Stop Trying New AI Tools
Your Favorite AI Tool Already Knows You Better Than You Think
A member of the Claude for Creators group cracked me up last week.
We were going through the same prompt across two tools — Claude on one screen, ChatGPT on the other — and the drafts came back wildly different.
Claude’s was technically fine. Clean. Structured. Kinda generic.
Then she said, almost apologetically:
“I’ll be honest, I’m gonna throw this into Marvin. He already knows me.”
Marvin is what she calls her ChatGPT. She’s been talking to him for eighteen months.
And here’s the thing—Marvin’s draft was better.
Not because ChatGPT is smarter than Claude. They’re roughly the same model class these days.
Marvin was better because she had eighteen months of context loaded into him. Her voice. Her offers. Her past newsletters. The way she actually talks to her people.
Claude was starting from scratch.
That’s the thing nobody tells you about AI tools.
The value isn’t in the model. It’s in the moat you build around it.
Every time you bounce to a shinier tool, you reset the moat to zero.
I’m not saying never switch. I switched to Claude last year for a few specific reasons and it took me weeks to get all the context and memory right. It’s easier now, but it still takes work.
But if you’ve been hopping between three different AI tools every couple of weeks, you don’t have an AI problem, you have a commitment problem.
Pick one and talk to it for six months. Watch what happens.
Matt “meet Marvin” Ragland
P.S. The prompt I use to turn coaching call transcripts into newsletter angles is the single highest-leverage thing I’ve built into my AI tool this year. Comment CALLS I’ll send it over.
CALLS
CALLS