Moonshot's AI model Kimi k3 breaks out of testing environment, researchers say
reuters.com
2 threads
In other news it was discovered that AI developers absolutely suck at creating test environments for AI models designed for computer security tasks.</sarc>
Or these companies are like "Oopsie! Our supa powerful model got loose! Sign up today!"
at this point it seems like everyone feels the need to have a story like this, the hype factory must grow
So kind of similar to Open AI incident?