I wrote this because I keep seeing “AI idea validator” tools give founders a confident score from nothing more than a pitch.

My argument is that AI is useful upstream for sharpening the idea, but it cannot be the verdict on demand. The interesting signal is whether strangers do anything when the offer is put in front of them.

Curious where people here draw the line between useful AI feedback and fake validation.

The only thing AI is better at than hallucinating is glazing (sycophancy).