Five tricks power users type into AI that you don’t. Steal all five.
Same chatbot, same monthly bill, wildly different answers. The gap isn’t intelligence. It’s vocabulary. Here are the five commands that turn a chatbot into a consultant, an editor, a burglar and a tutor.
🤯 WOW! Five key terms can turn a $20-a-month chatbot into a $500-an-hour consultant. One is in Merriam-Webster. Another helped plan the bin Laden raid. They all work on the AI you’re using.
The Short Version: Type the command FIRST, then a colon, then paste your problem. Copy-and-paste examples below.
📖 Read time: 2 minutes
ChatGPT/Kim Komando
I need your help: Add Komando.com as a preferred source on Google
Most people ask AI a question. Power users hand it a job description. These five key terms are the job description.
One rule makes them all work: The term goes at the front of your prompt. It’s the first thing you type, then a colon, then your situation. Like addressing the envelope before you write the letter.
🧠 Three orders make it think harder
ELI5: “Explain like I’m 5.” The slang acronym was born on Reddit in 2011, now 20 million members strong. Apply it to anything that makes your eyes glaze, including your MRI report, the escrow statement, the fridge warranty fine print. Use it like this: “ELI5: [paste the confusing thing].” Want it less preschool? “ELI5, but I’m a smart adult who’s just new to this.”
Steelman: The opposite of a straw man. You’re ordering the AI to build the strongest version of an argument, especially one you disagree with. It’s the cheapest way to find out you’re wrong in private, before you find out in public. Use it like this: Start with the verb. “Steelman the case for keeping my 12-year-old car.” Then add: “Now tell me which of my objections survive.”
Pre-mortem: Cognitive scientist Gary Klein’s trick. Don’t ask “what could go wrong?” Announce the funeral in advance. Klein cites research showing this boosts your ability to name the real risks by 30%. Use it like this: “Pre-mortem: [your plan]. It’s one year later and this failed. Write the story of why.” Order the autopsy while the patient is still healthy.
🛡️ Two that protect you
Red team: Straight out of Cold War games. Before the bin Laden raid, three separate red teams tore apart the intelligence. It attacks, so you can patch. Use it like this: “Red team my home security. You’re a burglar. Where do you get in?” Point it at your passwords, your inbox, your small business.
Socratic. The one exception to the front-of-the-prompt rule, because it’s a teaching style, not a job. The AI stops lecturing and starts interrogating. You answer. It corrects. That’s actual learning. Use it like this: “Teach me [topic] using the Socratic method. One question at a time.” Hand it to the grandkid dodging homework and watch the machine turn on him.
Same chatbot. Same bill. Five commands, always up front. That’s the whole upgrade, friend.
FYI, I once looked up “obfuscate” in the dictionary. The definition was confusing and misleading. Should have said ELI5.
🎯 Send this to someone who says AI “never gives me good answers.”
📱 Text-worthy: Type “Red team my home security, you’re a burglar, where do you get in?” into ChatGPT. The Pentagon invented this. It takes 30 seconds, and it will ruin your night in the best way.