OpenAI’s o3 Crushes Grok at Chess

Welcome to another edition of AI PlanetX.

OpenAI o3 sweeps Grok in AI chess clash; Nvidia unveils Cosmos Reason for robot planning; Microsoft launches Copilot 3D for rapid image-to-3D model creation.

• Hottest AI News

• Top AI & SaaS Tools

• AI Tutorial: Upscale Videos with Krea

• Top AI & Tech News

• AI Art Spotlight

• Prompt of the Day: Achievement Architect Prompt

• AI Video Tutorial

• F-R-E-E AI Course of the Day: LLM Course by Hugging Face

Hottest AI News

OpenAI

OpenAI o3 Sweeps Grok in Chaotic AI Chess Showdown

Sam Altman’s OpenAI o3 crushed Elon Musk’s Grok 4-0 in Google’s Kaggle Game Arena AI Chess Exhibition, but the real story was the poor play. Instead of showcasing cutting-edge AI, it became a display of digital blunders, with Magnus Carlsen likening both bots to beginners barely knowing the rules.

Details:

• From August 5-7, chatbots played chess with no special training—and it was a disaster. Carlsen rated them 800 ELO, likening it to “learning chess from a corrupted PDF.” Grok blundered pieces, misplayed openings, and marched its king into danger

• Grok’s collapse was stark after earlier praise from Nakamura. Musk had said chess was just a “side effect” for Grok—accurate. Even after o3 blundered, Grok’s endgame fell apart, giving o3 the sweep

• AIs needed four tries to make legal moves, often teleporting pieces or reviving pawns. Gemini claimed third after both sides in the bronze match failed to checkmate. Carlsen said they could collect pieces but not cook a win

These are the same AIs touted as near-human intelligence, yet they can’t play a 1,500-year-old game without cheating or forgetting rules—proof we’re safe from an AI takeover, at least until they learn how pawns move.

Practical AI for Business Leaders

The AI Report is the #1 daily read for professionals who want to lead with AI, not get left behind.

You’ll get clear, jargon-free insights you can apply across your business—without needing to be technical.

400,000+ leaders are already subscribed.

👉 Join now and work smarter with AI.

Stay ahead. Stay informed.

Nvidia

Nvidia Debuts Cosmos Reason to Boost Robot Planning and Physical AI

Nvidia is making a big move into robotics with advanced AI models and infrastructure unveiled at SIGGRAPH, aiming to power the next wave of physical AI and expand beyond traditional data centers into embodied intelligence.

Details:

• Cosmos Reason, a 7B-parameter vision language model for physical AI, gives robots memory, physics understanding, and decision-making skills. It supports data curation, planning, and video analytics for smarter real-world interaction

• Nvidia’s Cosmos Transfer-2 speeds synthetic data generation from 3D simulations, with a quicker distilled version. Together, they create text, image, and video datasets for training robots. Neural reconstruction libraries simulate 3D environments from sensor data, integrated with CARLA and the Omniverse SDK

• On hardware, Nvidia launched the RTX Pro Blackwell Server for unified robotics development and DGX Cloud for managing robotics workflows, built to handle intensive AI compute

Nvidia’s robotics push marks a strategic move beyond AI data centers, aiming at embodied AI and real-world applications as the next big GPU market.

Top AI & SaaS Tools

• Release0 (Life-time Deal): No‑code AI chatbot and agent platform for automated, personalized support and lead capture, with a visual builder, multichannel deployment, real‑time sentiment analytics

• Grok: xAI, Elon Musk’s latest AI model—and one of the top LLMs across multiple benchmarks—is currently available at no cost with usage limits [F-R-E-E]

• Nas io: A 24/7 “co‑founder” that accelerates launching and running a business—turning ideas into offerings, creating ads, finding leads, managing lists, and handling finances [F-R-E-E]

• Hera: Transform your animation prompts into professional, code-based animations you can fine-tune

• GLM-4.5V: Open-source visual reasoning model that sets the bar for its size, outperforming peers across 41 benchmarks (select it from left top) [F-R-E-E]

Fact-based news without bias awaits. Make 1440 your choice today.

Overwhelmed by biased news? Cut through the clutter and get straight facts with your daily 1440 digest. From politics to sports, join millions who start their day informed.

Sign up now!

AI Tutorial

How to Upscale Grainy Videos with Krea AI in Minutes

Krea AI’s video upscaler is designed to breathe life into grainy or low-resolution videos, producing sharp, clear visuals in minutes. Here’s how to make the most of it:

1. Visit Krea AI

   Go to Krea and either create an account or log into your existing one.

2. Open the Enhancer Tool

   From the dashboard, locate and select the Enhancer feature—this is where all the video upscaling options are available.

3. Choose AI Model

   At the lower left corner, select the AI model you want to use for upscaling. Krea offers multiple options—such as Krea Enhance, Upscale V1, and Topaz—each with unique strengths in quality, duration, and processing speed.

4. Upload Video Clip

   Pick a video you’d like to improve. The tool works best with clips that are concise enough for rapid processing.

5. Adjust Enhancement Settings

   Fine-tune your output by modifying parameters such as:

   • Upscaling ratio (e.g., 2x, 4x)

   • Frame rate (to make motion smoother)

   • Intensity (for sharper detail)

   You can also include an optional text prompt to influence the style or tone of the output.

6. Process the Video

   Select the Enhance button to start rendering. Processing time will depend on your chosen settings and video length.

7. Compare and Review

   Once complete, view the side-by-side comparison of the original and enhanced versions to ensure you’re happy with the result.

8. Download Improved Footage

   If satisfied, save the enhanced video to your device for use in projects, presentations, or social media.

Tip: Start with the highest-quality source video you have—even if it’s still low-res. AI upscalers work best when there’s enough base detail for the model to enhance.

Top AI & Tech News

• Microsoft released Copilot 3D, an AI tool that converts images into usable 3D models in a single step for integrations with games, animation, VR/AR, and more

• Apple is testing a more capable Siri that can perform actions inside third‑party apps using voice commands

• Nvidia and AMD have reportedly agreed to pay the U.S. government 15% of revenue from sales of certain high-end AI chips to China in exchange for export licenses

• GitHub CEO Thomas Dohmke resigned, and GitHub will be absorbed into Microsoft’s CoreAI group led by Jay Parikh, focused on AI platforms and tools

AI Art Spotlight

Model: Midjourney V7

Prompt:

Realistic photography, first - person POV, riding on dragon's back, holding reins with both hands. Below is bustling Magic City landscape. Oriental Pearl Tower has strong metallic texture, its glass curtain wall reflects bright light in sun. Crowds are vaguely visible on different floors' observation decks. Surrounding are tall modern skyscrapers with various materials like smooth stone and reflective glass on exterior walls. Traffic between buildings is like metallic rivers shuttling. Shiny dragon scales, no horns on head, dragon wings visible at edge of vision. Grand and full of epic feeling scene. 8K resolution, hyper - realistic photo, cinematic lighting, sharp details. No dragon body allowed, highlighting first - person POV and hands must be visible. --ar 9:16 --raw --v 7

Prompt of the Day

The Achievement Architect Prompt

This prompt turns a powerful dream into a concrete execution plan using the INEVITABLE System, guiding you from vision to daily action with precision. It helps you extract your true goal, reverse-engineer the end state, and map clear milestones across years, quarters, months, weeks, and days. You’ll build capability, habits, and accountability that remove guesswork and excuses, replacing hope with a structured pathway.

Top AI Video Tutorial

You’re Not Behind (Yet): How to Learn AI in 29 Minutes

Complimentary AI Course of the Day

LLM Course by Hugging Face

Hands‑on course on modern NLP and LLMs with Hugging Face (Transformers, Datasets, Tokenizers, Accelerate) and the Hub. Learn core Transformer concepts, use and fine‑tune Hub models, build data/tokenization pipelines, tackle sentiment analysis, NER, classification, translation, and publish an interactive demo.