The biggest AI problem no one is talking about

Hey guys,

I have some thoughts I’d like to share about the crazy time we’re living in.

We are living through a "step-function change" in what’s possible with AI

The common advice? “Learn AI or get left behind.”

So what do we do? we open 49 browser tabs. We watch 3-hour "State of AI" deep dives. We become the world’s leading experts at passing a multiple-choice test on which LLM is currently winning the benchmark wars.

But we aren't actually building anything. This is the new face of Procrastination. It’s sophisticated, it looks like "research," and it’s killing your growth.

Even the highest performers on the planet feel this. Tyler Denk, the founder of beehiiv, the company that makes this email software I’m using now, recently admitted that despite his fascination with AI and his desire to 100x his output, he’s been on the sidelines:

I doubt you work harder than Tyler does. The guy wakes up at 5:30 am and stays at his desk (including meals) until 9 pm every day. If a guy running a massive tech company feels "overwhelmed" and "on the sidelines," what chance do the rest of us have?

The Knowledge Trap vs. The Experiment

In my 22 years of podcasting and content creation, I’ve learned one thing: The world doesn't pay you for what you know; it pays you for what you ship.

The guy in the Facebook group who won't start his podcast because he is worried his dog will bark in the background is the same as the entrepreneur who won't use AI because they haven't "mastered the prompt engineering" yet.

They are both stuck in some modern form of analysis paralysis.

How to Break the Loop: Episode Negative One

The cure for AI overwhelm isn't more tutorials. It’s Episode -1.

In podcasting, Episode -1 is the recording you make just to see how you look and sound. No niche. No $400 mic. No logo. You don’t even have to publish it. You just need to break the seal.

I just did this with my brother to show you how "un-fancy" it should be:

I used Podsplice because it’s a tool we built to solve this exact friction—it’s browser-based and does the heavy lifting for you—but the tool doesn't matter. The action does.

We used AI to build Podsplice. But in order to build it, we had to actually start using AI instead of just reading about it.

I wrote about “Episode -1” here on my Substack

Stop “Learning”. Start Using

If you want to grow your business or your channel in 2026, stop trying to be the most "knowledgeable" person in the room. Be the person with the most "dirty" experiments under their belt.

The "Dog Test" Challenge: Stop reading about AI and plans to grow your online presence for the next 5 minutes. Don’t make any excuses equivalent to your fears of a barking dog in the background.

Start DOING and experimenting with what you know you should. Don’t worry if it is perfect in the beginning. It won’t be.

If you would like to promote something, now is easier and better than ever to start a podcast.

You don’t have to, but I highly recommend Podsplice. It’s what my brother and I built to use ourselves.

Whatever it is that you want to do, try to create your own version of Episode -1 instead of just reading and daydreaming about it.