Use AI to Organize the Idea, Not to Come Up With It

About this episode

Krysta wasn't going to make an AI episode this week — not every announcement needs its 15 minutes. Then the same news showed up in three or four Instagram carousels from accounts she doesn't even follow: Anthropic is building a machine-readable watermark into future Claude models, and this time it covers text. Not a logo across your caption — closer to the "AI info" banner already sitting under people's posts. Her knee-jerk reaction was probably yours: I used it like Google, I wrote the rest, why does that get a label?

Two Very Different Things Are About to Get the Same Label
• You take a selfie you were physically in, then use Canva's magic eraser to pull a trash can out of the background. Technically AI.
• You sit in your apartment in New York and generate a photo of yourself on a beach in Italy you've never been to. Also technically AI — and not remotely the same creative act.
• None of this is new. Adobe started the Content Authenticity Initiative in 2019 and called content credentials a nutrition label for digital media. What's new is text.

The Better Question Is What You Asked It to Do
• Krysta's process for this episode: brain dump into ChatGPT, ask it to pull the facts (the way we all used to use Google), then let it reorganize a 45-minute ramble into 20 usable minutes.
• Uses she stands behind: cleaning up grammar, summarizing hours of your own research or client calls, spotting the pattern in what you keep circling back to.
• The use that bothers her: "give me five controversial opinions that'll get engagement," pick one, have it write the script and the caption, film a talking head, call it your perspective.

If Everyone Has the Same Tools, Your Inputs Are the Advantage
• Nobody calls their wedding photos fake because the colors got adjusted. We start caring when the edit deceives us about what actually happened.
• "AI generated" still carries a verdict — you didn't make it, you cheated, maybe you don't even know your stuff. That may not survive the label becoming normal.
• Same tools, same prompts, so the tech isn't the advantage anymore. What's left is what you noticed this week, what pissed you off, what you changed your mind about.
• Reach for it every time an opinion is half-baked and you get less practice forming hot takes. Same reason your feed sounds identical.

Fix Tip of the Week
Run the wearable check-in in reverse. Before you open the Oura app, the Whoop, MyFitnessPal — anything handing you a number about your own body — decide how recovered you actually feel. Then look.

This Week's Exercise
Pull up the last three things you posted and name exactly what AI did in each one: research, grammar, reorganizing your own notes, or the whole idea. Then DM Krysta which bucket each one landed in — and whether an "AI info" label would bother you.

No Such Thing of the Week
There's no such thing as all AI use being exactly the same.

If this one landed, go back to No Such Thing as One Take EP3, where Krysta and Paige get into why AI has read everything and lived nothing.

Follow Krysta:

⁠@thekrystahuber⁠⁠

⁠⁠@thefitnessfyx⁠⁠

⁠@thespreadmktg ⁠