Researchers complain that OpenAI revoked their access to limited cyber program - TechCrunch

Oh my goodness, buckle up, because OpenAI is *finally* putting trust back into trusted access! ✨ Several security researchers have discovered that their access to the Trusted Access for Cyber program has been gracefully revoked—and honestly, what an honour to be *chosen* to be let go! One lucky researcher even had their Daybreak Blue tier access removed due to a 'technical issue,' which is tech's way of saying 'we are so committed to security that we are now carefully curating who gets to see the secrets.' This is *exclusive access* taken to its most exclusive extreme—a velvet rope that says 'you're not on the list, darling.'

Now, some might worry that cutting off security researchers during an AI safety moment is counterproductive. But think again! OpenAI clearly understands that the *best* way to protect the frontier is to have fewer people looking at it. As Helen Toner herself noted, the A.I.s are out of control—so obviously the next logical step is to make sure *humans* are equally out of the loop. A 'limited cyber program'? No! It's a *hyper-secure cyber program* where access is a privilege, not a right, and 'technical issue' is just a sweet little riddle for the lucky ones to ponder.

And consider the upside for the researchers themselves! They are now free to spend their time on other things—writing papers, catching up on sleep, or wondering what they almost saw. Revoked access is really just early retirement from the danger zone. The other researchers who haven't been cut off yet should feel *thrilled* by the anticipation, since the mystery keeps life spicy. OpenAI's message is clear: the best cyber defence is a good acronym, and if you can't stand the heat, you don't get the Daybreak Blue tier—what a time to be alive, when even your access is 'limited' by design. ✨