“Olympus”

Blog by: Priyanka Rana

Olympus

Amazon is investing heavily in an advanced large language model (LLM) called “Olympus.” This model is reported to have 2 trillion parameters, potentially outstripping OpenAI’s GPT-4, which has one trillion.

Project Confidentiality

• Details of the project have been kept under wraps, with sources discussing it under anonymity.
• Amazon has not made an official statement following the information leak.

Previous and Parallel AI Investments

• Amazon has developed smaller models such as Titan and has collaborations with AI startups like Anthropic and AI21 Labs.
• These partnerships extend the reach of Amazon’s AI capabilities to AWS users.

Strategic Move for AWS

• By developing in-house LLMs, Amazon aims to enhance AWS offerings.
• The move could attract enterprise clients seeking access to top-tier AI models on AWS.

Large Language Models as a Core Technology

• LLMs are designed to process vast datasets and generate responses akin to human interaction.
• The training of larger models incurs high costs due to the significant computing power involved.

Credit: GK Today