

You might be thinking: Rob, this sounds interesting, but what if I'm not advanced enough? Let me be clear: this program is for experienced developers. If you've been writing code professionally for 5+ years and you're comfortable in the terminal, you're ready.

You don't need prior AI expertise - that's what we're building.

What This Program Is

You'll be working with me, directly, on a weekly basis for 4 weeks straight. You'll have unlimited chat access (Signal, WhatsApp, Slack, emailโ€”whatever works) with a promise to get back to you within a few hours during work hours.

We'll meet for at least two hours every week: one 1:1 session where we work through your specific challenges, and one optional group session with other participants where we share wins, troubleshoot problems, and learn from each other's codebases.

The main thing we'll do is get you from "AI-curious" to "AI-fluent." Not fluent in the sense of knowing trivia about LLMs, but fluent in the sense of having a working system that makes you measurably more productive. The kind of fluency where you don't think about the tools anymore, you just use them.

At the end of the 4 weeks, you will either have a complete Claude Code system you actually use, or you'll have your money back. I shoot straight and am not in this for the money, but I'll be honest: if I can help developers like you get past this hump, that would be wonderful.

Ryan Dahl, the creator of Node.js

What This Program Is Not

I'm not here to blow sunshine about AI. I'm not going to tell you that Claude Code will 10x your productivity overnight or that you'll never write code again. That's marketing nonsense and we both know it.

I don't have magic prompts that unlock hidden capabilities. There's no secret formula. What I have is a systematic approach to using these tools effectively, built from years of working on AI coding tools and training thousands of developers.

This program is about you: your codebase, your workflow, your specific challenges. The only person who can make it a success is you, which is why I'm very, very selective with this program.

If you're still interested, read on...

The Agenda, Week by Week

Your Calendar

Week 1: Build Something Real

No lectures. No "let's start with the fundamentals." We're building a custom agent on day one.

Before we meet, I'll have you send me a quick overview of your codebase and your biggest pain points. When we sit down together, we'll pick the most annoying part of your workflow and build an agent that handles it.

Maybe it's a testing agent that knows your test patterns and catches the mistakes you always make. Maybe it's a review agent that checks for your coding standards before you commit. Maybe it's something specific to your domain that nobody else has built.

By the end of week 1, you'll have a working agent that solves a real problem. Not a tutorial example. Something you'll actually use tomorrow.

Week 2: Automate the Tedious Stuff

Now that you've seen what's possible, we expand.

Skills turn complex workflows into single commands. Deployment scripts, documentation generation, data migrations, code scaffolding. Whatever you do repeatedly, we'll turn it into a skill you can run with one command.

Hooks make automation invisible. Pre-commit checks that run your review agent automatically. Post-build triggers that update documentation. File watchers that regenerate tests when code changes.

By the end of week 2, you'll have at least two skills and one hook pipeline running in your actual workflow. The tedious stuff you used to do manually? It happens automatically now.

Week 3: Understand Why It Works

You've built real tools. Now let's understand the machine.

This is where we dig into how Claude Code actually works: the system prompt, CLAUDE.md, context management, token economics. This stuff would have been boring in week 1. Now that you've built agents and skills, it clicks immediately.

We'll also master metaprompting: having Claude write your prompts, instructions, and configuration. Why did your week 1 agent work so well? Because I helped you write good instructions. Now you'll learn to do that yourself, and have Claude help you iterate.

By the end of week 3, you'll understand why your tools work, and you'll have the metaprompting skills to build new ones without my help.

Week 4: Refine and Systematize

The final week is about polish and integration.

We'll review everything you've built: agents, skills, hooks, CLAUDE.md, path-based instructions. We'll identify what's working, what needs refinement, and what's missing.

Then we'll put it all together into a coherent system. Not a pile of disconnected tools, but an integrated workflow where everything talks to everything else.

We'll also talk about maintenance and evolution: how to iterate on your system as your needs change, how to onboard teammates if that's relevant, and how to stay current as Claude Code itself evolves.

By the end of week 4, you'll have a complete Claude Code system that you actually use every day. And you'll know how to keep improving it.

Who Am I and Why Do I Think I Can Help You?

Because I love this stuff. Not just AI tools, but the process of helping people figure things out. There's a moment that happens when something clicks for someone, when a concept that was fuzzy suddenly becomes clear, and their lives are changed.

This is the reason I do what I do. Connecting, listening, understanding, and problem-solving.

This focus is why I created This Developer's Life with Scott Hanselman years ago. It wasn't a technical podcast. It was about the human side of being a developer: getting fired, dealing with pressure, finding motivation, facing criticism. I wanted to connect with people, to hear their stories, letting the audience know they weren't alone. The idea worked great, and the podcast sat on top of Apple Podcast's technology category for a very long time.

I have learned a lot, and I love to share. I spent 3 years on the VS Code team at Microsoft working on Copilot strategy. I wrote The Imposter's Handbook, which has helped over 20,000 developers fill in their CS knowledge gaps. I created A Curious Moon, a sci-fi novel that teaches PostgreSQL. I co-founded Tekpub, which was acquired by Pluralsight in 2013.

But here's the thing: none of that matters if I can't help you.

I've worked with thousands of developers over the years. The ones who struggle with AI tools aren't lacking intelligence or experience. They're usually missing a mental model, or they haven't seen the right example at the right time, or nobody's taken the time to understand their specific situation.

That's what I'm here for. I want to understand your codebase, your workflow, your frustrations. I want to know what you've tried and why it didn't work. I want to build something with you that actually solves your problems, not generic solutions that work great in demos and fall apart in real life.

I'm not an influencer trying to build my follower count and I'm not going to tell you AI is going to replace developers or that you need to panic about your career. I'm going to listen, explain things clearly, and help you build something that works.

The Application

As I said above: I'm being selective here, I need to know that I can help you. If you're still coming to grips with AI and its influence on programming, let's wait a bit. This program won't work for you if I have to spend time trying to convince you that things are changing.

If, however, you're ready to embrace the inevitable and you want to get in front of this process, this program could be for you

We start with you copying/pasting the text below and letting me know a little bit about yourself. If I feel I can help you, we'll set up a time to chat face to face. If it works for both of us, we start.


Tell me a little about yourself.

(Your background, how long you've been coding, what your current role looks like, what languages/frameworks you work with.)

What's your current experience with AI coding tools?

(Which tools have you tried? What's worked, what hasn't? Be honestโ€”this isn't a test.)

What do you struggle with most right now?

(Getting useful results? Understanding how to configure things? Knowing what's even possible? Something else?)

If we were sitting down 4 weeks from now and you felt completely different about AI tools, what would that look like?

(What would change? How would you work differently?)

Why does this matter to you now?

(Why not wait another year, or just keep muddling through? Why now?)

This program is a $5000 investment. Are you in a position where that feels doable?

This pricing is for the "Founders Cohort", meaning that I've reduced it for the first few 1:1 rounds I do. This isn't about squeezing you, I just want to be realistic about the time commitment and cost.

Anything else you'd like me to know?


Questions?

I'm happy to answer any questions you have, but know that right now this program is already reduced as I get it off the ground, and I'm not in a financial position to go further. In the future I hope to open up pricing parity and other options for people who need help.