Suprmind’s neat trick: five AI models talk through the same question together....

https://domgodwinson38.gumroad.com/p/how-to-use-ai-disagreement-to-find-the-exact-part-that-is-wrong

Suprmind’s neat trick: five AI models talk through the same question together. They spot errors and push back on each other’s answers, solving tricky issues like the five-tabs problem where AIs get confused switching context