Anthropic launches new version of its AI chatbot, Claude

In the generative () arms race, Anthroc is one of the 10 most noteworthy startups and one of OpenAI’s strongest competitors.

Claude, the new version of the AI ​​chatbot that Anthropic is calling easier to talk to, is finally available for more people to try out. The company has announced that everyone in the US and UK can test the new version of the conversational bot, Claude 2, on its website.

Its public availability enables Claude to join the ranks of ChatGPT, Bing, and Bard, all of which are available to users in many countries. But Anthropic emphasizes that "think of Claude as a friendly, enthusiastic colleague or personal assistant who can help you complete many tasks by guiding him using natural language."

Claude is also described by Anthropic as being "helpful, harmless, and honest" and can do things like create summaries, write code, translate text, and more. While this sounds a lot like Google's Bard or Microsoft's Bing chatbot, Anthropic says it's built differently than these bots. It has a more conversational tone than its counterparts and is said to even have a sense of humor. Not only that, Claude also follows a set of charters to modify its responses on its own, rather than relying on human moderation.

While Google-backed Anthropic initially launched Claude in March, the chatbot was only available to businesses via request or as an app in Slack. With Claude 2, Anthropic has made some improvements to the chatbot's existing capabilities. In addition to being able to craft longer responses, Claude 2 is also more proficient in math, coding, and reasoning than previous Claude models.

For example, Anthropic says Claude2 got 76.5 percent of the multiple-choice sections of the bar exam correct, while its predecessor, Claude 1.3, only got 73 percent. Claude2 is also two times better at "responding harmlessly," according to Anthropic. That means it should be less likely to respond with hurtful content when you interact with it than the previous model, though Anthropic doesn't rule out exceptions.

Unlike Bard and Bing, however, Claude2 is still not connected to the internet and is being trained on data until December 2022. While this means it can't display information about current events, its dataset is still more recent than the one used by the free version of ChatGPT. Anthropic representative Sandy Banerjee told The Verge that you can still give Claude' a recently published website or webpage, and it should be able to query it.

Additionally, Anthropic recently expanded Claude's context window to about 75,000 words. This means you can upload dozens of pages, or even an entire novel, to the robot for it to parse. So if you need a quick summary of a complex and lengthy research paper, Claude is the way to go. Other models are much more limited, with ChatGPT capped at about 3,000 words.