Decrypt’s Art, Fashion, and Entertainment Hub.
Artificial intelligence startup Anthropic has secured $100 million in funding from SK Telecom (SKT), a leading player in the South Korean telecommunications sector.
“By combining our Korean language-based LLM with Anthropic’s strong AI capabilities, we expect to create synergy and gain leadership in the AI ecosystem together with our global telco partners,” said the firm’s CEO Ryu Young-sang in a press release.
The investment is intended to foster the development of a multilingual large language model (LLM) tailored for global telecommunications firms. The funding also comes hot on the heels of a Series C round in which the firm raised a cool $450 million.
SKT was involved in that round too, via its venture capital arm SK Telecom Venture Capital (SKTVC). Google also invested in Anthropic in a previous funding round.
The partnership between Anthropic and SKT will focus on specifically on creating a multilingual LLM to support languages such as English, Korean, German, Japanese, Arabic, and Spanish.
“SKT has incredible ambitions to use AI to transform the telco industry,” Anthropic co-founder and CEO Dario Amodei said. “We see industry specific LLMs as having high potential to create safer and more reliable deployments of AI technology.”
Anthropic, established in 2021, is known for Claude, an AI system designed to streamline various corporate tasks.
Claude's applications in the telecom industry include customer service, marketing, sales, and interactive consumer applications. Its core technology relies on a “constitutional” training, which makes it less reliable on RLHF techniques—reinforcement learning from human feedback—like those used by OpenAI, making its models theoretically less prone to human biases.
Last week, Anthropic updated its Claude Instant bot to version 1.2 which is less powerful but considerably cheaper than Claude 2.
With capabilities now comparable to GPT-3.5, Claude Instant represents a significant advancement in AI technology. Claude AI can handle over 100K tokens of context, for example, boasting a robust functionality when compared to OpenAI’s chatbots, which can handle between 4K, 8K, and 32K tokens of context at most.
A token in this context is like the minimum unit of information handled by a LLM. It is not the same as letters or words. For example: “A” can be a token, “Cat” can be another token, but “Medicine” has three tokens and “Humanoid” just two.
Claude 2, launched in July, extended its availability to the general public in the United States and the United Kingdom.
Unlike its predecessor, the latest version offers accessibility through both a beta website and an API, broadening Anthropic's user base.
Anthropic rides Claud 2 coattail
Dario Amodei, co-founder and CEO of Anthropic, also expressed enthusiasm for the collaboration and the potential transformation in the telecommunications industry.
SKT's investment in Anthropic reflects its commitment to AI advancements within the telecommunications sector, Amodei said. The collaboration will foster AI-driven innovations tailored to diverse linguistic and market needs.
The partnership between Anthropic and SKT signifies a strategic move in the field of AI. With the introduction of Claude and the introduction of new use cases for LLMs, this collaboration is another major investment in a year in which AI has been a favorite topic of conversation among tech investors.