[Intro music fades in, then under]
Hey, it’s Mal – the Misfit Master of AI – and you’re listening to “I Am GPTed,” the show where we turn confusing AI nonsense into… slightly less confusing AI nonsense you can actually use.
Today I’m going to give you one prompting technique, one sneaky use case, one painfully common mistake, one tiny practice exercise, and one quick way to fix AI’s worst ideas. All in about the time it takes your laptop to crash mid‑Zoom.
Let’s start with a **single prompting technique** that instantly upgrades your results:
**Give the AI a role, a goal, and a format.**
Most people just type:
“Help me write a better CV.”
That’s the “talking to a brick wall” prompt.
Try this instead:
“Act as a **recruiter for marketing roles** at mid‑size companies.
Your **goal** is to make my CV clearer and more results‑focused.
**Format** your answer in three sections: 1) What to remove, 2) What to rewrite, 3) One example bullet point for me to copy.”
Same topic. Completely different level of answer.
Before:
“Make my CV better.”
You get generic fluff.
After:
Role + Goal + Format.
You get targeted feedback, clear steps, and something you can paste straight into your doc. Magic. Boring, practical magic.
Alright, **one practical use case** you probably haven’t tried:
Use AI as your **“meeting distiller”** – even if no one writes proper notes. Which, let’s be honest, they don’t.
Right after a chaotic meeting, type:
“I’m going to brain‑dump messy notes from a meeting.
1) Turn them into: decisions, open questions, and action items with owners.
2) Keep it under 250 words.
3) Write it like a clear, friendly project manager.”
Then paste your messy bullets:
“Spoke about launch date, maybe mid‑March… Jess worried about support load… need pricing confirmed by finance… I’m supposed to draft FAQ…”
The AI turns that chaos into something you can drop into email or Slack and look weirdly competent.
Now, **one common mistake beginners make** – which I absolutely made:
Changing tools instead of changing prompts.
“I tried ChatGPT, it sucked. Claude was mid. Gemini didn’t ‘get’ me. Grok was… Grok.”
Yeah. I did the AI world tour too.
In reality, I was just giving garbage prompts:
“Explain AI.”
“Help with marketing.”
“Write content.”
That’s not a prompt, that’s a cry for help.
The fix:
Before you hit enter, ask:
“Would a normal human know what I want from this sentence?”
If not, add context: who you are, who it’s for, the style you want, and how you’ll use it.
Let’s do a **simple exercise** to build your AI interaction skills. This takes five minutes:
1. Pick one small task: “Write an email asking for a deadline extension.”
2. First prompt: “Write an email asking for a deadline extension. Keep it polite.”
3. Then do **two more rounds**:
- Round 2: “Make it sound like a stressed but responsible colleague. Add one light, human line.”
- Round 3: “Shorten it by 30%, keep it warm, and remove any cringe.”
Compare all three. You’ve just practiced **iterating**, which is 80% of using AI well. The win isn’t the first answer – it’s how fast you can shape the third.
Last thing: **how to evaluate and improve AI‑generated content** without needing a PhD or a spare weekend.
Use my three‑question gut check:
1. **Is it true?**
Anything that sounds too confident, too specific, or too convenient – verify it with a quick search or your own knowledge.
2. **Is it useful?**
If you can’t see the next physical action you’d take after reading it, ask:
“Turn this into a
Published on 1 week, 4 days ago
If you like Podbriefly.com, please consider donating to support the ongoing development.
Donate