"Mythos has laid bare thousands of vulnerabilities in commonly used applications for which no patch or fix exists, prompting the San Francisco-based AI startup to form an alliance with cybersecurity specialists to bolster defenses against hacking and withhold wide distribution."
"The oldest of the vulnerabilities uncovered by Mythos dates back 27 years, and none were ostensibly noticed by their makers before being pinpointed by the AI model, according to Anthropic."
"AI models have reached a level of coding capability where they can surpass all but the most skilled humans at finding and exploiting software vulnerabilities, Anthropic said in a blog post."
"As a precaution, Anthropic has shared a version of Mythos with cybersecurity companies CrowdStrike and Palo Alto Network."
Anthropic's upcoming AI model, Claude Mythos, has identified thousands of software vulnerabilities, many of which lack existing patches. The company has decided not to release Mythos publicly, instead collaborating with cybersecurity experts to enhance defenses against potential hacking threats. The vulnerabilities uncovered include long-standing issues that had gone unnoticed by developers. Anthropic emphasizes the model's advanced capabilities in detecting subtle flaws, which could have significant implications for cybersecurity, public safety, and national security.
Read at www.theguardian.com
Unable to calculate read time
Collection
[
|
...
]