Kris Puckett – Neuma Course Premium

Here’s What You Get:

For designers, writers, and founders who never learned to code. Your first build is a personal AI operating system that runs your day. What you keep is bigger: the confidence to build whatever you think of next, in files you own. No CS degree, just you and a blinking cursor.

• 20 modules: 9 foundation, 9 advanced, 2 bonus
• ✓Starter kit: 15 template files
• ✓Expert files + advanced templates
• ✓Advanced track: agents, orchestration, custom MCP servers, ship a real iOS app
• ✓Lifetime access, future updates included

01How It Works:

Watch. Build. Keep.

Self-paced video course. Build a real system on your machine. Keep everything forever.

// 01

Video Lessons

~6 hours of focused video across the foundations, with advanced videos rolling out. Each module demos the real system first, then walks you through building your version.

20 modules, Self-paced

// 02

Hands-On Building

Every module ends with you building something real. A personal operating system that runs on your machine. Markdown files and Claude Code.

Starter kit included

// 03

Own Everything

Files on your machine, no subscription, no platform lock-in. The system is yours, and so is everything you build after it.

Lifetime access

02.What You’ll Build

Your first build:

a personal OS

The personal OS is how you learn the skill. It’s a real system you’ll actually use, and building it teaches you everything: how to give AI persistent context, how to connect it to your tools, how to make it bring you things before you ask. By the time you finish, the OS is running on your machine and the loop that built it is running in your head.

Websites, apps, dashboards, client tools. Same loop, different surface.

03.Methodology:

Not a prompt course. A way of building.

Most AI courses hand you prompts. This one teaches you how to think about problems, describe what you want, and build real things through conversation. The personal OS is your first project. The mental models are what you keep forever.

// Typical AI Course

• ✕“Summarize this document”
• ✕Copy-paste prompt templates
• ✕Tips and tricks that expire in weeks
• ✕Starts fresh every conversation
• ✕You learn to use one tool

// This Course

• →Frame problems as conversations
• →Build systems that compound
• →Mental models that transfer
• →Context that persists across months
• →Build anything you can describe

20 modules. Your first build. Then anything you want.

~6 hours of focused content. Each module teaches you a concept, demos it live, then walks you through building your own version. The personal OS is the project. The frameworks are yours to keep.

Claude Code from Zero

For people who’ve never opened a terminal

~35 min

Install Claude Code, open a folder, and build a working dashboard by talking. No syntax, no setup rituals, no pretending you know what a for-loop is. By the end you’ll have the same blinking cursor everyone else has, except it won’t feel like a threat anymore.

You’re not learning a tool. You’re learning a loop: Describe, Review, Redirect. The tools will change. The loop doesn’t.

// when you’re done:

• ✓Installed Claude Code and ran your first session from a real project folder
• ✓Built a working HTML dashboard through conversation, no code written by hand
• ✓Ran the Describe / Review / Redirect loop on your own feedback and watched it land

Your Workspace: The Foundation

One file turns a brilliant stranger into a collaborator who knows you

~28 min

CLAUDE.md is a markdown file Claude reads at the start of every session. You’ll build yours by conversation, not a template, and end up with a workspace where the AI knows your role, your constraints, your communication style, and what it should never do. Same prompt, night-and-day response.

The skill isn’t in how you ask. It’s in what Claude already knows before you ask.

// when you’re done:

• ✓Created a project folder with your own CLAUDE.md, five sections deep
• ✓Built the file through conversation instead of copying a template
• ✓Watched the same prompt return a generic answer, then a personal one, once context was loaded

Identity: Teaching Your AI Who It Is

The shift from assistant to thought partner

~38 min

Your CLAUDE.md gets three new sections: a role definition, a values layer, and a coaching stance. This is where Claude stops executing and starts pushing back. Ask it a question you’re avoiding and it’ll name it instead of answering it.

I keep the agency. You keep me honest. That’s the whole relationship in two sentences.

// when you’re done:

• ✓Defined Claude’s role and posture, not just its tasks
• ✓Wrote values that act as tiebreakers when your own preferences conflict
• ✓Tested the coaching stance on a real decision and got the question under the question

Memory: Making It Remember

The layer that makes every session build on the last

~40 min

Three layers of memory: auto-memory Claude writes itself, daily logs you keep like a ship’s log, and a knowledge base where patterns get promoted to permanent understanding. Raw data flows upward and refines into the rules you actually live by. After three months, the system remembers things you’ve already forgotten.

The write is the memory. Not the conversation. Not the context window. The file.

// when you’re done:

• ✓Turned auto-memory on and learned to edit what Claude writes about you
• ✓Started logging daily entries as ground truth, not reflection
• ✓Promoted your first real pattern to a knowledge base page

The Daily Briefing

Orientation over reaction. Start the day with your own context

~45 min

One document, generated every morning, that synthesizes your health, your calendar, your tasks, and your energy into a single view of the day ahead. You read it before email. You close it out at night. Every close-out makes tomorrow’s briefing sharper.

A calendar app shows you what’s scheduled. The briefing tells you what you’re actually capable of today.

// when you’re done:

• ✓Generated your first daily briefing in plain language, matched to how you actually feel
• ✓Ran the evening close-out so tomorrow’s briefing has something to read
• ✓Connected at least one data source so the briefing stops requiring manual input

Energy-Based Task Management

Work organized by capacity, not arbitrary priorities

~32 min

Map your real energy windows for a week, categorize tasks by the kind of brain they need, and teach Claude the rules in your CLAUDE.md. Then capture tasks in plain language and let the system sort them into the right time of day. When plans break, it tells you what to cut, not what to add.

Importance without capacity is guilt. Capacity-aware importance is a plan you can actually follow.

// when you’re done:

• ✓Mapped your own energy windows from a week of honest observation
• ✓Wrote energy-based task rules into your CLAUDE.md
• ✓Captured tasks in natural language and had Claude place them in the right window

The Coaching Layer

A mirror with memory. The part that catches what you forget

~40 min

Build a small library of honest self-knowledge: your baseline, your patterns, your guardrails, the questions you’re sitting with. Then teach Claude to check every frustration and every “I’ll skip it tonight” against those files. It stops being a chatbot and starts being the friend who says “you’re doing that thing again.”

A friend might notice your patterns occasionally. This system sees them daily. It doesn’t forget. It doesn’t get tired of reminding you.

// when you’re done:

• ✓Wrote an honest baseline: who you are at your best and under stress
• ✓Set guardrails in advance so tired-you can’t override clear-headed-you
• ✓Watched Claude flag a real pattern in the moment and name it directly

Making It Yours

Your system evolves. Here’s how to let it

~35 min

Skills encode the things you explain every session. Hooks automate the guardrails you never want to think about. You’ll build one of each, not five, and learn to listen to friction as the signal for what to build next. Derive your system from your own repetition, absence, and annoyance, not from mine.

Copy my setup and you get a system shaped for my brain. Derive the structure and build what’s actually true for you.

// when you’re done:

• ✓Built one real skill from something you kept re-explaining
• ✓Installed one hook as a safety net, not an optimization
• ✓Practiced killing features after three weeks instead of three days

Going Always-On

Bonus. From a tool you pick up to an environment you inhabit

~50 min

Three layers of ambient: scheduled triggers on a clock, event-driven triggers that respond to changes, and the rich context you’ve built all course. Most people only need the first and the third. Your briefing shows up before you do, your close-out runs when you’re ready, and the system earns the right to speak by knowing when to shut up.

Build on principles. The products will die. The principles won’t.

Optional. Everything in Modules 0–7 works with just Claude Code and a folder of markdown files on any laptop.