It’s Not Them, It’s You

Working with AI—whether it’s generating images, crafting blog posts, or answering life’s weirdest questions—can sometimes feel like trying to explain your vision to a slightly confused intern. Sometimes the AI nails it; other times, it gives you a surreal mess that leaves you staring at your screen wondering, “How did we get here?”

But here’s the truth: it’s not them, it’s you.

Now, before you close this tab in frustration, let me clarify: it’s not your fault in the blame-y sense. It’s just that prompting an AI isn’t magic—it’s a skill. And like any skill, it takes clarity, patience, and a little trial and error to get it right.

The good news? This skill isn’t just for tech nerds or AI enthusiasts—it’s for everyone. Whether you’re an AI artist trying to generate the perfect image, a writer wrangling content from a chatbot, or just someone messing around with AI tools, getting better at prompting makes the whole process smoother—and way more fun.

So let’s break it down. We’ll cover the basics, explore why patience is key, and finish with one golden rule: pick something fun and dive deep.

1. Garbage In, Garbage Out (GIGO): The Golden Rule of Prompting

When you ask an AI something vague like, “Make me an image that looks nice,” you’re going to get an equally vague result. Why? Because AI isn’t psychic—it’s predictive. It’s trained to respond to patterns in your prompts, and if your instructions are unclear, it’ll fill in the blanks… often poorly.

How to Fix It: Be Specific, Be Intentional

• Instead of: “Make me an image that looks cool.”

• Try: “Create an image of a futuristic city skyline at sunset, with flying cars and neon lights reflecting off glass skyscrapers.”

The more context and direction you provide, the more the AI can align with your vision.

2. Context is King (Or Queen, Or Supreme Overlord)

Imagine walking up to someone and saying, “Make me something great.” Without context, they might give you a sandwich, a painting, or perform an interpretive dance. AI works the same way—it thrives on context.

How to Provide Context:

• Purpose: What are you using this for? (e.g., a podcast intro, a social media post, a blog).

• Tone/Style: Casual? Professional? Playful?

• Constraints: Word count, genre, or specific details.

For example:

• “Create an image of a cozy cabin in a snowy forest, with smoke rising from the chimney and warm lights glowing from the windows.”

Context isn’t just helpful—it’s everything.

3. Patience is (Still) a Virtue

AI prompting isn’t a one-and-done thing. It’s an iterative process, like tuning an instrument or tweaking a recipe. The first draft might not hit the mark, and that’s okay.

The Iterative Process:

Iteration is like having a conversation with your AI where every reply gets a little smarter and a little closer to what you want. In short, iteration means refining your prompts step-by-step until you hit that sweet spot.

1. Start broad.

2. Analyze the AI’s response.

3. Refine your prompt based on what worked and what didn’t.

4. Keep repeating until you hit gold.

For example:

• First Prompt: “Make me a landscape image.”

• Refined Prompt: “Make me an image of a mountain landscape during sunrise.”

• Further Refined: “Make me an image of a snow-capped mountain landscape during sunrise, with soft pink and orange tones in the sky.”

Each step sharpens your results.

4. Know the AI’s Limits (For Now)

AI tools are powerful, but they’re not perfect. They might misunderstand nuance, get facts wrong, or offer oddly generic answers. Sometimes, they experience what’s known as 'AI hallucinations'—when the system generates false or nonsensical information that sounds plausible. Biases, often inherited from the training data, can also skew responses in unexpected ways.

To ensure accuracy, it’s a good idea to fact-check AI outputs using tools like another LLM, Perplexity, or Stanford University’s Co-Storm before publishing. However, don’t rely solely on AI tools for verification—traditional research sources, such as academic papers, trusted websites, and primary sources, are equally important for cross-referencing information.

Additionally, keep in mind that different AI models have varying capabilities and limitations. A prompt that works seamlessly with one model might produce less accurate results with another.

When you hit a roadblock, don’t quit—adjust your approach and try again.

5. Pick Your Playground: One Fun Thing to Rule Them All

Here’s where things get fun—and, honestly, where you’ll grow the most as a prompter.

The Big Idea:

Pick one fun thing to focus on. It could be AI art, joke-writing, recipe generation, or storytelling. Whatever it is, dive deep and explore how far you can push your prompts in that area.

Why This Works:

When you focus on one playful area, you reduce the pressure. It’s easier to experiment, make mistakes, and learn.

Some Ideas to Get You Started:

• AI Art: “Create an image of a cat DJ at a neon-lit party.”

• Comedy Gold: “Write a dad joke about a robot in therapy.”

• Recipe Roulette: “Create a vegan pasta recipe using only three ingredients.”

• Storytelling: “Write a sci-fi story where humans discover an alien planet made of candy.”

Play, iterate, refine, and see where it takes you.

Final Thoughts: AI is a Collaboration, Not a Transaction

Start small, pick a playful project, and give yourself room to learn. Over time, you’ll notice your prompts becoming sharper, your results more accurate, and your frustration levels way lower.

Because in the end, it’s true: it’s not them, it’s you.