Amazon enters AI chatbot race (finally)

After reports came out, around this time last year, that Amazon was building its very first flagship Large Language Model (LLM), codenamed Olympus, it seems that its debut is right around the corner: Next week, at its AWS conference, re:Inven.

Insiders have leaked that Olympus will have all the multimodal capabilities that LLMs from Google, OpenAI, and Microsoft have, ie. it will be able to process and understand text and images, but with a difference. Olympus will allegedly be able to find specific visual moments in videos or images, via simple text prompts. For example, if you were watching a game of basketball, and you asked Olympus to “find the winning goal” it would search the footage and deliver the goal that won the match.  

This will not only allow Amazon to finally enter the LLM race, but it will also reduce its reliance on third-party LLMs, from the likes of Anthropic, which they just invented $8B in.