Company Amazon Web Services (AWS) has just started its conference re:Invent in Las Vegas. At the same time, most of the announcements (unsurprisingly) concern the most popular technology of this year – artificial intelligence.
AWS introduced in particular the latest generation of chips designed for training AI models and for running trained models. Trainium2, which is apparently intended for training models, has been designed to provide up to 4x more performance and 2x more energy efficiency compared to its predecessor. Amazon promises these chips will enable programmers train models quickly and at lower cost thanks to reduced energy consumption. Amazon-backed OpenAI competitor Anthropic has already announced plans to build models using Trainium2 chips.
On the other hand Graviton4 is intended rather for general use. These processors are based on the Arm architecture, but they consume less energy than Intel or AMD chips. Amazon promises 30% increase in general performance using the trained AI model built into the Graviton4 processor. It should reduce costs to cloud computing for companies that regularly use AI models and offer slight increase in speed for regular users.
Amazon also announced of the new AI chatbot business-focused called Q. It is described as a new type of generative AI personal assistant and is specifically designed to help streamline work projects and customer service tasks. It can be tailored to fit any business and offers relevant answers to commonly asked questions. Amazon Q can also self-generate content and perform actions based on customer requests. It even tailors interactions based on the user’s role in the business.
It will be deployed v communication applicationssuch as Slack, and in text editing applications, which are commonly used by software developers. To that end, Q can actually change source code and can connect to more than 40 enterprise systems, including Microsoft 365, Dropbox, Salesforce, and Zendesk.