OpenAI is Building "her"

Welcome to the Highest Quality AI & Tech Newsletter!

Share With Your Smart Friends, They Will Thank You For It 😉 

In Today’s Edition

📱 ChatGPT4o Unveiled
🚨 Google I/O 2024 Highlights
🤔 Ilya Leaves OpenAI
🧑‍🚀 The Robot Astronaut
🌀 A Portal Between US & Ireland

AI

Unveiling GPT-4o: OpenAI’s Omnimodel

OpenAI unveiled their latest flagship model (GPT-4o) through a new and updated user interface, as well as an API. This model, named "Omnimodel" significantly advances conversational AI by seamlessly integrating capabilities across text, audio, and vision.

The Omnimodel Advantage 🤳
GPT-4o is a leap forward from previous models that combined intelligence, text-to-speech, and audio functionalities. It reasons fluidly across voice, text, and visual inputs, making interactions more conversational and natural.
Notably, GPT-4o can understand the context and tone of conversations and respond appropriately, even when interrupted, enhancing the overall user experience.

Advanced features 📱 
Real-time vision (images and videos)
Memory capabilities
Other GPTs
Web browsing
Advanced data analysis

GPT-4o offers several substantial improvements 🧑‍💻 
Faster Performance: Twice as fast as previous models.
Cost Efficiency: 50% cheaper than its predecessor.
Higher Rate Limits: Five times the rate limits compared to GPT-4 turbo.
Broad Accessibility: Available for free to all users through both the UI and API.

The model supports over 50 languages, significantly reducing latency and making it the smartest, most natural assistant OpenAI has showcased live.

GPT-4o will be available to all ChatGPT users through the Web UI, including those on the free tier. It will also be accessible via the API and the playground.

Full Details

Google I/O 2024 (Highlights)

Google I/O, the annual developer conference, kicked off with a slew of exciting announcements. Here's a rundown of the most significant news:

AI Developments 🤖 

LearnLM Models: Google unveiled LearnLM, a family of generative AI models designed for educational purposes. Integrated with Google Classroom, these models assist in lesson planning and provide conversational tutoring.

YouTube AI Quizzes: New AI-generated quizzes on YouTube allow viewers to interact with educational content more dynamically, thanks to the Gemini model’s long-context capabilities.

Gemma Model Expansion: Responding to developer requests, Google is launching a 27-billion-parameter version of its Gemma model, optimized for next-generation GPUs and TPUs.

Google Play Enhancements: Updates include a new discovery feature, enhanced Play Points, and tools like the Engage SDK for personalized, immersive app experiences.

AI-Powered Features 👾 

Scam Detection: An upcoming Android feature will use the Gemini Nano model to detect scam calls in real-time, alerting users to potential threats.

Google Photos' Ask Photos: Rolling out this summer, this feature leverages AI to allow natural language searches within users’ photo collections.

Gmail Enhancements: Gemini AI will help users search, summarize, and draft emails, and automate complex tasks like processing e-commerce returns.

Gemini 1.5 Pro: This upgraded model can analyze longer documents, codebases, videos, and audio recordings, supporting up to 2 million tokens.

Gemini Live: A new feature for in-depth voice interactions with AI, adapting to users’ speech patterns in real-time and responding to their surroundings.

Gemini Nano: Integrated into Chrome, this small AI model will power features like Gmail’s “help me write” tool.

AI Integration in Google Maps: The Places API will now offer generative AI summaries of locations based on community insights.

AI in Search and Media 🔍️ 

AI-Powered Search: Google is enhancing its search capabilities with AI-generated overviews and generative AI summaries for better trip planning and information aggregation.

Imagen 3: The latest generative AI model for image creation, Imagen 3, promises more accurate, creative, and detailed image generation with fewer errors.

Full Story

OpenAI Co-Founder Ilya Sutskever
Resigns a Day After ChatGPT-4o Launch

What did Ilya see? 🤔 
We don’t know yet.

Stay Tuned

Space & Tech

The Portal: Bridge of Art and Tech

🌀 Step into the remarkable world of real-time international connection with "The Portal," an art installation that has creatively linked New York City and Dublin.

🌎️ Open until fall 2024, "The Portal" offers a daily dose of international interaction. It’s an invitation to experience firsthand the unity and shared reality of our interconnected world.

Full Story

Taikobot: The Robot Astronaut

🚀 In an exciting development in space exploration, China's Tiangong Space Station is poised to welcome an innovative addition to its crew: Taikobot. This humanoid robot is designed to serve as both a caretaker and an assistant aboard the station.

Capabilities and Design of Taikobot 🤖 
Taikobot stands at approximately 5 feet and 5 inches and weighs around 55 pounds. It is equipped with an anthropomorphic dual-arm system featuring six degrees of freedom and dexterous hands, enabling it to perform multiple tasks simultaneously. The robot boasts advanced sensors and cameras that aid in navigation and task performance, including delicate operations its human counterparts undertake.

Preparation and Testing 🛰️ 
Before its deployment, Taikobot is going through final tests, testing its capability to assist with everyday tasks such as fetching, carrying, and maintenance, and even handling tools like hammers and electric screwdrivers.

Full Story

Movie of the Week