Daily AI Edge
Posts
🎨This AI can create 1-minute cartoons

🎨This AI can create 1-minute cartoons

+ Google’s AI can now see and search with images 👀

Leo Grundström
April 9th, 2025

Today’s menu 🍽️👇️

Length: 5 minutes ⏰

NVIDIA and Stanford’s one-minute AI cartoons 🤖
Microsoft starts testing Copilot Vision update that can ‘see’ your screen and apps 😲
Google’s AI Mode can now see and search with images 👀
Amazon unveils a new AI voice model, Nova Sonic 🗣️
Create professional high-quality thumbnails with Recraft 🎨
Top AI Tools of the week! 🛠️

BREAKING AI NEWS

NVIDIA and Stanford’s one-minute AI cartoons 🤖

NVIDIA and Stanford researchers have developed a groundbreaking AI technique called Test-Time Training (TTT), which enables the generation of one-minute-long animated videos with improved consistency and storytelling. This method uses neural networks as memory, allowing the AI to maintain coherence across multiple scenes 🎥

As a demonstration, the team created AI-generated Tom and Jerry cartoons, showcasing dynamic motion and character interactions. These animations are produced directly by the model without the need for editing or stitching, marking a significant leap in AI video generation ✅

Microsoft starts testing Copilot Vision update that can ‘see’ your screen and apps 😲

Microsoft has begun testing an exciting update called Copilot Vision, which enhances its AI assistant's capabilities on Windows. This feature allows users to share their screen or apps with the AI, enabling it to guide them through tasks like using Adobe Photoshop or analyzing photos and webpages. Initially limited to the Edge browser, Copilot Vision is now expanding to other apps on your PC 💻️

The update also introduces a file search feature, allowing users to ask the AI about the contents of various file types, such as .docx, .xlsx, and .pdf. These features are currently being tested with Windows Insiders in the U.S., with a broader rollout planned for the coming weeks 📆

Google’s AI Mode can now see and search with images

Google has introduced a significant update to its AI Mode, enabling it to process and respond to questions about images. This feature combines the power of Gemini AI and Google Lens, allowing users to upload or snap a photo and receive detailed, contextually relevant responses 🔍️

The AI can analyze the entire scene in an image, understanding objects, their relationships, materials, colors, and more.

This multimodal capability uses a "fan-out technique," issuing multiple queries about the image and its components to provide nuanced answers. For example, it can identify books in a photo, suggest similar titles, and even offer recommendations with positive reviews 🖼️

Initially available to Google One AI Premium subscribers, this feature is now rolling out to millions more users in the U.S. through the Labs program. It can be accessed via the Google app on Android and iOS.

Amazon unveils a new AI voice model, Nova Sonic 🗣️

Amazon has introduced Nova Sonic, a cutting-edge AI voice model designed to deliver human-like voice interactions. This model integrates speech understanding and generation into a single system, enabling smoother and more natural conversations. Nova Sonic is available through Amazon Bedrock, the company's platform for building enterprise AI applications.

Some standout features include:

Speed and Accuracy: Nova Sonic boasts industry-leading speed with an average latency of 1.09 seconds and a word error rate (WER) of just 4.2% across multiple languages.
Cost Efficiency: Amazon claims it is 80% less expensive than OpenAI's GPT-4o.
Advanced Conversational Abilities: It adapts to pauses, interruptions, and speaking styles, making dialogues feel more fluid.

Nova Sonic is already powering Amazon's upgraded Alexa+ and is part of Amazon's broader strategy to develop Artificial General Intelligence ✅

CREATE HIGH-QUALITY THUMBNAILS

Create professional high-quality thumbnails with Recraft 🎨

Now you can use Recraft to transform your simple layout designs into professional-looking thumbnails by combining images, text, and AI-generated elements; all in one place ✅