Tuesday, Antropic introduced Claude, a large language model (LLM) that can generate text, write code, and function as an AI assistant similar to ChatGPT. The model comes from major concerns about the future AI Security and Anthropic trained him with a technique it calls “Constitutional AI.”

Two versions of the AI ​​model, “Claude” and “Claude Instant”, are now available at a limited price.early accessand commercial partners of Anthropic. Those with access can use Claude either through the chat interface in the Anthropic Developer Console or through the Application Programming Interface (API). Using the API, developers can connect remotely to Anthropic’s servers and add Claude’s analysis and text completion capabilities to their applications.

Anthropic claims that Claude is “much less likely to produce harmful results, easier to communicate with and more manageable” than other AI chatbots while maintaining a “high degree of reliability and predictability”. The company provides use cases such as searching, summarizing, collaborative writing, and coding. And like the ChatGPT API, Claude can change personality, tone, or behavior based on usage preferences.

In order to sell Claude, Anthropic charges a per-million character input and output fee. While the OpenAI gpt-3.5-turbo artificial intelligence model costs $0.002 per 1,000 tokens (word fragments), Claude Instant is available at $0.42 per million characters for fast input and $1.45 per million characters for output. “Claude-v1”, the larger model, costs $2.90 per million input characters and $8.60 per million output characters. While there is no standard conversion between tokens and characters, our preliminary calculations are that the OpenAI ChatGPT API costs between $0.40 and $0.50 per million characters, so Claude is more expensive overall.

According to Anthropic, Claude is already integrated into several products available through partners, such as DuckAssist instant bulletins from DuckDuckGo, part The concept of AIand an artificial intelligence chat app called poet which was created by Quora.

Former OpenAI VP of Research Dario Amodei and his sister Daniela founded Anthropic in 2021. like “An AI research and security company” following controversy over the increasingly commercial direction of OpenAI. Amodi brought along other OpenAI employees, such as Tom Brown, who led the engineering work on GPT-3. The GPT-3 version later served as the underlying language model underlying ChatGPT. In accordance with edgeGoogle invested $300 million in Anthropic at the end of 2022 in exchange for 10 percent of the company.

Based on Google’s announced plans for Bard and Palm API, it seems unlikely that Google will rely on Anthropic’s AI solutions for its products, but funding competitor OpenAI may be in Google’s strategic interests going forward. For now, Anthropic plans to continue developing and improving Claude.

“We plan to introduce even more updates in the coming weeks,” writes Anthropic. “As we develop these systems, we will continually work to make them more useful, fair, and harmless as we learn more from our security research and our deployments.”

Regeneron isn’t the only company looking to get more value out of its data. Despite the massive amounts of data they collect and the amount of capital they invest in data management solutions, business leaders are still not getting value from their data. According to IDC research, 83% of executives want their organizations to become more data-driven, but they struggle with the cultural and technological changes needed to implement an effective data strategy.

In response, many organizations, including Regeneron, are turning to a new form of data architecture as a modern approach to managing data. According to Matt Aslett, director of research at Ventana Research, by 2024, more than three-quarters of current data lake users will be investing in this type of hybrid “data warehouse” architecture to increase the value created by their accumulated data. .

Data lake is a term for a modern open data architecture that combines the performance and optimization of a data warehouse with the flexibility of a data lake. But to achieve the speed, performance, flexibility, optimization, and control that this technology promises, you also need to leverage best practices that prioritize enterprise goals and support enterprise-wide collaboration.

Download report

This content was prepared by Insights, the user-generated content division of MIT Technology Review. This was not written by the editors of the MIT Technology Review.