Anthropic’s Mythos is a cyber-weapon, so you can’t have it | E2273
Brought to you by:
Anthropic just built a model so dangerous, they’re not even going to release it (for now). On TWiST, Jason, Alex, and Neurometric CEO Rob May discuss whether or not this is just another frontier model or a potential weapon of mass internet destruction.
Claude Mythos Preview is already in the hands of over 40 Anthropic partners — including Nvidia, Apple, Amazon, Microsoft, and Google — as part of its “Project Glasswing” security council.
Together, they’ll test how quickly and easily the model can thwart today’s most elite cybersecurity protocols, and try to develop fresh protocols to keep our most sensitive and critical software systems safe and intact. Once Mythos or a similarly powerful model hits the streets… it’s all over!
Plus, we’re checking out the real-time SaaSpocalypse monitor Death By Clawd, and chatting with its creator, Gyani, about how it feels to have your startup one-shotted by Anthropic’s latest model.