Sundry Pictures
Amazon (NASDAQ:AMZN) is investing thousands and thousands in coaching a big language mannequin in an try to compete with prime AI fashions from the likes of Microsoft (MSFT)-backed OpenAI and Alphabet (GOOG) (GOOGL), Reuters reported citing individuals with information of the matter.
The mannequin, dubbed Olympus, has two trillion parameters, the report added, which might make it one of many largest fashions being skilled. OpenAI’s GPT-4 is reported to have one trillion parameters.
The group is led by Rohit Prasad, former head of Alexa, who now stories to Amazon’s CEO Andy Jassy. As head scientist of synthetic common intelligence, or AGI, at Amazon, Prasad introduced collectively researchers who had been engaged on Alexa AI and the Amazon science group to coach the fashions.
Amazon has already skilled smaller fashions resembling Titan and in September stated it will make investments as much as $4B in San Francisco-based AI startup Anthropic, and decide a minority stake within the firm. Amazon has collaborated with Anthropic and AI21 Labs, to supply companies to Amazon Internet Companies, or AWS, customers.
Google has additionally dedicated to take a position $2B in Anthropic.
Amazon thinks that having self-developed fashions might make its choices extra enticing on AWS, the place enterprise purchasers need to entry high-performing AI fashions. Nevertheless, there was no particular timeline for launching the brand new mannequin, based on the report added.
In an earnings name in April, Amazon executives famous the corporate would enhance funding in LLMs and generative AI whereas decreasing on success and transport in its retail unit, the report famous.
Generative AI companies have taken the world by storm, because the launch of OpenAI’s ChatGPT final 12 months. Globally, firms have launched their very own LLMs which may present companies resembling content material and picture technology, to call just a few.
Alibaba’s (BABA) Tongyi Qianwen 2.0 and Tongyi Wanxiang, Baidu’s (BIDU) Ernie Bot, OpenAI’s text-to-image device DALL·E 3, Meta Platforms’ (META) AudioCraft, SeamlessM4T, and Llama 2, Google’s Bard, and Getty Pictures’ (GETY) mannequin referred to as Generative AI by Getty Pictures, are a number of the LLMs.