While OpenAI took a stumble last week with a much-publicized ouster and ultimate return of its CEO Sam Altman, the company’s cross-town rival Anthropic was launching a new version of its Claude chatbot.
Claude 2.1 provides a 200K token context window, meaning users can relay up to 200,000 tokens, which equals 500 pages of data in an interaction.
For example, a user could upload the entire Iliad and then be able to have an interactive conversation with the chatbot about the text.
The new version also has a 2x improvement in reducing model hallucinations and is much better at summarizing large complex documents like contracts without presenting erroneous answers.
Like OpenAI, Anthropic is beta-testing adding interfaces to embed its chatbot into proprietary applications via APIs.
The company is also expanding its developer environment to allow coders to embed and test the chatbot with prompts.
Anthropic is seen as one of the biggest rivals of OpenAI and is at the center of much of the debate on how AI will be used in the world.
Anthropic was founded by a splinter group that broke away from OpenAI in 2021 to create a more ethical model.
The company has often been tied to the effective altruism movement, which is a philosophical movement that centers on the idea of using evidence and reason to help as many people as possible.
The differences between OpenAI and Anthropic seem to be one of the topics at the center of the recent dust-up at OpenAI where Sam Altman, prior to his attempted ouster, faced off with OpenAI board member Helen Toner.
Toner, who is a Director at Georgetown’s Center for Security and Emerging Technology (CSET), reportedly wrote an academic paper praising Anthropic’s approach as compared to OpenAI.
Anthropic is seeking an additional $2 billion in funding to augment several rounds of investment from Amazon, which pledged $4 billion, and Google which invested $2 billion.