Anthropic built an AI that can break software at the level of a skilled human security researcher. Then gave it to the defenders before anyone else could use it. Claude Mythos Preview. Vulnerability identification at the expert human level. Exploit chaining that turns minor bugs into catastrophic attack vectors. The capability that changes the security calculus for every piece of critical software on earth. Project Glasswing is the strategic response. AWS. Apple. Google. Microsoft. NVIDIA. Linux Foundation. All of them now using this capability to proactively harden OpenBSD, Linux, and the digital infrastructure civilization runs on. The gap between offensive AI capability and defensive AI deployment is the most consequential security gap in history. Project Glasswing is the bet that defenders can close it before the attackers realize it exists!