As part of their partnership, Anthropic and SK Telecom will work together for building large language models (LLMs) for telcos.
US-based artificial intelligence startup Anthropic PBC announced that it has secured a staggering $100 million in investments from South Korea’s largest telecommunications company SK Telecom. With this investment, SK Telecom also seeks to bolster its telecommunications-driven AI business.
The two players have entered into a partnership wherein Anthropic will build large language models (LLMs) customized or telcos. Anthropic is already competing with players like OpenAI in building different AI foundation models. It also stands to be one of the most well-funded AI firms in the market. earlier this year in May 2023, Anthropic raised $450 million from investors such as Spark Capital and Alphabet Inc’s Google.
Established in 2021 by CEO Dario Amodei and his sister Daniela, both of whom were former senior team members at OpenAI, Anthropic’s mission revolves around creating trustworthy AI systems and conducting research on the potentials and challenges of AI.
Unlike OpenAI, Anthropic takes a different approach. While OpenAI employs reinforcement learning from human feedback (RLHF) to enhance safety and minimize risks, Anthropic utilizes reinforcement learning from AI feedback (RLAIF). This feedback mechanism involves one AI model correcting another based on a set of “rules or principles,” known as “Constitutional AI” by Anthropic.
As part of their collaboration, SK Telecom and Anthropic are set to co-create a multilingual Large Language Model (LLM), catering to various languages like Korean, English, German, Japanese, Arabic, and Spanish.
The two entities, SK Telecom, and Anthropic, are collaborating to integrate this multilingual LLM into the Telco AI Platform, which is currently under development by the Global Telco AI Alliance.
Anthropic’s Claude 2 Model
Anthropic’s Claude series of large language models (LLMs) have been in direct competition with OpenAI’s ChatGPT. Recently, Anthropic unveiled its Claude 2 model in order to compete with ChatGPT-4.
Anthropic asserts that Claude 2 boasts enhanced performance, longer response capabilities. Users can now access them through an API, along with its new public beta website called Claude.ai. Marketed as a relatively “harmless” AI system, Claude can handle a diverse range of conversational and text-related tasks.
The company further added:
“We have heard from our users that Claude is easy to converse with, clearly explains its thinking, is less likely to produce harmful outputs and has a longer memory. We have made improvements from our previous models on coding, math and reasoning”.
Merely a week ago, Anthropic introduced an upgraded edition of Claude Instant, a swifter and more cost-effective text-generation model accessible via an API. The company asserts that this new version exhibits “substantial” enhancements in domains like mathematics, coding, reasoning, and safety.
next