As the world grappled with the consequences of Anthropic’s Cloud Mythos model, a new challenge emerged. A new AI startup, DepthFirst, has now claimed that its AI model has discovered vulnerabilities that even Cloud Mythos was unable to find. The special thing is that Cloud Mythos is considered so powerful that it can hack any software or operating system. Now, DepthFirst claims its own model is more powerful. This development has raised new concerns regarding cyber security.
**What did the DepthFirst AI model do?**
According to reports, DepthFirst has launched an initiative called “Open Defense” to help developers detect vulnerabilities in their software. Its purpose was to fix these shortcomings before they fell into the hands of hackers. The CEO of DepthFirst claims that their AI model has discovered more vulnerabilities than Mythos. The CEO further says that this model is very economical, and is capable of delivering better results at a much lower cost than other models.
**AI model exposes old vulnerabilities**
According to media reports, the model has detected a vulnerability in NGINX – one of the most used web servers in the world. This particular flaw has existed since 2008, and has been affecting the Internet for years. Additionally, DepthFirst claims to have discovered security vulnerabilities in both Linux and Google Chrome; Google has acknowledged these vulnerabilities and has since released security patches. The company also claims that it has discovered 12 vulnerabilities in the FFmpeg platform, which runs major services like YouTube, Netflix, Spotify, Instagram, and Facebook.
**Why are concerns being raised about such models?**
Some people may wonder: If these AI models are helping to detect security vulnerabilities, why are they considered a threat? In fact, the answer lies in an important question: who controls these models? If these models fall into the wrong hands, they can become a threat to the entire world. As a result, entire online systems may be at risk. This is why Anthropic has not released its Cloud Mythos model to the general public.
