Anthropic, an AI startup founded by ex-OpenAI executives, has released its second-generation AI chatbot called Claude 2. The new model is an improvement over its predecessor, with better performance in multiple-choice exams, coding tests, and math problem-solving. Claude 2 was trained on more recent data and can analyze a large amount of text, making it more capable. However, it still has limitations, including the potential for generating irrelevant or harmful responses. Anthropic has implemented constitutional AI to guide Claude 2’s behavior and make it more understandable and adjustable. The company is cautious about using the model in high-stakes situations or applications involving health and well-being. Anthropic plans to make Claude 2 available through its own API and Amazon’s generative AI hosting platform. The company continues to offer a faster and cheaper version of Claude called Claude Instant. Anthropic has raised $1.45 billion in funding and estimates it will need $5 billion over the next two years to fully develop its chatbot.
