New AI model from Amazon?

Look what you missed in this AI-week! Amazon unveils powerful new AI models, Japanese companies level up with generative AI for research, and Meta’s compact moderation model is making waves. Let’s dive in!

1: ”Amazon launches a family of multimodal AI models”

Amazon announces Nova, a new family of multimodal AI models

AWS announced Nova, a new family of multimodal generative AI models available via its Bedrock platform. The lineup includes text, image, and video generation models, with standout features like massive context windows, fine-tuning capabilities, and built-in safeguards. AWS also teased future speech-to-speech and “any-to-any” models for 2025.

2: ”Japanese firms adopting generative AI for information searches”

Japanese firms begin adopting generative AI for information searches

Japanese companies like Lion are leveraging generative AI to streamline research, enabling faster and more precise access to internal databases. These tools enhance efficiency and knowledge sharing; still, challenges like copyright remain a burden that requires improved regulations and AI literacy. How do they use it? Read the article!

3: ”A new moderation model”

Meta AI Releases Llama Guard 3-1B-INT4: A Compact and High-Performance AI Moderation Model for Human-AI Conversations

Meta’s Llama Guard 3-1B-INT4 revolutionizes AI safety moderation with a compact 440MB model optimized for mobile devices. Using advanced compression, it delivers top-tier safety and multilingual performance, outperforming larger models like GPT-4 in key moderation tasks.