Anthropic
On Tuesday, Anthropic announced Claude, a large-scale language model (LLM) that can generate text, write code, and act as an AI assistant similar to ChatGPT. The model was born out of a major concern about the safety of future AI, and Anthropic used a technique called “Constitutional AI” to train the model.
Two versions of the AI model, Claude and “Claude Instant”, are currently available to a limited “early access” group and Anthropic’s commercial partners. Users with access can use Claude through Anthropic’s developer console chat interface or application programming interface (API). The API allows developers to remotely connect to her Anthropic’s servers and add Claude’s analytics and text completion features to their apps.
Anthropic says Claude is “much less likely to produce harmful output, easier to talk to, and easier to manipulate” than other AI chatbots, while maintaining a “high degree of reliability and predictability.” claim. The company cites use cases such as searching, summarizing, co-authoring, and coding. Also, similar to ChatGPT’s API, Claude can change its personality, tone, or behavior depending on usage preferences.
Promotional video introducing Anthropic Claude.
To sell Claude, Anthropic charges royalties for every million characters input/output. OpenAI’s gpt-3.5-turbo AI model is $0.002 per 1,000 tokens (word fragments), while Claude Instant is available at $0.42 per million characters for prompt input and $1.45 per million characters for output. The larger model, Claude-v1, costs $2.90 per million characters of input and $8.60 per million characters of output. There is no standard conversion between tokens and characters, but our rough estimate is that OpenAI’s ChatGPT API is around $0.40 to $0.50 per million characters, so Claude is generally more expensive.
Arstecnica
According to Anthropic, Claude is already integrated into several products available through partners, including DuckAssist instant summaries for DuckDuckGo, part of Notion AI, and an AI chat app called Poe made by Quora.
Former OpenAI VP of research Dario Amodei and his sister Daniela founded Anthropic in 2021. As After disagreements about OpenAI’s increasingly commercial direction, it became known as an “AI safety and research company”. I came. A revised version of GPT-3, he later served as the underlying language model at the heart of ChatGPT. According to The Verge, Google has invested his $300 million in Anthropic in exchange for his 10% of the company in late 2022.
Based on Google’s announced plans for the Bard and PaLM APIs, it seems unlikely that Google will rely on Anthropic for AI solutions in its own products, but funding OpenAI’s rivals is likely to be Google’s next step. It may be in your strategic best interest. For now, Anthropic plans to continue developing and improving Claude.
“We will be rolling out more updates in the coming weeks,” wrote Anthropic. “As we develop these systems, we will continue to work to make them more useful, honest, and harmless, while learning more from safety research and deployment. “
