Amazon trains Olympus AI model and hopes to compete with OpenAI

Amazon is investing millions of dollars to train an ambitious AI-powered large language model (LLM), hoping it can compete with models from OpenAI and Alphabet. This is reported by Reuters with reference to knowledgeable sources.

The model, codenamed Olympus, has 2 trillion parameters. It is being developed by a team led by former Alexa project manager Rohit Prasad. For this purpose, he engaged Amazon’s science team and researchers who worked with Alexa AI.

Recently, it became known that Amazon is making a significant step in the field of artificial intelligence, investing up to $4 billion in a well-known startup Anthropic. As part of the agreement, Amazon employees and customers of its cloud services will be among the first to have access to Anthropic’s technologies. This will allow them to integrate Anthropic’s capabilities into their products.

Anthropic has also promised to primarily use Amazon’s cloud services, which includes training its future artificial intelligence models using a huge number of chips purchased from Amazon.