Amazon is investing millions of dollars in training its own LLM, called Olympus, according to Reuters. This model should become a serious competitor to Google and OpenAI.
According to sources of Reuters, Amazon is training Olympus, a model that has 2 trillion parameters. This would make the model larger than OpenAI’s current GPT-4 model. GPT-4 has 1 trillion parameters.
According to sources, Amazon strongly believes in developing its own AI models. With the models, the tech giant wants to make its AWS services more attractive, Reuters writes. In the public cloud environment, enterprise customers are eager to have access to high-performance AI platforms.
Development gets boost
Amazon is reportedly currently freeing up millions of dollars to train the AI model. The team engaged in these training activities is led by head scientist of artificial general intelligence Rohit Prasad. He was the former head of digital assistant Alexa.
The development of LLM Olympus is not the only AI activity being undertaken by Amazon. The online retailer and tech giant already has its own LLM, Titan, and is working with AI model developers Anthropic and AI21 Labs. The LLMs of these two AI specialists will be offered on AWS.
When Olympus will become available is not known.