Anthropic shifted its focus away from chatbots at the end of last year, according to head of science Jared Kaplan. In a recent interview, Kaplan explained that the company is now prioritizing more complex tasks like research and programming. This new direction is also reflected in the latest Claude 4 models, which are designed with agent-based applications in mind. Kaplan acknowledges that tackling these kinds of advanced tasks comes with a higher risk of models behaving unpredictably. As a result, Anthropic is putting a strong emphasis on minimizing those risks. Programming remains a particular strength for the company’s models—a core capability that has made Anthropic especially popular with developers. The approach appears to be paying off: Anthropic's annual revenue has doubled, reaching two billion dollars.
