Anthropic's latest AI, Mythos, is so powerful that they aren't releasing it to the public. They are just making it available to other software companies because it can easily find security bugs. In fact, it found security holes, some over 20 years old, in every operating system and web browser. They are hoping that companies can use it to patch flaws before other AIs are powerful enough to help hackers.
An interesting this is that the government may not like Anthropic from sharing it with companies, since the government collects flaws so that they can hack enemy systems.
© 2025 Praveen Puri
