This week’s prompt will talk through a piece of feedback you’ve been holding back from someone on your team.
Don’t soften it.
Drop it into your AI exactly how you’d say it if tone didn’t matter. Blunt. Direct. No hedging.
Your Prompt
Rewrite this feedback so it actually lands with a real person. Preserve the standard. Improve the delivery. Explain what would likely offend them and why.
Now don’t stop there.
Take the rewritten version and tell me:
"What part of this feedback is about clarity… and what part is about how I’m choosing to deliver it?”
Against The Grain
“AI is great because it doesn't get offended.” My buddy Byron brought this up in a presentation he gave this week.
You can be blunt. Direct. Even a little rude if you want. No awkward pause. No tone to manage. No relationship to preserve.
Just input… output… done.
And I get it. I'm a boss. And a dad. And a husband. I deal with a lot of emotions
And I'm human - so I'm also aware I have my own double-standards. I want people to be gracious in their feedback toward me, while also wishing I could just say what I need to say without tiptoeing around their feelings.
So I understand, it's can be refreshing to have a chatbot that doesn't stop working because of the tone I used.
But it’s worth asking…
Isn't it actually a good thing that humans get offended?
Remember just a couple years ago, when we were joking about the opposite with AI?
Typing “please” and “thank you” into ChatGPT… just in case it remembered us in the impending AI uprising.
Half joking… half habit.
Because we’re used to interacting with things that respond to how we say things.
But at some point, most of us stopped hedging our bets, and just started responding, "No. Stop it. You're making that up. Do it right."
(Probably around the same time we realized how sycophantic AI could be.)
We’ve learned that the system works better when we strip all the frills out.
Cleaner prompts. Faster iterations. No emotional friction.
But since this newsletter is about training us... we need to see that something subtle is happening alongside that shift.
We’re now getting hours of reps where our primary conversation partner never pushes back on how we talk to it.
No misinterpretation of your tone. No defensiveness. No clarifying how you actually meant something to come across.
You just respond with another prompt and move on.
It’s so much more efficient.
It also may be a loss.
Because when a human gets offended, it’s usually not random.
It’s pointing at something. Something we might need in the future to unlock real potential.
Maybe your expectation wasn’t clear. Maybe you challenged something they care deeply about. Maybe you were right… but delivered it in a way that made it unusable.
AI doesn’t show any of that.
And if feedback is a gift… this is one we're opting out of.
If most of your day you're interacting with systems that don’t react to you…
What happens the next time you actually have to lead?
Not delegate to a tool. Not refine a prompt.
Lead. Actual humans. Offline, IRL
Where how you say something will determine the outcome. Where deals are lost because of body language. Where buy-in matters as much as accuracy.
Where being right doesn’t count if no one follows you.
