Anthropic, a renowned artificial intelligence (AI) lab, has made a significant announcement with the launch of Claude 2, the latest rival to ChatGPT. The improved model, a successor to the original Claude, is set to be accessible to the general public in both the United Kingdom and the United States.
Following the concerns raised on the societal risk that AI poses, Claude-2 is built on ‘Constitutional AI.’ This innovative technology aims to create a human-like conversational experience, making the model sound less like a machine. Claude-2 has undergone extensive training to respond appropriately to questions while politely declining harmful or inappropriate inquiries. According to internal testing conducted by Anthropic, the chatbot is capable of successfully avoiding harmful questions twice as often as other existing models.
Unlike its predecessor, users can access Claude-2 through an application programming interface (API) as well as a user-friendly public-facing website. Residents of the United Kingdom and the United States can simply visit the company’s website and sign up to engage with the chatbot.
Anthropic stressed that Claude-2 is superior in terms of reasoning, coding, and mathematical capabilities. Demonstrating its prowess, the chatbot achieved an impressive 76.5% score on the bar exam, surpassing Claude 1.3 by 3.5 percentage points.
Claude-2 —a better alternative to ChatGPT
Anthropic aims to replicate the remarkable advancements made by OpenAI with ChatGPT and GPT-4 through the development of Claude-2.
Compared to GPT-4, Claude-2 offers a more cost-effective solution. Moreover, this advanced chatbot surpasses Claude-1 in its ability to perform coding and qualitative reasoning tasks. With Claude-2, users can experience quicker and more convenient access to the chatbot, as it is readily available on both the company’s website and APIs.
Ethan Mollick, one of the testers for Claude-2, hails it as not only the most humane AI personality but also a highly knowledgeable chatbot when it comes to document handling. Claude-2 showcases its proficiency in managing lengthy texts by effectively condensing them into concise formats.
Anthropic, the creator of Claude-2, asserts that users can input a remarkable volume of up to 100,000 tokens in a single interaction—an impressive increase compared to ChatGPT’s limit of 8000 tokens. Despite being in the developmental stage, Anthropic firmly believes that Claude-2 offers a more reliable and secure option in terms of performance, making it a commendable alternative to ChatGPT.
As development progresses, it holds the promise of becoming an even more powerful and dependable tool for various applications.
Photo Credit: Possessed Photography
