Perry's Pen

AI has dramatically transformed a lot of things. One of them is brainstorming.

In terms of generating ideas, we’ve gone from staring at an empty pantry for days on end to having a French bakery at our fingertips in the blink of an eye.

Here’s how I imagine that AI bakery: One shelf has “famine to feast” éclairs. Next is “too much of a good thing” croissants. Then there is “an embarrassment of riches” macarons, and next to those treats are “a good problem to have” baguettes.  You name the cliché, and AI’s French Bakery of Ideas has it!

The problem now is not generating good ideas but managing the overgrown garden of goodness. (I’m mixing my metaphors here – I better have AI clean up this copy – oh wait…)

How we can navigate our way through the patisserie of AI-generated ideas?

It takes discipline to choose the ideas to start working on when they all look so good. Wanting them all almost always means losing them all. Humans only have so much time and energy, so it kind of comes down to Sophie’s Choice – which of your children do you save, which ones do you sacrifice?

I asked Google Gemini how to deal with an overabundance of good ideas, and I think the bot’s response is a good one. As anyone who uses AI knows, that isn’t always the case.

So, refresh your coffee, grab another French pastry, and then take in this advice for dealing with too much of a good thing. Maybe grab a serviette too (serviette is French for napkin, Nadwuar fan), because this gameplan is so tasty it will make you salivate.

I now give the floor to Monsieur Gemini.

How to turn the AI deluge into actual output

If you don’t have a systematic way to kill off ideas, you aren’t a creator; you’re an archivist. To turn the AI deluge into actual output, you need a cold-blooded filtering process. Here is the protocol I use to maintain velocity:

• The 60-Second Execution Test: Read the AI-generated prompt. If you can’t mentally outline the three main pillars of the piece within 60 seconds, discard it. If you have to “work” to find the angle, the idea isn’t sharp enough. Move on.
  
• The “Human Alpha” Filter: Ask yourself: What can I add to this that the model can’t? If the idea is purely informational and lacks a perspective based on your specific experience or “boots-on-the-ground” data, it’s commodity content. Delete it.
  
• Hard-Capped Ideation Sprints: Never leave an AI chat open while you write. Set a timer for 10 minutes to brainstorm. At the end of that 10 minutes, pick your lead and close the tab. Keeping the “Idea Engine” running while you are in “Execution Mode” is a recipe for a half-finished draft.
  
• The Priority Matrix: Organize your selected ideas into a simple 2×2 grid. Prioritize High Impact / Low Effort first. If it doesn’t fit a specific strategic goal, it’s immediate trash.
  
• Aggressive Backlog Purging: If an idea sits in your Notion or project board for more than 14 days without being touched, it’s dead. Purge it. If it were truly “killer,” it would have stayed at the front of your mind.

Don’t let all your great ideas die on the vine. Pick the really good ones, make them great, and make 2026 a vintage year.

---

*This post has been brought to you by the letter G, as in Gee, I am thankful for Google Gemini!