Anthropic Claude Mythos Has Helped Mozilla Find 271 Vulnerabilities on Firefox 150

Anthropic's Claude Mythos Preview is already proving to be one of the most powerful tools for finding security problems. Mozilla has shared that its use of the Claude Mythos has helped it discovered as many as 271 vulnerabilities on Firefox 150.

Thanks to Anthropic's latest Project Glasswing development, Mozilla was able to find the said vulnerabilities, and the company has revealed that they have already patched these discoveries in the browser.

Anthropic's Claude Mythos Helps Discover Firefox Vulnerabilities

Mozilla announced in its latest blog post that it used the new Claude Mythos Preview via its continued collaboration with Anthropic on its latest version of the Firefox browser to test its capabilities.

The result led to the discovery of as many as 271 vulnerabilities in the initial evaluation using the technology. Despite it being on preview, the Claude Mythos was able to identify the risks present on the browser.

Engadget revealed that this latest post from Mozilla shared a third-party look into Anthropic's technology, applied in real-world settings. Moreover, Mozilla shared their real experience, saying that "So far, we've found no category or complexity of vulnerability that humans can find that this model can't."

While Mozilla held the Claude Mythos Preview in such high regard and allowed it to discover hundreds of vulnerabilities, the company noted that the AI was only able to find threats that human researchers can also find on their own when given enough time and resources.

That said, Mozilla's latest comments are a testament to Anthropic's claims about the Project Glasswing initiative, one that focuses on helping clients find issues or problems that lie within their tech that have remained unchecked.

Mozilla Patches Firefox 150's 271 Vulnerabilities

According to Mozilla, the company has recently worked on patching all the vulnerabilities with the latest Firefox 150 release.

The latest unveiling of Project Glasswing is a massive achievement for Anthropic as the company was able to reassert its commitment to AI security in the industry, following its unwillingness to drop safeguards to secure a US government contract.