AWS re:Invent 2023: Amazon Q, new AI chip, and more announced
AWS re:Invent 2023: Amazon Web Service (AWS)'s annual flagship event, re:Invent, is currently underway. Yesterday was Day 1 of the event, and during its opening sessions, AWS CEO Adam Selipsky made a series of new announcements and updates. The main session on Day 2 also concluded with a couple more announcements. There is a clear focus on artificial intelligence (AI) at the event being held in Las Vegas. The event will run until December 1st, and we expect more announcements to be made. But these are the 5 biggest announcements from the AWS event so far. Let's take a look.
AWS re:Invent 2023 Announcements
AWS Transcription Platform Gets AI: The event also made a specific announcement from Amazon Transcribe, highlighting an updated multi-million-parameter speech foundation model, which expands automatic speech recognition to more than 100 languages. AWS said this managed automatic speech recognition (ASR) service simplifies the integration of speech-to-text capabilities into applications.
Amazon Q: Amazon Q is a business chatbot where customers can ask specific questions about their companies. It works as an artificial intelligence assistant that allows users to obtain useful information based on their own data. It serves as a tool for tasks such as retrieving company guidelines or understanding code, streamlining the process by presenting information without performing extensive document searches. Access it through various platforms, including the AWS Management Console, company documentation pages, developer environments like Slack, and third-party apps.
Trainium 2 AI Chip: The company's CEO announced significant advancements in its Trainium chip technology. The company highlighted this by announcing a new AI chip that can train fundamental models twice as fast. The company claims the second generation of Trainium 2 boasts a remarkable four times faster model training speed than its predecessor.
AI chip to solve primary problems in quantum computing: During a re:Invent keynote, Peter Desantis, general manager of AWS EC2, revealed a new chip designed to correct errors, like reversing binary codes from 1 to 0, that can occur in quantum computing queries. AWS is currently experimenting with it, and not a finished product.
AWS and Nvidia enlarge their AI generation partnership: AWS is ready to deploy Nvidia's upcoming GH200 chips, making it the first cloud provider to do so. These chips, unveiled by Nvidia CEO Jensen Huang at AWS re:Invent, feature three times more memory than the widely used H100 chips.