Information about a new model that Anthropic, which operates in the field of artificial intelligence, was working on, emerged as a result of the company’s internal materials accidentally leaking to the internet. It was understood that the company, known for its chat bot named Claude, was testing a more advanced system than its existing models.
Features and possible effects of the new model
Anthropic states that this new artificial intelligence model it is working on is the most powerful system they have developed so far. The behavior and possible risks of the model, which has not yet been disclosed to the public and is only accessible to a limited number of priority customers, are being evaluated. In the information shared by the company, it is emphasized that there is a remarkable leap in the performance of this model and that it exceeds the previous generation systems.
In the leaked documents, the model named “Claude Mythos” is mentioned and it is pointed out that the system may have the ability to detect software vulnerabilities and exploit them. This has raised various debates in the field of cybersecurity. Although the details regarding the content and capabilities of the model have not yet been finalized, it is being discussed what risks the system may pose in terms of security.
Anthropic currently offers three different levels of artificial intelligence models called Opus, Sonnet and Haiku. These models differ in size, capability and price. The leaked materials also state that a new layer called “Capybara” is planned, which will be even more advanced and larger than Opus.
First reactions in companies and markets
With the news following the leak, there was a remarkable decline in the shares of some important companies operating in the cyber security and software industry in the USA. Share values of companies such as Palo Alto Networks, Crowdstrike and Fortinet fell by four to six points in a short time. A similar decline was seen in broad technology software indices.
In addition to temporary market fluctuations, such leaks have also had an impact on digital asset markets. Bitcoin showed a significant decline in a short time with the leak information and retreated to a level slightly lower than the peak it reached in the previous hours.
The main reason for the information leak in question is that approximately three thousand Anthropic blog content-related materials were accidentally uploaded to a public data repository. The company found that these documents included announcement drafts and internal company content that had not yet been published.
Anthropic was founded in 2021 by former OpenAI employees, and the company’s focus is on ethical AI development, secure systems, and scalable AI applications. Anthropic’s work, which closely monitors the developments in the technology sector, is followed with interest, especially by large organizations and regulatory institutions.
