I'm BACK. And I need to tell you something...

I've been a maniac for the past few months.

In the lab.

Experimenting.

Teaching.

Building things I'm soooo excited about.

But this email list sat in the corner of my dashboard, unopened, starting to wonder if I was ever coming back.

I'll tell you what I've been doing in a second, but first, something that might sting...

You probably have a folder somewhere. Maybe in Notion. Maybe buried in your X bookmarks.

Maybe it's a $99 "prompt library" sitting in your Downloads folder next to a headshot you no longer use and a content calendar from 2019.

That folder knows you're avoiding it.

It's full of prompts you saved because they looked useful. Viral threads. LinkedIn carousels. "Game-changing" templates from people you've never heard from again.

You've used maybe five of them. The rest just sit there, getting dusty, starting to wonder if you're ever coming back.

And with prompts, saving them feels productive. You found something good. You captured it. Little hit of dopamine. Box checked. Your brain registers this as progress.

But using them? That's where it falls apart.

You paste one in. It gives you something that sounds right but like it was written for everyone. You go back to what you've always done, rewrite the whole thing and think:

So you go back to scrolling. Saving. Collecting.

The pile grows and the guilt compounds.

I'm describing this because I think you recognize yourself (I sure as hell did.) And because the fix looks different than most people expect.

More prompts won't solve this. Better prompts won't either.

What actually works is changing how you think about AI.

​

Four shifts.

I'm going to show you each one over the next four days. Real examples with things you can use today.

Tomorrow: Why that viral prompt didn't work for you (and the 60-second fix).

Stay Curious,

P.S. If you want to do something useful right now, try this. Pick ONE prompt from your graveyard and paste it into Claude or ChatGPT with this addition at the top:

I'm about to give you a prompt I saved from somewhere. Before you run it, I want you to interview me.

Ask me 5-7 questions about:

• What I'm actually trying to accomplish (not what the prompt assumes)
• Who will see the output and what they care about
• What I've already tried that didn't work
• What constraints I'm working within (time, format, tone)
• What "good" looks like for this specific situation

Then, adapt the prompt to my actual context before running it.

That interview is the difference between generic output and something you'd actually use. See what happens.

Check out my SubStack to find the...

​​​

​