Amazon Team to Train New LLM AI Model "Olympus"

Companies around the world are increasingly adopting artificial intelligence technology, and the potential for AI to drive global economic growth and investment opportunities is enormous. AI is rapidly evolving, with new regulations and business applications emerging constantly. Although AI is still a relatively new technology, there are many uncertainties in terms of security and ethics. However, companies that completely avoid AI may face difficulties in the future, similar to the situation where companies that avoided the internet in the early days are now at a disadvantage.

The economic impact of AI may quickly become apparent and bring about significant changes. With the combination of human intelligence and machine capabilities, employees and businesses in various industries have the potential to enjoy advantages that can completely transform the way we work, accelerate productivity growth, and provide positive returns for investors.

Amazon Joins the Battle of Robots

Global giant Amazon is investing millions of dollars to train a new large-scale language model (LLM) called Olympus, which may compete with the best models from OpenAI and Google.

Olympus has 20 trillion parameters, surpassing any other LLM currently being trained. This is nearly twice the number of parameters in OpenAI's GPT-4 model, which is widely regarded as the top LLM in the market. Parameters are crucial for AI performance, so the additional capacity of Olympus is likely to enable it to generate more accurate and human-like responses, surpassing GPT-4.

What Benefits Will Olympus Bring?

Rohit Prasad, former head of Alexa, now reports directly to CEO Andy Jassy and serves as the Chief Scientist for Artificial General Intelligence (AGI) at Amazon. In this role, Prasad has assembled a team of researchers with expertise in Alexa AI and Amazon Science. The team's current focus is on training models, effectively integrating AI projects within Amazon, and providing dedicated resources for these efforts.

Amazon has already trained smaller models like Titan and has partnered with AI model startups such as Anthropic and AI21 Labs to offer their models to users of Amazon Web Services (AWS). Amazon emphasizes the development of internal models because they believe it enhances the attractiveness of their products on AWS.

On AWS, enterprise customers typically require high-performance models, and Amazon aims to meet this demand. However, insiders have pointed out that there is no specific timeline for the release of the new models.