(Reuters) – Anthropic, the artificial intelligence startup backed by Google, has announced significant updates to its chat program Claude, expanding consumer access and enhancing its underlying technology. The company claims that “Claude 2” is now even better at tasks such as computer coding and arithmetic. Additionally, Anthropic has improved the system’s ability to avoid offensive or dangerous responses, making it safer for users. Consumers in the U.S. and UK can now chat with Claude online, whereas access was previously limited.
In the rapidly evolving field of AI technology, companies like Anthropic are constantly striving to gain a competitive edge. This latest release from the San Francisco-based startup follows a wave of product launches in Silicon Valley, with major players like Google and Microsoft also pushing advancements in AI. Inflection AI and various other startups are also vying for prominence in this space.
The immediate benefit of these AI advancements is the significant increase in speed for tasks such as reading and writing. Generative AI, like that employed by Anthropic, can draft blog posts and analyze existing text at an accelerated pace. One particular feature that sets Anthropic apart is its system’s ability to process up to 75,000 words at a time, allowing the AI to delve into lengthy business documents.
Amid growing concerns over regulation, spurred by copyright issues and instances of AI-generated misinformation, Anthropic has prioritized the safety of its technology. In the latest upgrade to Claude, the company claims to have doubled the model’s performance on safety evaluations. OpenAI, the creator of ChatGPT, has maintained its leadership position within the industry. Its recent GPT-4 model is “multimodal,” capable of responding not only to text but also to images provided by humans. GPT-4 achieved an impressive score of 75.7% on the Multistate Bar Exam, which assesses whether U.S. attorneys are qualified to practice law.
Anthropic, founded by former executives of OpenAI, has announced that its new and improved Claude 2 model now scores 76.5% on the multiple-choice section of the Bar, a significant improvement from the previous model’s score of 73%. Although it is unclear whether this result is directly comparable to OpenAI’s performance, the achievement demonstrates Anthropic’s progress in the field. Sandy Banerjee, a go-to-market official at Anthropic, revealed that the company is exploring multimodal systems and harnessing Claude’s ability to comprehend vast amounts of content rapidly.
With its impressive speed and capacity for quick decision-making, Anthropic’s systems are highly compelling for businesses and end customers alike, according to Banerjee.
Credit: The Star : Tech Feed
