Bilgesu Erdem
In a bid to rival top models from OpenAI and Alphabet, Amazon is reportedly investing millions in training a new large language model (LLM) codenamed “Olympus.” With a staggering 2 trillion parameters, Olympus could be one of the largest models being trained, surpassing OpenAI’s GPT-4 model, which is estimated to have one trillion parameters.
Led by Rohit Prasad, former head of Alexa and now head scientist of artificial general intelligence (AGI) at Amazon, the Olympus project brings together researchers from Alexa AI and the Amazon science team to develop cutting-edge AI models. This consolidation of AI efforts across the company aims to provide dedicated resources and expertise for the project.
Amazon’s investment in LLM development stems from its belief that having homegrown models could make its cloud computing offerings, particularly Amazon Web Services (AWS), more attractive to enterprise clients seeking access to top-performing AI models. While there is no specific timeline for Olympus’ release, the company’s commitment to this project signals its determination to stay ahead in the AI race.
The pursuit of larger LLMs is driven by the notion that more parameters equate to greater understanding and the ability to better connect various nuanced concepts. However, training such models poses significant challenges. The sheer amount of computing power required increases costs and can lead to slower results. Additionally, the energy consumption associated with training large LLMs is a growing concern.
Despite these challenges, the potential benefits of larger LLMs are compelling. Amazon believes that Olympus could significantly enhance the capabilities of its other AI products, such as Alexa. The company has already hinted at a “smarter and more conversational Alexa” powered by generative AI, and Olympus could play a crucial role in realizing this vision.
As Amazon joins the AI race with Olympus, it remains to be seen whether the model will live up to its ambitious goals. However, the company’s commitment to AI development and its willingness to tackle the challenges posed by large LLMs demonstrate its determination to remain a leader in the field.
