Amazon has addressed the issue of the lack of high-capability LLM models by investing $4 billion in Anthropic to access the latest version of the Claude model, which outperforms GPT-4.
While the funding may solve the short-term problem, questions arise (even among employees themselves) as to whether a tech giant like Amazon is unable to build its own LLM model, or if (currently Amazon has its own model named Titan for use on AWS, which may not rival Claude).
The Verge reports that Amazon actually has a team called Artificial General Intelligence (AGI) led by Rohit Prasad, SVP in charge of developing a new model named Olympus, a project that has been discussed previously.
According to reports, Olympus is massive, with parameters in the hundreds of billions, and has the capability to outperform Claude, with a planned release around mid-2024.
Source – The Verge
The image of the Olympus mountain in Greek mythology is created by the rival model Google Gemini, as Amazon’s Titan is not available for free use.
TLDR: Amazon invests in Anthropic to access the Claude model, plans to release the Olympus model to rival Google Gemini.
Leave a Comment