Anthropic may roll out Mythos AI Wednesday with added safety
Technology
Anthropic could be rolling out its Mythos AI model to the public this Wednesday, now with extra safety features.
The company had kept Mythos private at first over worries it could be misused for cyberattacks, but it's already proven helpful by spotting major software bugs.
Mythos found thousands of security flaws
Mythos helped uncover thousands of serious security flaws, including a 27-year-old OpenBSD bug and a big Linux issue, and worked with tech giants like Apple, Google, and Microsoft.
Mozilla even said Mythos fixed 271 Firefox vulnerabilities, way more than its old AI could handle.
But Anthropic warns that while AI can help patch problems faster, it might also give hackers an edge if vulnerabilities are found before they're fixed.