Forget "Prompt Engineering." You Need AI Fluency
Announcing a 5-part series on the real skills that turn generative AI from a novelty toy into an augmented part of your mind.
TL;DR:
Most of us use AI like a search engine, “asking” for answers and getting frustrated by mediocre, generic, or wrong replies. We get stuck in a “correction spiral,” trying to fix the AI’s mistakes. We’ve been told “prompt engineering” is the secret, but that’s a lie.
The real solution is “AI Fluency”—a shift from being an “asker” to an “architect.” This is how you become truly “AI augmented,” directing the AI as a tool rather than begging it for answers.
To teach this, I’m launching “Towards AI Fluency,” a free 5-part email series. It’s a new operating manual for working with AI, focused on five core skills:
The Architect’s Skill: How to decompose any complex problem into small, “AI-sized” tasks before you ever write a prompt.
The Navigator’s Skill: How to recognize when an AI chat is “polluted” and why restarting is often faster than iterating.
The Expert’s Skill: How to use your own domain knowledge as your most critical shield to detect and defeat confident AI hallucinations.
The Curator’s Skill: How to provide high-signal, curated context (the “gold”) instead of “data dumping” (the “noise”) to get precise results.
The Collaborator’s Skill: How to master the “DIY-AI-DIY” workflow, where you act as the Architect, the AI as your Intern, and you as the final Editor-in-Chief.
This series will teach you to stop “chatting” and start directing. Subscribe to the newsletter to get all five articles and turn AI from a novelty into a powerful, professional collaborator.
Towards AI Fluency
You’ve felt it, haven’t you?
That flicker of frustration. You’re using the most powerful creative and analytical tools ever built, and yet, half the time, the results are... meh.
You get generic, high-school-essay answers. You get code that looks right but doesn’t run. You get politely confident, completely wrong “facts” that you only catch because of a gut feeling.
We’ve all been there. We’re stuck in the “correction spiral,” typing “No, that’s not what I meant” and “Please follow the instructions” until we want to throw our laptops out the window.
Here’s the problem: We’ve been told that the key to AI is “prompt engineering.” That if we just find the magic combination of “magic words”—act as, you are an expert, I will tip you $200—we’ll unlock the machine’s genius.
This is a lie.
Prompting is a tiny, final step in a much larger process. Relying on “prompt hacks” is like trying to become a Michelin-star chef by only learning how to plate the food. You’re missing the entire kitchen.
Your biggest mistake with AI is your first one: you treat it like a search engine or a digital oracle. You ask it for answers.
True fluency begins before you ever write a prompt. It starts with a total shift in your workflow. It requires you to stop being an asker and start being an architect.
This is the path to becoming AI Augmented.
The “AI Augmented” Professional Isn’t an Asker. They’re a Director.
The “AI Augmented” professional doesn’t see AI as a replacement for their skills. They see it as a force multiplier—an incredibly powerful, endlessly patient, and slightly clueless intern that, when directed properly, can elevate their work to a new level.
This isn’t a futuristic fantasy. This is a practical, learnable skillset. It’s the difference between asking, “Help me plan a marketing campaign,” and directing the AI through a 15-step process that results in a campaign you can actually launch.
But this new, augmented workflow demands new skills.
It’s not about typing. It’s about thinking. It’s about decomposition, navigation, expert verification, curation, and collaboration. It’s about leveraging your hard-won human expertise as the most critical part of the system.
In a world flooded with generic AI content, your personal domain knowledge—the stuff you actually know—is your greatest shield and your most valuable asset.
This is why I’m launching “Towards AI Fluency,” a free 5-part series designed to teach you these five core, transferable skills. This isn’t a list of “10 Fun Prompts.” This is a new operating manual for your brain.
We’re going to move past the hype and give you the framework to become the architect, not just the asker.
Here’s the roadmap.
Article 1: The Architect’s Skill — Decomposing Problems Before You Prompt
The root of all AI frustration is a poorly defined problem.
We approach a powerful Large Language Model with a vague, massive goal: “Write a blog post about my product,” “Plan my vacation,” or “Refactor this entire codebase.” Then we get a vague, massive, and mediocre result. We blame the AI. We are wrong.
The problem is that we’ve given a “human-sized” problem to the AI. A fluent user knows they must first act as an Architect, breaking that complex goal into a series of clear, specific, “AI-sized” tasks.
In this first article, we will stop “chatting” and start architecting.
You’ll learn the “Project Brief” Method, a simple, offline process for defining your final output and identifying the core building blocks needed to get there. We’ll contrast a vague goal like “Help me plan a marketing campaign” with a proper, decomposed brief:
“Analyze the tone of my last 3 newsletters.”
“Generate 10 audience personas based on this target demographic...”
“Draft 5 Twitter hooks for Persona A about this specific topic.”
“Draft a 4-part email sequence based on Hook #3.”
This decomposition is the single most important skill in AI-augmented work. It transforms you from a passenger to the driver, giving you a concrete plan of action. This roadmap is your new workflow.
Article 2: The Navigator’s Skill — When to Iterate and When to Restart
You’re three replies deep. You’ve corrected the AI twice. It still isn’t getting it. It keeps reverting to a previous mistake or ignoring a key constraint.
You are stuck in the “Correction Spiral,” and you’re about to lose your mind.
A novice user will keep fighting. They’ll keep correcting, tweaking, and clarifying, sinking more and more time into a failing conversation. A fluent Navigator knows when a chat is lost.
In this article, you’ll learn about the “Polluted Context Window.” We’ll explore how an AI’s short-term memory gets contaminated by your own vague prompts and corrections, leading it to latch onto flawed assumptions.
I’ll give you a simple heuristic—the “Three-Strike Rule”—to help you recognize when a chat is “polluted” beyond repair. You’ll learn why restarting is not failure; it’s an efficiency tactic.
But we won’t just hit “new chat.” We’ll master the “Restart & Refine” strategy. You’ll learn to diagnose why the last chat failed, harvest any good ideas it produced, and craft a single, superior first prompt that incorporates all your learnings from the start. This skill alone will save you hours of frustration.
Article 3: The Expert’s Skill — Spotting Confident Hallucinations
An AI can generate a beautiful, persuasive, and utterly false statement with unshakeable confidence.
It will invent statistics. It will cite non-existent legal cases. It will create fake quotes from real people. And it will do it all in a smooth, authoritative tone that makes it sound true.
In the age of generative text, your personal expertise is no longer just a tool for creation; it’s your most critical shield. AI fluency isn’t about trusting the AI’s output. It’s about knowing when not to.
In this article, we’ll tackle the most dangerous part of AI: hallucinations. We’ll explore why they happen (hint: AIs are “plausibility engines,” not “truth engines”) and the tell-tale signs of a fabrication.
More importantly, we’ll lean into the idea that to be an augmented professional, you must first be a “master of some.” You cannot be an expert in everything, but you must be the expert in your own domain.
We’ll learn the “Master of Some” Verification Method. You’ll stop asking, “Is that true?” and start challenging: “Provide the source for that statistic.” You’ll learn to trust your gut. Your hard-earned experience is infinitely more reliable than the AI’s plausibility. This skill is how you maintain quality and credibility in a world of AI-generated noise.
Article 4: The Curator’s Skill — Providing High-Signal Context
“Garbage in, garbage out.” This has never been more true than with AI.
We’re told to give AI “more context,” so we dump entire 50-page documents, messy spreadsheets, or 10,000-word articles into the prompt window and hope for the best. This is the equivalent of trying to drink from a firehose.
More context is not better. Better context is better.
The fluent AI user isn’t a data dumper; they are a master Curator. They know that their job is to maximize the “signal” (key facts, examples, constraints) and minimize the “noise” (irrelevant details, redundant information).
In this article, you’ll learn to provide curated, high-signal context. We’ll break “context” down into three actionable categories:
Instructional Context (The ‘How’): Clear rules, like “Use Markdown” or “The tone is professional.”
Exemplar Context (The ‘What’): A specific, high-quality example of the output you want.
Knowledge Context (The ‘Source’): The raw information the AI needs, curated by you.
We’ll deploy the “Researcher’s Workflow,” a process where you do the human work first. You find the source, you extract the key quotes and data points, and you feed the AI only that “gold.” This is how you stop getting generic summaries and start getting surgically precise analysis.
Article 5: The Collaborator’s Skill — Mastering the DIY-AI-DIY Workflow
The fear is that AI will take your job. The reality is that it changes your job.
This final article integrates all four previous skills into a single, powerful, repeatable workflow: DIY-AI-DIY (Do It Yourself - AI - Do It Yourself).
This framework redefines your relationship with AI. You are not replacing yourself. You are hiring an intern. This loop ensures your expertise remains the most important part of the process.
We’ll break down the three roles you now play:
You as the Architect (DIY): This is where you start. You use your judgment to Decompose the Problem (Skill 1) and Curate the Context (Skill 4). You define the entire strategy before the AI writes a single word.
AI as the Intern (AI): This is the execution phase. The AI’s job is to do the “first draft” grunt work on the hyper-specific tasks you created. It brainstorms, writes boilerplate code, or summarizes your curated notes, and here you spot the Polluted Context Window (Skill 2).
You as the Editor-in-Chief (DIY): This is where you finish. You take the AI’s raw output and apply your critical judgment. You Spot Hallucinations (Skill 3), refine the language, inject your unique insights, and integrate the pieces into a final, polished product that is uniquely yours.
This workflow is the future of professional, creative, and technical work. It combines the speed of AI with the quality, judgment, and insight of a human expert. It is the very definition of being AI Augmented.
Your New Workflow Begins Now
These five skills—Architect, Navigator, Expert, Curator, and Collaborator—are the foundation of true AI fluency. They are not hacks. They are a durable, transferable framework for thinking and working alongside these new tools.
This series is your roadmap. It’s for the developer who wants to build faster, the marketer who wants to create better campaigns, the researcher who needs to synthesize information, and any professional who is ready to move from “tinkering with AI” to building a true cognitive partnership.
Don’t miss the first article.
To get all five articles in the “Towards AI Fluency” series delivered directly to your inbox, subscribe to the newsletter below.
The journey starts next week. It’s time to stop asking and start directing.