💪NVIDIA ASAP: Humanoid Robots Leap Forward

PLUS: 🌐ChatGPT Search Goes Public: AI Takes on Google!

In partnership with

Welcome to AI Entrepreneurs

In this issue we explore pivotal advancements in AI and robotics. This issue delves into NVIDIA's ASAP, enhancing humanoid robot training, Figure AI's new direction in AI model development, ChatGPT's public search feature challenging Google, ByteDance's OmniHuman-1 for animation, Google's Gemini 2.0, OpenAI's engagement in India, and the latest in AI healthcare innovations.

NVIDIA

Nvidia’s ASAP: Breakthrough in Humanoid Robot Training

Nvidia and Carnegie Mellon University have launched ASAP (Aligning Simulation and Real Physics), a framework that reduces robot motion errors by 53%, allowing for accurate, complex actions like jumps and kicks. It has shown impressive results with the Unitree G1 humanoid robot, mimicking movements of sports legends like Cristiano Ronaldo and Kobe Bryant.

Key Highlights:
  • 53% Reduction in Motion Errors – Enhances real-world accuracy of robotic movements.

  • Dual-Stage Learning Approach – Trains in simulation, then refines movements with real-world data.

  • Human Motion Mimicry – Successfully replicates athletic feats like jumps, kicks, and spins.

  • Open-Source Code – Available on GitHub for further AI and robotics advancements.

  • Potential for Versatile Robotics – Paves the way for agile humanoids in real-world applications.

ASAP operates in two phases: First, it trains robots in a simulated environment using human motion data. Then, a specialized model fine-tunes the robot’s actions by adjusting for real-world differences. This dynamic feedback mechanism ensures precise, fluid movements that closely mimic human motion.

ASAP represents a major leap in robotics, enabling more natural and responsive humanoids for industries like healthcare, manufacturing, and entertainment. While challenges like motor overheating remain, this breakthrough brings us closer to AI-powered humanoid robots that can seamlessly integrate into everyday life.

FIGURE AI

Figure AI Breaks Ties with OpenAI to Build Its Own AI Models

Figure AI, the humanoid robotics startup, has officially ended its partnership with OpenAI, choosing to develop its own AI models instead. Founder Brett Adcock stated that integration challenges led to this decision, emphasizing that solving embodied AI at scale requires vertical integration. He believes outsourcing AI, like hardware, is not an option for building truly intelligent robots.

Source: Figure AI. Edited on Canva

Despite OpenAI being a key investor, Figure AI is moving forward independently, focusing on industrial applications like BMW’s South Carolina plant. The company plans to showcase groundbreaking new capabilities within the next 30 days, hinting at a major leap in humanoid robotics. With OpenAI rebuilding its robotics team, this split could signal a brewing competition in the AI-driven robotics space.

AI BYTES

ChatGPT Search Goes Public: AI Takes on Google

OpenAI has made ChatGPT’s search feature publicly accessible without login, enabling real-time web searches with verified sources. This move positions ChatGPT as a direct competitor to Google, fueling debates on AI-driven search efficiency. While praised for speed and convenience, some users worry about biases and limited exploration compared to traditional search engines.

OmniHuman-1: ByteDance’s AI Brings Images to Life

ByteDance has unveiled OmniHuman-1, an advanced AI framework that transforms still images into lifelike animated videos using text, audio, and body poses. Trained on 19,000 hours of video, OmniHuman-1 excels in generating natural movements and realistic mouth synchronization, surpassing previous models. It can animate real people and cartoon characters, expanding creative possibilities. Its video length is memory-dependent, enabling large-scale AI content creation.

Google Unveils Gemini 2.0: AI Just Got Smarter

Google's Gemini 2.0 series, Flash, Flash-Lite, and Pro Experimental, aims to enhance AI speed, performance, and cost-efficiency. Flash is widely available, Flash-Lite is in public preview, and Pro Experimental targets advanced coding. These models integrate with Google services like YouTube, Search, and Maps. Available via Google AI Studio and Vertex AI, Gemini 2.0 is seen as a game-changer in AI applications.

OpenAI and India: A Tech Collaboration

During his visit to India, OpenAI CEO Sam Altman met with top officials, including Minister Ashwini Vaishnaw, to underscore India's rising prominence in the AI space. India's user base has tripled in the past year, making it a key player in global AI innovation. A notable discussion during the visit focused on collaborating with India on the AI stack, including GPUs, models, and apps. Talks also aimed at creating a comprehensive AI ecosystem to reduce costs and boost adoption, marking a new era of tech partnership between OpenAI and India.

AI HEALTHTECH INSIDER - Issue #34

🚀 AI Breakthroughs in Medical Imaging & Smart Prescriptions

In Issue # 34, we explore how AI is changing healthcare. Companies like NVIDIA and Siemens are improving medical imaging, and AI tools are helping doctors prescribe antibiotics more effectively. This issue also includes important updates from the HIMSS25 conference and news about Rad AI's recent funding.

Stay ahead with the latest insights shaping the future of medicine—subscribe now!

Interested in AIHealthTech Insider?

Are you interested in receiving the AIHealthTech Insider newsletter directly to your inbox? Stay updated on the latest AI-driven healthcare innovations.

Login or Subscribe to participate in polls.

ADVERTISEMENT: THE RUNDOWN AI

Start learning AI in 2025

Everyone talks about AI, but no one has the time to learn it. So, we found the easiest way to learn AI in as little time as possible: The Rundown AI.

It's a free AI newsletter that keeps you up-to-date on the latest AI news, and teaches you how to apply it in just 5 minutes a day.

Plus, complete the quiz after signing up and they’ll recommend the best AI tools, guides, and courses – tailored to your needs.

AI CREATIVITY

Midjourney’s Moodboards Just Got a Major Upgrade!

Midjourney has rolled out two exciting new features for moodboards, making creative image generation even more powerful. You can now combine multiple moodboard codes and mix them with sref codes for enhanced control over your aesthetic direction.

How to Use the New Moodboard Features:
1️⃣ Access the Personalize Page – Head to Midjourney’s website to start.
2️⃣ Upload & Select Images – Use personal uploads, URLs, or Midjourney’s gallery.
3️⃣ Mix & Match Codes – Blend different moodboard codes with sref codes for greater creative flexibility.

Ready to level up your AI-generated imagery? Try the new moodboard features today and unleash your creativity!

GIVEAWAY

🎁$2000 in AI Tools, Courses & Templates— For FREE

We’re giving away over $2,000 worth of AI software, courses, and templates to help you 10X your personal and business goals.

This giveaway includes a free business consultation, exclusive tools like MegaSEO, and the $1 Product Challenge Course to launch profitable products without spending on ads. Plus, get access to expert-led workshops, hundreds of templates, and much more. Don’t miss out—this opportunity is 100% free, but only for a limited time!

Reply

or to participate.