.png){kind=logo}
Amazon Web Services (AWS) has revealed one major AI advancement at its flagship event re:Invent. The tech giant has introduced new models, infrastructure, and tools designed to accelerate the development and deployment of AI.
Revamping the AI landscape, AWS has released the long-awaited product: the new AI model, Nova 2 Omni. This innovation can handle text, speech, images, and videos simultaneously.
The model, currently in preview, is part of the Nova family alongside Nova 2 Lite and Nova 2 Sonic. Nova 2 Omni can combine several AI models into one and access a wide range of applications across various industries.
In conjunction with Nova 2 Omni, AWS has launched Nova Forge, which allows customers to build their own AI models. According to AWS CEO Matt Garman, this move ensures greater accuracy and personalization, allowing enterprises to meet certain needs with more precision.
AWS has announced enhancements to Amazon Bedrock AgentCore, which it claims is the most advanced platform for securely building and deploying AI agents at scale. New features include:
Policy in Preview: Allows teams to establish clear boundaries on an AI agent’s actions using natural language.
AgentCore Evaluations has 13 pre-built evaluators tracking real-time interactions for correctness, safety, and performance alerts.
AgentCore Memory provides episodic memory, allowing agents to learn from experience and improve decision-making.
Also Read: Samsung Galaxy S24 Now Available at Rs. 38,957: Check the Amazon Deal
AWS has broadened Amazon Bedrock’s model portfolio by introducing 18 open-weight, fully managed models. Four of these are Mistral AI models that are exclusive to Bedrock. Consequently, the total number of serverless models is about 100, providing broad access to the best AI technologies.
The firm has made the EC2 Trn3 UltraServers with Tranium3 chips available, delivering a whopping 4.4 times the compute power, 4 times the energy efficiency, and 4 times the memory bandwidth compared to the previous generation. AWS has also provided a sneak peek of at least 6x the processing speed, with improved FP8 performance, to accelerate AI training and inference.
AWS has unveiled a new class of scalable, autonomous AI agents: Kiro, the virtual developer; AWS Security Agent; and AWS DevOps Agent. AWS AI Factories has also been introduced. This infrastructure product combines NVIDIA and Trainium chips with AWS networking and AI services to give rapid AI app development.
With innovative services and technical upgrades, AWS solidifies its leadership in AI infrastructure and tooling, pioneering the next-gen wave of AI-driven transformation across several sectors.
