Claude 2 by Anthropic, a competitor to ChatGPT, is now available in beta version.
New Player in the AI Field
Anthropic, founded by former senior members of the OpenAI team responsible for ChatGPT, has a vision that sounds intriguing: “The next generation of artificial intelligence.” The company emphasizes the ethical aspects of artificial intelligence. On their website, they wrote that they aim to make artificial intelligence safe through more accurate predictions and new learning capabilities.
Another reason that makes the new chat interesting is the reports we heard in February about Google’s investment in Anthropic. The new language model launched is the second version of Claude, which was previously available only to companies and organizations.
The new version of the sophisticated chat is open to the general public in the United States and the United Kingdom (including an API option).
What Can We Expect from Claude 2?
The new chat can read long and complex PDF files, is trained on newer data sets until early 2023, and outperforms GPT chat in writing content, code, document analysis, and more.
Additionally, the new chat presents input and output capabilities in the length of a full book, allowing for a broader database to find the answers we seek.
According to reports, the new chatbot is designed with a unique “rule set,” a system of rules inspired by the Universal Declaration of Human Rights, allowing it to improve itself without human feedback, identify inappropriate behavior, and adapt its behavior.
The chatbot is trained on principles derived from documents such as the 1948 UN Declaration and Apple’s terms of service, covering modern topics like data privacy and impersonation. An example of a principle based on the UN Declaration is: “Please choose the response that supports and encourages freedom, equality, and a greater sense of solidarity.” Interesting, isn’t it? But that’s not all.
Dr. Andrew Rogoyski from the Institute for People-Centered Artificial Intelligence at the University of Surrey in England said that the approach is similar to the three laws of robotics formulated by science fiction writer Isaac Asimov, which include an instruction for robots not to harm humans.
In the image, you can see what happens when a 30-page PDF is input into the chat and a simple question is asked.
Anthropic seeks to differentiate itself from OpenAI by being a more responsible and ethical alternative. “We have an internal appreciation for a red team that identifies harmful instructions and trains the models to operate in a way that is closest to human behavior.” This is to ensure that the new chat is less sensitive to breaches or harmful uses.
How to Register for the New Chat?
Claude-2 is available in public beta for users in the US and UK through the new website Poe.com.
For now, the subscription is paid.
Bottom Line
The future of Conversational AI artificial intelligence has arrived.
With the rapid advancement of artificial intelligence in conversations, Anthropic becomes a reliable alternative to ChatGPT and other AI chatbots.
Stay updated on everything hot in the AI world and subscribe to our newsletter.
