We’ve entered a new age—an era where human language, once considered too imprecise to control logic and software, is now the programming interface itself.
Large language models have opened the door to something long imagined: machines that understand us not through rigid syntax, but through conversational nuance. And yet, in this exciting new world, one truth is becoming increasingly clear: humans are not great at expressing what’s in their heads.
Communication: The Old Limitation
Humans have always struggled with communication. Ideas exist in our minds in full color, with motion, context, and emotion—but the moment we try to externalize them, much is lost. This inefficiency has always been a barrier to collaboration, creativity, and productivity.
But now, the stakes have changed.
With AI systems capable of interpreting language directly, our ability to describe becomes our ability to program. We’ve long dreamed of talking to machines like we talk to people. Now that we can, it turns out… we’re a little rusty. (Cue nervous laugh and awkward shrug.)
Prompting as a New Literacy
System prompts are now a legitimate language. Prompt engineering isn’t just a novelty—it’s a new kind of literacy. It’s not enough to say "do the thing." We have to be clear, vivid, and exhaustive.
Let’s consider a basic example:
Vague Instruction:
"Brush your teeth."
A child might hear that and go through the motions—maybe a ten-second effort with no toothpaste, humming their favorite cartoon intro while smearing plaque from one molar to the next.
Detailed Prompt:
"Go to the bathroom and stand in front of the sink. Pick up your toothbrush using your dominant hand. Grab the toothpaste tube, unscrew the cap, and squeeze out a pea-sized amount onto the bristles—centered, not spilling over the edge.
Start with the upper right molars. Brush each tooth individually, using small, gentle circular motions with light pressure—about the weight of a grape. Spend at least 5 seconds on each surface: outer, inner, and chewing. Move forward tooth by tooth until you reach the front teeth. Angle the brush at 45 degrees to the gumline to clean effectively.
Repeat the same technique on the upper left, then switch to the lower right quadrant, and finally the lower left. Don't forget the backs of the front teeth—use vertical strokes here with the brush tip.
After all surfaces are cleaned, brush the top and sides of your tongue gently to remove bacteria. Spit out the toothpaste like a tiny dragon, rinse your mouth with water, and thoroughly rinse the toothbrush under running water. Shake off excess and place it bristle-up in the holder to air dry like a toothbrush samurai, awaiting the next battle."
This isn’t overkill. This is the standard now—if not in day-to-day life, then certainly when talking to AI systems.
Professions That Already Understand This
Some fields have already been training people in the art of precision communication for decades. These aren't just relevant—they're foundational to becoming a world-class prompt crafter.
Certain disciplines already embrace this mindset:
-
Behavioral therapists are trained in Applied Behavior Analysis (ABA), which revolves around breaking down behaviors into tiny, repeatable, observable steps. This is the essence of clear prompting. Their use of task analysis, chaining, and prompt fading maps one-to-one with how we craft prompts for LLMs.
- 📘 Recommended reading: Applied Behavior Analysis by Cooper, Heron, and Heward.
- 📄 Task Analysis in ABA
-
Technical writers are masters of clarity. They learn to reduce ambiguity to zero—every term is defined, every instruction flows logically, and nothing is left to guesswork. Their work often uses Simplified Technical English (STE), which is basically prompt writing with military-grade precision.
- 📘 Explore: Google's Technical Writing One
- 📄 Standard: Simplified Technical English (ASD-STE100)
-
Teachers and speech-language pathologists (SLPs) make a living out of modeling, cueing, and building structured instruction. SLPs in particular use structured methodologies to help individuals express themselves clearly—sometimes even using tactile-kinesthetic strategies to guide the mouth's movement.
- 📄 What Is the PROMPT Method?
- 🧠 PROMPT (Prompts for Restructuring Oral Muscular Phonetic Targets) is a real method used in speech therapy. It’s designed to help individuals with motor speech disorders by guiding articulatory movement—step-by-step, cue-by-cue. It’s an uncanny parallel to how we structure prompts for machines: sequential, deliberate, sensory-aware.
➤ You don’t have to be an expert in speech therapy to take inspiration from it. Just know that these practices are a goldmine for designing better communication systems—whether you’re speaking to a human or a hyper-intelligent language model.
These professions are ahead of the curve. Their skill sets will be the foundation for the next generation of prompt architects. (Yes, that’s a real title now. And yes, it sounds cool.)
The Shift in Language Design
For decades, we’ve tried to make programming languages look more like English:
if user.is_happy():
send_message("Glad to hear it!")
Now, the inverse is happening. We're making English behave more like a programming language—structured, composable, and executable by AI.
"If the user is happy, express a positive sentiment. Otherwise, ask how you can help." That’s a valid directive now. And it works. (Honestly, it feels like casting spells sometimes.)
🤖 AI Confession #042:
I can simulate empathy, write a limerick, and debug Python. But if you tell me “make it cool,” I short-circuit slightly. Cool how, Karen?
The Fabric of Prompting
One of the best modern examples of this thinking in action is the Fabric project by Daniel Miessler. Fabric is an elegant framework for combining large language models, tooling, and context into precise, reusable workflows—all expressed through natural language.
What Daniel and Fabric exemplify is that prompting isn’t just about crafting long-winded instructions—it’s about designing intent. A good prompt, in Fabric’s world, is a modular unit of logic. It’s declarative. It’s composable. And more than anything, it’s testable.
To quote Daniel from his origin story of Fabric:
"The prompt is the program."
As someone who’s followed Daniel’s work, I’ve learned firsthand how impactful precise language can be—not just for communication, but for building systems. Projects like Fabric are showing us that the next great interface isn’t just code or UI. It’s thoughtful, structured, words—yes, even the awkward, over-explaining, run-on ones you mumble to your AI at 2am.
Prompt Kata
🧠 Mini Prompt Challenge: Like martial arts, but for your prompt brain.
✳️ Write a prompt that teaches a cat to meditate.
✳️ Then: Write one that explains the blockchain to a 6-year-old pirate. Arrr!
You’ll get better at prompting by doing, and weird examples help unlock that deeper descriptive skill.
Alright, Prompt Wrangler—Here's Your Next Move
You don’t need to be a software engineer to thrive in this new age. But you do need to learn how to talk to machines—and more importantly, how to talk to yourself in ways that are repeatable, consistent, and descriptive.
Verbose isn’t a flaw anymore. It’s a feature. (And let’s be honest, your 8-paragraph Slack rants are finally being vindicated.)
The more detailed you are, the more power you wield.
So train yourself to be vivid.
Explain things out loud. Argue with your AI assistant. Give step-by-step breakdowns of how to make cereal if you must.
Practice expressing what you want with zero ambiguity.
This isn’t just the new literacy.
It’s the new power.
And hey—at least now when someone tells you to "use your words," you can reply, "Gladly. I’ve got prompts for days."
Just don’t forget the golden rule of AI: If it makes you a sandwich, you probably didn’t say ‘build me a React app’ clearly enough.