AWS re:Invent 2023
AWS re:Invent 2023: The annual flagship event of Amazon Web Service (AWS), re:Invent is currently underway. Yesterday, November 28, marked day 1 of the event, and during its opening keynote sessions AWS CEO Adam Selipsky made a number of new announcements and updates. The day 2 keynote session has also wrapped up with a couple more announcements coming along the way. There is a distinct artificial intelligence (AI) focus in the event that is being hosted in Las Vegas. The event will run through December 1 and we expect more announcements to be made. But these are the 5 biggest announcements to come from the AWS event so far. Let us take a look.
5 biggest announcements from AWS re:Invent 2023
- Amazon Q:Amazon Q is a chatbot for businesses where clients can ask questions specific to their companies. It functions as an AI assistant letting users get actionable insights based on their own data. It serves as a tool for tasks such as retrieving company guidelines or comprehending code, streamlining the process by presenting information without the need for extensive document searches. The tool is accessible through various platforms, including the AWS Management Console, company documentation pages, developer environments like Slack, and third-party applications.
- Trainium 2 AI chip:During AWS re:Invent 2023, the company CEO announced significant advancements made toward its Trainium chip technology. The company highlighted this as it announced a new AI chip that can train foundational models twice as fast. The second-generation Trainium 2 boasts a remarkable fourfold increase in model training speed compared to its predecessor, claims the company. Selipsky also said that the Trainium 2 features three times the memory capacity of its predecessor, which allows it to handle more complex and larger datasets.
- AWS and Nvidia expand their gen AI partnership:AWS is set to deploy Nvidia’s upcoming GH200 chips, making them the first cloud provider to do so. These chips, introduced by Nvidia CEO Jensen Huang at AWS re:Invent, boast three times the memory of the widely used H100 chips. Additionally, AWS users can anticipate the arrival of Nvidia’s DGX Cloud, an AI “factory,” adding a notable dimension to the collaboration between the two tech giants in advancing AI capabilities.
- An AI chip that can solve key quantum computing problems?During a re:Invent keynote, Peter Desantis, AWS EC2’s general manager, revealed a new chip designed to correct errors, such as flipping binary codes from 1 to 0, which can occur in quantum computing queries. AWS is currently experimenting with it and it is not a finished product. Desantis emphasized that mitigating these errors could potentially enhance the usability of quantum computing.
- AWS’ transcription platform gets AI: AWS re:Invent 2023 also made an Amazon Transcribe-specific announcement, highlighting an upgraded multi-billion parameter speech foundation model, expanding automatic speech recognition to over 100 languages. This managed automatic speech recognition (ASR) service simplifies the integration of speech-to-text capabilities into applications, the company stated. The foundation model, trained with self-supervised algorithms, learns universal speech patterns across languages and accents.