Read full article about: South Korea's SK Telecom bets big on AI with $100 million investment in Anthropic
SK Telecom invests $100 million in AI company Anthropic, joining investors including Google and Spark Capital. Anthropic develops basic AI models and competes with OpenAI. The Korean telco is working with Anthropic to develop a global telecom-focused, multilingual, large language model and AI platform. Anthropic recently released Claude 2, a chatbot on par with GPT-3.5, but with about 6x the context window. Anthropic also offers a faster, less powerful, but cheaper chatbot with Claude 1.2.
Read full article about: Chat your way to better AI safety with Stable Chat's research preview
Stability AI has launched Stable Chat, a free website for AI safety researchers and enthusiasts to interact with and evaluate the company's large language model, Stable Beluga (formerly known as "Free Willy"). The platform is designed to help improve the safety and performance of LLMs. Stable Chat invites users to provide feedback on the safety, usefulness, and quality of LLM responses, and to report biased or harmful content. You can try the research preview for free.
Comment
Source: Stability AI
Read full article about: Claude Instant 1.2 shows notable gains in math, coding, reasoning, and safety, says Anthropic
Anthropic, the AI startup co-founded by former OpenAI leaders, has released an updated version of its text generation model, Claude Instant 1.2, available through its API.
It boasts improved performance in areas such as math, coding, reasoning, and security, with better resistance to jailbreaking and hallucination. Claude Instant 1.2 aims to compete with entry-level offerings from OpenAI, Cohere, and AI21 Labs. However, it's less powerful than Claude 2, which is on par with OpenAI's GPT-3.5, according to a newly released chatbot benchmark.
Introducing our latest version of Claude Instant, version 1.2, available now through our API!
Claude Instant is our faster, lower-priced yet still very capable model, which can handle a range of tasks including dialogue, analysis, summarization, and document comprehension. pic.twitter.com/p9M2d7O9K9
— Anthropic (@AnthropicAI) August 9, 2023
