Book Brew 167: The DREAMS Manual for Non-Slop Content

I spent hours on a lead magnet that Damien eventually described as 'polished diarrhea’; I didn’t even make it to solid turd level.

Enthusiastic. Productive-feeling. Completely misguided.

That's the most accurate description of the afternoon I spent building what I was absolutely convinced was going to be a knockout lead magnet. I was in a state of flow only those with AuDHD can understand. The document was growing and Gemini was cooperating beautifully.

Damien read it.

It was bad. Like, polished diarrhea bad. The kind of output that looks like something until someone who knows better reads it out loud and the whole thing deflates like a sad birthday balloon.

I still found a use for it.

And the reason I almost didn't has nothing to do with AI.

The Problem Wasn't the Tool

I went into that session with energy (possibly too much squirrel energy, but that’s a different story…) and no plan. I knew the general topic.

However, I didn’t follow Dalio’s 5-step process of Goals → Problems → Diagnosis → Design → Do

I did not have a defined problem, a clear end goal, or any real sense of who the thing was actually for.

AI is extraordinary at executing, but not so not great at caring whether you've thought things through first. Hand it a vague brief with good intentions and it will produce something that technically answers the question you didn't quite ask.

The output reflects the input: Garbage In, Garbage Out

And if the input starts with "I have this general idea and a free afternoon," the output is going to be exactly as focused as that sounds.

Start With the End, Then Work Backward

Cheryl Einhorn's DREAMS framework is one of the cleaner ways I've seen this problem named and solved.

The framework isn't complicated, but it requires something most of us skip when we're excited about an idea: actual thinking before we start.

• Define your problem. What specific gap does this piece of content fill? What does someone need to know, decide, or feel differently about after reading it?
• Refine your decision as a question. Turn the fuzzy goal into something answerable. "I want to make a lead magnet about AI" is a topic. "What's the single most common mistake my audience makes in their first month of using AI tools?" is a question you can actually build toward.
• Examine your motivation. Why this, why now? Is this genuinely useful for your audience or does it just feel satisfying to make?
• Activate your options. What formats could this take? What angles haven't been covered? What does your audience actually respond to? Generate the options before you commit to one.
• Manage your biases. This is the one people frequently skip. What assumptions are you bringing in that might be wrong? Do a pre-mortem: imagine the final piece flopped completely, and work backward to figure out why. You'll catch things in five minutes that would have taken weeks to discover in the wild.
• Select your path forward. Now you build. With a plan, a defined problem, and a specific person in mind.

It may feel slower than just opening a chat window and going for it But it also can be the difference between a document that Damien can hand back with a look on his face, and one that actually does what you needed it to do.

The Slop Wasn't Useless

"How you perceive and respond to challenges shapes your outcomes". - Ankur Warikoo

Despite the polished turds, that afternoon wasn't entirely wasted.

The document I made was too vague to be a lead magnet. It turned out to be a decent brainstorm and gave me raw material to build with.

AI output without a plan isn't always garbage. The mistake is treating "it came out of a productive session" as the same thing as "it's ready to publish." Those are very different things, and the gap between them is exactly where the DREAMS framework lives.

Know what you're building before you build it. Then use the tool to execute the plan.

Your audience can tell the difference. So can Damien.

Ponder This

1. Look at the last thing you made with AI that didn't land, did you have a defined problem going in, or a general idea?
2. If you ran a pre-mortem on your next AI project right now, what's the first thing you'd predict going wrong?
3. What's one piece of content you made that you thought was the thing but turned out to be raw material for the actual thing?

Books/Newsletters

• Strengthening your decision-making with Geneartive AI - Cheryl Einhorn
• Do Epic Shit - Ankur Warikoo
• Principles - Ray Dalio