- Daily AI Skills
- Posts
- What happened at NVIDIA GTC 2025?
What happened at NVIDIA GTC 2025?
PLUS: JP Morgan's Python Training Github Repo
Welcome back to Daily AI Skills.
Here’s what we are covering today:
1. Adobe’s Enterprise Agents
2. Roblox Open Source AI 3D Generator
3. NVIDIA GTC 2025 - AI’s Super Bowl
+ JP Morgan’s Python Training Program
Adobe’s Enterprise Agents
Adobe has introduced a groundbreaking AI agent strategy centered on its new Experience Platform Agent Orchestrator, designed to elevate enterprise-level customer experiences and marketing workflows.

Key Highlights
Agent Orchestration Across Ecosystems – The Agent Orchestrator helps businesses manage AI agents across Adobe and third-party platforms, rooted in a deep understanding of enterprise data, content, and customer journeys.
Specialized AI Agents – Adobe's suite of AI agents covers key areas like audience targeting, content creation, site optimization, and B2B account management—streamlining complex tasks and improving customer engagement.
Brand Concierge for Personalization – The Brand Concierge enables businesses to create immersive and conversational experiences, guiding customers from exploration to confident purchase decisions. AI-driven traffic to retail sites increased by 1,200% in February.
Integration with Microsoft 365 Copilot – The Adobe Marketing Agent now works directly within Microsoft 365 Copilot, giving teams seamless access to Adobe’s AI capabilities inside familiar Microsoft applications.
Generative AI Through Firefly – Adobe is expanding its generative AI platform with Firefly models integrated across creative and marketing workflows, optimizing the content supply chain.
Roblox Open Source AI 3D Object Generator
Roblox has introduced Cube 3D, a new open-source AI model designed to create 3D objects and scenes from text prompts, alongside several enhancements to its AI-assisted game development tools.

Key Highlights:
Direct 3D Object Generation – Cube 3D generates fully functional 3D objects from text prompts, using native 3D data instead of traditional image-based reconstruction.
Simple Commands – Developers can create assets with basic prompts like "/generate motorcycle." Future updates will include the ability to generate assets from image inputs.
3D Tokenization for Shape Prediction – Cube 3D predicts and generates shapes similarly to how language models predict text, setting the foundation for future 4D scene generation.
Studio Enhancements – Roblox Studio has been updated with improved performance, real-time collaboration features, and new monetization tools to support developers.
Roblox announced Cube 3D, an open-source AI for generating 3D objects and scenes from text prompts
Trained on native 3D data, Cube uses 3D tokenization to create functional 3D outputs in seconds
Support for image inputs is also on the way!
— Rowan Cheung (@rowancheung)
6:45 AM • Mar 18, 2025
Jensen Huang Kicks Off Nvidia’s GTC 2025 with Major AI Announcements
Jensen Huang opened Nvidia’s GTC 2025 conference with a two-hour keynote, calling it “AI’s Super Bowl” and unveiling major updates on upcoming chips, robotics, autonomous vehicles, and more.

Key Highlights:
New GPUs: Nvidia announced the Blackwell Ultra coming in late 2025, Vera Rubin in 2026, and Feynman in 2028 — each promising significant performance boosts.
AI Scaling: Huang noted that the computational demand for AI is increasing rapidly, now projected to be 100x higher than last year’s estimates.
Humanoid Robot Model: Nvidia introduced Isaac GR00T N1, the first open foundation model for humanoid robots, alongside a large physical AI training dataset.
AI Workstations: The new DGX Spark and DGX Station deliver data center-level AI computing for personal use — described by Huang as “the computer for the age of AI.”
Robotics Engine: Nvidia unveiled Newton, a robotics physics engine developed with Google DeepMind and Disney — demonstrated using Blue, a Star Wars-inspired robot.
Self-Driving Cars: Nvidia announced a partnership with GM to develop the automaker’s first fleet of autonomous vehicles.
JP Morgan’s Python Training Program
📩Forward it to people you know who are keeping pace with the changing AI world, and stay tuned for the next edition to stay ahead of the curve!