Anthropic uses Google Cloud for potential ChatGPT rival; a prelude to M&A?

“We’re thrilled to be able to partner with Google Cloud on training and deploying useful AI systems,” said Dario Amodei, CEO of Anthropic. “We’ve been impressed with Google Cloud’s open and flexible infrastructure. We are excited to partner with Google to scale up our systems in the coming months.”

Google Cloud has prior experience with large-scale AI systems, so Anthropic plans to use Google Cloud’s custom machine learning algorithms to handle demanding workloads and continue its research in developing secure and responsible AI technology. By partnering with Google Cloud, Anthropic will work on the same infrastructure running Google’s search engine and YouTube video streaming service.

“Google Cloud is providing open infrastructure for the next generation of AI startups, and our partnership with Anthropic is a great example of how we’re helping users and businesses tap into the power of reliable and responsible AI,” said Thomas Kurian, CEO of Google Cloud.

Anthropic’s Claude becomes a new rival to ChatGPT

In the whole ecosystem of generative AI, cloud service providers have benefited the most first and foremost by hosting the models and related code. The benefits are extending outward as services like ChatGPT not only use Microsoft cloud computing power to support its users (reportedly in the neighborhood of 100 million and growing), but Microsoft is integrating ChatGPT into its Bing search engine and eventually other Microsoft products such as Office 365. Anthropic, for its part, will be using Google Cloud’s computing power to support the next generation of generative AI.

Anthropic was established in January 2021 by the former employees of OpenAI, where they developed GPT-3 and Reinforcement Learning from Human Feedback. Since its inception, the company has published 14 research papers pertaining to construction language models.

After nearly two years, the company released a public language model assistant called Claude, which operates on Google Cloud and other platforms. Anthropic is collaborating with several technology partners to implement Claude with safety measures in large-scale AI systems. The goal of the company is to build a transparent and easily understandable AI system that can be securely deployed on a large scale.

Anthropic Claude is an AI-based conversational assistant that uses a natural language processing technique similar to ChatGPT. Both conversational AI methods rely on reinforcement learning to train a model over their outputs and use preferred generations to fine-tune the output.

Data-labeling provider Scale AI performed a head-to-head evaluation between Claude and ChatGPT on various parameters, such as calculation, factual knowledge and reasoning, analysis of fictional works, mathematical reasoning, code generation, comedic writing, and text summarization.

ChatGPT was found to be a better conversational AI for code generation or reasoning about code, as Claude appears to have more errors. When it comes to calculation and reasoning of logical problems, both of the generative AI techniques have performed with a similar accuracy level.

Analysis

Claude could become an alternative to ChatGPT in the near future as it still requires a lot of improvement and fine-tuning. That being said, sufficient progress would make Anthropic an attractive target for an acquisition, with Google at the front of the line. For its part, Google has been at the forefront of AI development, but from a product and marketing perspective has been put on its back foot with the popularity of ChatGPT. A strategic acquisition of Anthropic would be one obvious way to rectify perceived shortcomings, though not something we would expect to see this year.

Article Topics

acquisition  |  AI/ML  |  Anthropic  |  ChatGPT  |  cloud  |  Google  |  model development