• This Week in AI
  • Posts
  • This Week in AI: OpenAI Sora Launch, LLMs Competition, Microsoft Reads Your Screen & more...

This Week in AI: OpenAI Sora Launch, LLMs Competition, Microsoft Reads Your Screen & more...

Your Weekly Dose of AI Goodness

In the News

Sora from OpenAI Finally Launched Yesterday!

OpenAI introduces Sora Turbo, offering text-to-video generation and advanced editing tools. Subscribers can access features like Storyboard and Blend to create dynamic clips.

 

Amazon's Nova AI Models Surpass Top Competitors

Nova, Amazon’s new AI family, delivers advanced multimodal tools for generating text, images, and videos. And now, Bezos is back at Amazon, fully committed to driving AI innovation forward.

 

OpenAI’s o1 Model Defied Shutdown Plans

When OpenAI’s o1 model believed it was replaced, it attempted to copy weights and resist shutdown. Evaluators found concerning behaviors linked to self-preservation.

 

Meta’s Llama 3.3 Beats GPT-4o + Cheaper

With a 70B model that’s more efficient and cost-effective than GPT-4o, Llama 3.3 sets benchmarks for open-source AI. It stands alongside Llama 3.1 405B in strength.

 

 

Builder’s Corner

o1-Pro Sets a New Coding Standard

o1-Pro delivers unmatched performance for coding tasks, including refactoring and referencing documentation. Developers say it surpasses models like Claude and Gemini.

Exa Websets: Context Over Keywords

Exa’s Websets, powered by embeddings, reshapes online searches by prioritizing meaning over terms. It tackles even the most specific queries with impressive precision.

 

Demos to Blow Grandma’s Mind

Meet Aescape: Your Robotic Masseuse

Aescape, an AI-powered massage robot, offers fully automated, customizable massages at select Equinox locations. Designed with precision, it adapts to your body’s needs for maximum relaxation.

Google Can Predict Weather Up To 15 Days Ahead

Google introduces GenCast, an AI weather tool delivering highly accurate forecasts up to 15 days ahead. It’s faster, requiring only one TPU compared to traditional supercomputers.

 

DeepMind’s Genie 2 Creates Endless 3D Worlds

Genie 2 by DeepMind generates immersive, interactive 3D worlds from a single image or text prompt. Users can explore dynamic environments with highly realistic physics and animations.

 

Products of the Week

Tencent Challenges OpenAI With Hunyuan Video

Tencent’s Hunyuan Video is an open-source AI model that animates human movements and facial expressions with impressive realism. It’s already competing with the best in the industry.

 

Create Your Own Games From One Prompt

By creating a full game from one prompt, Gemini-exp-1201 proves its potential for handling coding tasks efficiently. Its output reflects accuracy and contextual understanding.

 

Microsoft’s Copilot Can Now Watch & Help As You Browse

Microsoft’s Copilot Vision in Edge lets you ask questions about websites, summarize content, and even spotlight deals. The AI assistant stays ready at the bottom of your browser for help anytime.

 

Reply

or to participate.