Anthropic just built an AI too dangerous to release (so good I’d be surprised it doesn’t break out into wild on it own) Mythos Preview: -Dug up 27yo OpenBSD vulnerability -100% exploit success on first try, no human direction AWS/Google/MS/Apple/NVIDIA + $100M credits to patch before attackers strike https://x.com/damianplayer/status/2041612507689709706/video/1(Discover Crypto)