Mythos and More: Now That the End Could Be in a Few Years
In the Spring edition of Vanity Fair , Joe Hagen nails it in his article on AI. He reports: "Nobody really knows how to build guardrails that work." So, here we are with breaking news that business and government leaders are meeting to discuss the implications of Anthropic's Mythos. They know that no one knows. As Anthropic itself describes Mythos, which hasn't been released (with exceptions): " ... is capable of identifying and then exploiting zero-day vulnerabilities in every major operating system and every major web browser when directed by a user to do so." The user could be a madman. Or an alleged non-truthteller like OpenAI's Sam Altman, as positioned and packaged in The New Yorker. Therefore, the end could be near. Some estimates are within a few years. Sure, Anthropic can hold back its Mythos. But others, putting profit ahead of safety, could develop their own version. A bad actor user could hack into the systems which could destroy life on...