OpenAI Partners with Oracle AI Cloud

Today’s top AI Highlights:

1. OpenAI expands beyond Microsoft Azure, adds Oracle to its infrastructure

2. Samsung showcases AI-era vision and latest foundry technologies

3. Humanoid robot that can sit in a car and drive it like a human

4. Google’s new LLM helps you understand your health data from wearable devices

& so much more!

Read time: 3 mins

Latest Developments 🌍

OpenAI Taps Oracle for More Compute Power ⚡️

OpenAI is expanding its AI infrastructure by partnering with Oracle Cloud Infrastructure. This means OpenAI will now be using Microsoft’s Azure AI platform on Oracle’s infrastructure to provide additional capacity for OpenAI. OpenAI's CEO, Sam Altman, said this move will help them keep up with the growing demand for their services.

This new deal with Oracle signals that OpenAI is looking for more computing power to support ChatGPT. They want to avoid the outages that users have experienced in the past. While Microsoft remains a key partner, OpenAI is clearly looking for additional resources to keep its AI models running smoothly.

Samsung Unveils Roadmap for the Future of AI 🌟

Samsung recently held its annual Samsung Foundry Forum (SFF), focusing on the company’s vision for the future of AI and showcased its latest advancements in semiconductor technology. Samsung announced two new process nodes – SF2Z and SF4U – designed to improve performance and power efficiency for AI chips. The company also revealed its Samsung AI Solutions platform, which integrates its Foundry, Memory, and Advanced Package (AVP) businesses to offer comprehensive AI solutions.

Key Highlights:

1. Advancing process technology with its 2nm node, SF2Z - This node features optimized backside power delivery network (BSPDN) technology, which places power rails on the backside of the wafer to reduce bottlenecks and enhance performance. SF2Z is slated for mass production in 2027.

2. A High-Value 4nm Variant - SF4U offers improved performance and power efficiency through optical shrink. SF4U is scheduled for mass production in 2025.

3. Samsung AI Solutions - A turnkey platform that leverages the company’s strengths across its various businesses, offers customers comprehensive AI solutions. This platform provides high-performance, low-power, and high-bandwidth solutions tailored to specific customer needs.

4. GAA Technology for AI Performance - Samsung continues to invest in its gate-all-around (GAA) technology, critical for meeting the power and performance demands of AI chips. The company’s GAA process is entering its 3rd year of mass production and is demonstrating continuous improvements in both yield and performance. Samsung plans to mass produce its second-generation 3nm process (SF3) in the second half of 2024 and deliver GAA on its upcoming 2nm process.

Humanoid Robot Takes the Driving Wheel 🚗

Autonomous driving by humanoid robots is an active field of research, but current robots lack the flexibility and human-like proportions needed to naturally interact with vehicles designed for humans. However, a new project using a musculoskeletal humanoid robot named Musashi is making progress.

Musashi mimics the human body and is built with a structure such that it can perform tasks like sitting on a car seat and operating a steering wheel with both arms, unlike previous humanoid robots.

Key Highlights:

1. Hardware for Human-Like Interaction: Musashi's hardware features a modular design for easy adjustments, special units that make its body flexible, and movable eyes with high-resolution cameras. This helps Musashi interact with a car’s interior more naturally and efficiently.

2. Learning-Based Software for Complex Control: Musashi’s software combines online and offline learning to control its movements. It learns from sensor data to achieve desired positions and movements, and it can handle complex tasks by training on data from random commands.

3. Real-World Experiments: The researchers have conducted real-world experiments with Musashi driving a small electric vehicle, showing its ability to operate pedals and a steering wheel, recognize traffic lights and humans, and even recover from slipping while using the pedals.

LLM to Help You Sleep Better and Get Fitter 🏋️

Google published two papers on a new LLM specifically designed for personal health and wellness. This model can understand and reason about individual health data and questions, offering personalized recommendations based on data from mobile and wearable devices. The model’s capabilities include analyzing time-series data, contextualizing information, and providing insights to improve sleep and fitness habits.

Key Highlights:

1. Personal Health LLM (PH-LLM): This model, fine-tuned from Gemini, excels in generating insights and recommendations for sleep and fitness by interpreting time-series data from wearable sensors. It performs on par with human experts.

2. Agent Framework for Health Insights: An AI agent based on Gemini Ultra 1.0 uses code generation and information retrieval to analyze wearable data and provide personalized health recommendations. It achieves 84% accuracy in numerical health queries and outperforms standard models in open-ended health queries.

3. Comprehensive Evaluation: PH-LLM and the agent were tested on case studies, expert domain knowledge, and self-reported assessments of sleep quality, where they demonstrated impressive ability to provide accurate and personalized health insights.

Tools of the Trade ⚒️

1. Autodesigner 2.0: Create, prototype, and iterate UI designs using plain text prompts. You can quickly generate multi-screen prototypes, modify components, and create themes, making it ideal for turning your app ideas into interactive designs in seconds.

1. LLM Red Teaming by Promptfoo: A CLI that helps you test your LLM application for vulnerabilities by generating several hundred adversarial inputs across many categories of potential harm. You can use it to detect potential failures and receive reports with suggested fixes.

2. GenType by Google: Create customized alphabets on different themes and styles with simple text prompts. This is based on Imagen 2 API.

1. Awesome LLM Apps: Build awesome LLM apps using RAG for interacting with data sources like GitHub, Gmail, PDFs, and YouTube videos through simple texts. These apps will let you retrieve information, engage in chat, and extract insights directly from content on these platforms.

Hot Takes 🔥

1. Remember when Scarlett Johansson told OpenAI not to use her voice, but they cloned it and used it anyway? Now imagine what they can do with your phone data, even if you don't allow them to use it. ~DogeDesigner

2. I don’t think people quite realize the significance of this moment, so I will spell it out: OpenAI has been bluffing all along, and their CTO just gave away the fact that they are not holding aces. 75% chance that the GenAI bubble will have burst by 12 months from today. ~Gary Marcus

Meme of the Day 🤡

gpt-5 hope day

That’s all for today! See you tomorrow with more such AI-filled content.

Real-time AI Updates 🚨

⚡️ Follow me on Twitter @Saboo_Shubham for lightning-fast AI updates and never miss what’s trending!

PS: I curate this AI newsletter every day for FREE, your support is what keeps me going. If you find value in what you read, share it with your friends by clicking the share button below!