Amazon is in turn entering the artificial intelligence market. The large American firm yesterday unveiled its own AI model, simply titled Nova.
While the race for artificial intelligence is in full swing between Microsoft, Google, Meta and OpenAI, a new player could well find its way between the digital giants. Amazon in fact announced during a conference this Wednesday, December 3, that it was developing its own AI model, called Nova.
From Las Vegas, Amazon Web Services (AWS) successively detailed all the capabilities of its new tool, ranging from text generation to image and video creation. At first glance promising, this novelty intends to enter the very competitive artificial intelligence market.
An AI with multiple functionalities
If Nova looks interesting, it does not revolutionize the AI world. Like its competitors, Amazon deploys an all-terrain model, capable of responding to all kinds of requests and objectives:
- New Micro: AI capable of generating texts.
- Nova Lite: AI capable of generating texts from an image, video or other text.
- New Canvas: AI capable of generating images from prompts.
- Nova Reel : AI capable of generating short videos (up to 6 seconds) from a prompt. Amazon promises the ability to create 2-minute videos by 2025.
- Nova Premier : AI capable of responding to sophisticated queries, especially in the professional context. This module is currently in development and will be released in early 2025.
Assets that make Nova a solid alternative
With the launch of Nova, Amazon could quickly gain ground in the artificial intelligence market. In addition to offering a fast, comprehensive and efficient tool, the American firm offers much lower prices than its competitors. According to Amazon, the first three modules of Nova would be “at least 75% less expensive than the best performing models in their category […]in addition to being the fastest“.
In just a few seconds, Nova’s different domains will be able to respond to any query. The generation of images and videos will produce qualitative content in record time. The rendering is really convincing and detailed, as evidenced by this video from Amazon, which features Pasta City.
Already additions planned for the future
In addition to the six different Nova models, Amazon plans to add two more options. The first should be able to transform a textual, video or audio request into a result in the chosen form (written, illustrated, or audio). The second will be of type “speech to speech”that is to say it will be possible to chat with Nova, like with Siri or Alexa.
Nova’s strengths are undoubtedly its ability to operate in 200 different languages, but also the attention it pays to the ethics of its program. Each of the six models incorporates protection and security measures, thanks to AWS AI Service Cards. This system will make Nova secure, with rules and limitations on responsible use.
Exclusively reserved for businesses and AWS client developers, access for the general public is unfortunately not planned for Nova.
