Meta Launches 'Universe of AI’ Across Apps

At the Meta Connect conference, Mark Zuckerberg announced extensive AI enhancements coming to Instagram, Facebook, WhatsApp, and more! From advanced conversational assistants to a diverse AI 'cast of characters', the company is integrating state-of-the-art AI into the apps that billions interact with every day. The improvements are designed to spark creativity and enhance user interaction, but Zuckerberg suggests this is just the beginning, with “lots of limitations” still to be refined.

List of Features:

• Meta AI Chatbot: A beta rollout is enabling advanced conversational assistance on WhatsApp, Messenger, Instagram, and even on Ray-Ban Meta smart glasses and Quest 3, providing real-time information and image generation.

• AI ‘Cast of Characters’: Meta introduced 28 AI personalities in beta, some personified by celebrities like Snoop Dogg and Tom Brady, each bringing unique interests and backgrounds to the user experience.

• AI Studio Platform: A new platform allowing businesses to create AI chatbots for Meta’s messaging services, and a sandbox tool is on the horizon to allow anyone to experiment with creating their own AI.

• Generative AI Stickers: Users can now edit and co-create images on Instagram with AI editing tools, turning text prompts into stickers in just a few seconds. The tool uses Llama 2 and Meta’s new image generation model, Emu, and turns text prompts into stickers in just a few in seconds.

• Ray-Ban Smart Glasses: Equipped with Meta AI, these glasses, starting at $299, allow users to spark creativity, gather information, and control the glasses using just their voice.

The Big Picture: Mark Zuckerberg highlighted that these AI advancements are in their early stages, with substantial room for improvement and refinement. The incorporation of AI into Meta’s platforms signifies a transformative phase in enhancing user interaction and experience, paving the way for a more interconnected and intelligent social media landscape.

Generative Modeling Tools Renders 2D Sketches in 3D

Overview: A research team at Carnegie Mellon University's Robotics Institute has developed pix2pix3d, a tool leveraging generative AI that can convert 2D sketches into realistic 3D models, making the creation of custom 3D content accessible for both amateur and professional designers.

How it works: Pix2pix3d is a unique 3D-aware conditional generative model. It allows users to input a 2D sketch or more detailed information from label maps, such as a segmentation or edge map, synthesizing a 3D volumetric representation of geometry, appearance, and labels. The synthesized 3D model can be rendered from multiple viewpoints to create realistic, photograph-like three-dimensional images. Users are also able to interactively edit and modify the 3D image in real-time, adjusting the original 2D sketch as needed, making it a versatile tool for designing and adjusting products in fields like manufacturing.

Why it matters: Pix2pix3d is a game-changer for 3D content creation. It's not just about transforming sketches into 3D models but about making 3D modeling more inclusive and accessible, allowing even those with minimal artistic ability to create detailed and accurate 3D representations. This development could revolutionize industries, from customized household furniture creation to the development of video game content and film effects, by providing a user-friendly platform for transforming imaginative concepts into 3D reality, fostering innovation and creativity in design across various domains.

• 🥽 Review: This ultraportable VR headset gave me a taste of Vision Pro at a fraction of the cost

• 🏥 For your health: Virtual reality vs. opioids: clinical trial of non-pharmacological pain management during burn wound care

• 👀 Watch: A Hackers' Guide to Language Models

• 🦺 Manufacturing: 3D insights into an innovative manufacturing process

• 🔮 Future: Computer vision's next breakthrough

• 🌐 Beyond Just Devices: The Path to Spatial Computing

• 👯 AR Glasses Spawn a Whole New Social Dynamic