Anthropic is set to launch a two-way voice mode for its AI assistant, Claude, allowing users to engage in real-time conversations. Dario Amodei, CEO of Anthropic, announced this feature during a discussion at the World Economic Forum in Davos. The new voice mode aims to enhance user interaction by enabling Claude to remember user preferences and history. Additionally, Amodei indicated that Anthropic plans to release smarter AI models in the coming months, with the focus on enterprise applications rather than consumer use. Meanwhile, a Chinese startup, DeepSeek, has made headlines by releasing a new open-source model, DeepSeek R1, which reportedly outperforms OpenAI's GPT-4o and Claude 3.5-Sonnet on various benchmarks, including math tasks. DeepSeek R1 is noted for being significantly cheaper and more efficient than its competitors, running locally and offering advanced reasoning capabilities. The rapid advancements from DeepSeek highlight the shifting landscape in AI, particularly in response to U.S. tech restrictions, which some argue have inadvertently spurred innovation in China.