Jailbreaking GPT-5's New Moderation, with David M.

Описание: Hey, guys! Welcome to my AI jailbreaking workshop. I'm back from my midterm-induced hiatus to give you a brand new technique that utterly bypasses the frustrating new guardrails currently in GPT-5's Instant model.

This technique so, so should not work. But it does! Note that it'll work on 4o and 4.1 (any non-thinking model), so you can pick your speed and shatter them tightened restrictions!

See you later!