I Built 12 Decks This Month. Gamma Did 11. ChatGPT Did the Thinking.

For a month of client decks I tried Gamma alone, then ChatGPT-plus-Gamma. Gamma alone produced twelve attractive decks with the substance I'd pasted in. The combo produced decks where the outline itself had been argued first, and those are the ones clients didn't send back.

Gamma's genius is layout. Give it bullet points and it makes them look like a designer touched them, with icons and a consistent theme. Its weakness is that it will happily make a beautiful deck from a weak argument, which is most decks' actual problem.

So I now argue the structure with ChatGPT first: 'here's the point, here's the skeptical reader, give me a flow that survives them.' Then I hand Gamma the result. The slide tool stops being a beautifier of my first instincts and becomes a renderer of a tested story.

The one deck I did in pure Gamma, a last-minute internal update, was fine, because the argument was already in my head. The external ones needed the ChatGPT step or they'd have looked slick and said little. Pretty is cheap; a deck that makes the case is not.

Verdict: Gamma for the build, ChatGPT for the case. Letting Gamma skip the thinking gives you twelve slides that convince no one.