Post

Hilarious to read someone characterizing the exaggeration of the severity of a discovered vulnerability as evidence of the malfeasance of AI companies and not, like, the established practice of the security industry for several decades now

25/10/26, 4:44 pm0 reposts0 likes

twotwotwo.bsky.social

@twotwotwo.bsky.social

yeah "you have a domain and a brand name for *this*??" unfortunately describes an extremely human tendency

Coda

@codahale.com

Not exactly a deep cut from the training data, no

Coda

@codahale.com

“AI companies want you to be scared” or maybe spicy autocorrect has understood the assignment in ways which might not make us feel good about ourselves