Your Prompt
Interview me about my work and identify one AI exercise I can practice every day for the next week.
The exercise should take less than 10 minutes, build useful AI skills, and be repeatable.
Focus on consistency over productivity.
Once you've suggested the exercise, explain what skill it is actually training.
Against The Grain
Friends ask me all the time: "What's the best instrument to learn first?"
Piano, for that classic, foundational music education?
Guitar, for the cool factor and the solos?
Drums, for the excuse to bang out your aggression on an instrument all day?
My answer is always "Whichever instrument you're going to pick up every single day."
Sure, some instruments teach more transferable skills than others.
But eventually, a musician learns how to glean something from any instrument. My acoustic playing improved dramatically when I started learning bass. Over the past twenty years, I've added in electric guitar, piano, hand percussion, mandolin, and more. Each one taught me something I could bring back to the others.
None of that matters - you'll never get to that point - if your first choice ends up collecting dust in the corner after three weeks.
Practice beats optimization.
It's easy to make the same mistake with AI.
A few weeks ago, I wrote that AI platforms should be considered like streaming services: You shouldn't lock yourself into one ecosystem and pretend it's the only one worth using.
I do believe that.
But I should have added a caveat...if you're just getting started, I'd give almost the opposite advice.
Pick one and stick with it.
Not forever. Just long enough to build a habit.
The biggest mistake isn't choosing the wrong platform. It's bouncing so frequently between platforms you never learn to use any of them well.
ChatGPT this week. Claude next week. Gemini on the watchlist. Every new release feels like the answer must be back over there, where I just left.
The FOMO is real.
The value isn't in knowing which model won the latest benchmark. It's learning how to ask better questions, provide useful context, challenge an answer, and recognize when something sounds impressive but isn't actually helpful.
Once you've built a few hundred reps on that muscle, then you can go explore. Try other platforms, compare outputs. Learn their strengths and quirks.
At that point, you're a musician picking up a second instrument.
But too many beginners spend their time shopping for “the right instrument” instead of practicing the one they own.
The goal is to get past the point of “Should I try AI for this?”
And get to the point where you're asking, “How do I use AI for this?”
How do you get to Carnegie Hall?
Practice
