Anthropic, a startup founded by former OpenAI employees and backed by Google, has unveiled Claude, a new AI chat assistant that experts already see as a major rival to OpenAI’s ChatGPT. Claude can be accessed through a chat interface and handle various conversational and text processing tasks, including summarizing, searching, co-writing, Q&A, coding, and more.
Anthropic has taken a more pragmatic approach in developing its chatbot, making it “easier to navigate” and “much less likely to produce harmful output” than ChatGPT. The company claims Claude is committed to building trustworthy, interpretable, and maneuverable AI systems, using “Constitutional AI” to create “Benefits.” According to the Anthropic paper detailing Constitutional AI, the process has stages of supervised learning and reinforcement learning, resulting in non-intrusive AI assistants that engage by explaining objections to harmful queries.
Anthropic has been working with partners like Notion, Quora, and DuckDuckGo to increase its functionality.
Anthropic is feeling pressure from investors to recoup hundreds of millions of dollars spent on AI technology. The company received a $580 million tranche from investors, including ousted FTX founder Sam Bankman-Fried, Caroline Ellison, Jim McClave, Nishad Singh, Jaan Tallinn, and the Center for Emerging Risk Research. Google also invested $300 million in Anthropic and acquired a 10% stake in the startup, making Google Cloud its “preferred cloud provider.” The two companies also agreed to co-develop an AI computing system.
Anthropic plans to offer Claude via an API to support businesses and nonprofits. Currently, two types of Claude and Claude Instant have been released, and pricing for API access has yet to be revealed (Sign up for early access). With the backing of Google and a growing list of partners, Anthropic’s Claude looks set to give OpenAI’s ChatGPT a run for its money.