• Smarter with AI
  • Posts
  • SunBrief#48: Musk’s xAI Sues OpenAI Over Alleged Trade Secret Theft

SunBrief#48: Musk’s xAI Sues OpenAI Over Alleged Trade Secret Theft

OpenAI touts GPT-5’s expert-level skills as Google DeepMind debuts a “thinking” robotics AI

September 28, 2025

Smarter with AI banner

Welcome to the SunBrief

Today in SunBrief 🌞

  • 🚀 Revolutionize Learning with AI-Powered Video Guides

  • Elon Musk’s xAI Sues OpenAI Over Alleged Trade Secret Theft

  • Google DeepMind Unveils First “Thinking” Robotics AI

  • Stock Updates

  • OpenAI Says GPT-5 Rivals Human Experts in Wide Range of Jobs

  • AI Highlights of the Week

  • Too Important to Miss

🚀 Revolutionize Learning with AI-Powered Video Guides

Sponsored

Upgrade your organization training with engaging, interactive video content powered by Guidde.

Here’s what you’ll love about it:

1️⃣ Fast & Simple Creation: AI transforms text into video in moments.
2️⃣ Easily Editable: Update videos as fast as your processes evolve.
3️⃣ Language-Ready: Reach every learner with guides in their native tongue.

Bring your training materials to life.

The best part? The browser extension is 100% free.

Elon Musk’s xAI Sues OpenAI Over Alleged Trade Secret Theft

Lawsuit Claims OpenAI Poached Staff to Access Secrets Behind Grok Chatbot

Elon Musk’s xAI is suing OpenAI, accusing it of stealing trade secrets by poaching staff to gain insights into its Grok chatbot and operations.

Key Points:

  • Allegations of Poaching: xAI claims OpenAI recruited former staff, including engineers and a senior finance executive, to gain access to confidential technology and business plans.

  • Trade Secret Concerns: The suit alleges OpenAI sought insights into xAI’s source code and data center operations, inducing staff to breach confidentiality agreements.

  • OpenAI Response: OpenAI denied the claims, calling the suit part of Musk’s “ongoing harassment” and reiterating it has no interest in competitors’ trade secrets.

  • Broader Legal Battles: Musk has recently filed other lawsuits against OpenAI and Apple over alleged anticompetitive conduct, reflecting intensifying competition in the AI sector.

Why It Matters:

The lawsuit highlights Silicon Valley’s fierce fight for AI talent and IP, with Musk using legal battles to position xAI against OpenAI and other tech giants.

Would you trust AI developed by companies engaged in frequent lawsuits?

Login or Subscribe to participate in polls.

Google DeepMind Unveils First “Thinking” Robotics AI

Gemini Robotics Models Aim to Bring Agentic Reasoning to Robots

Google DeepMind’s Gemini Robotics 1.5 and ER 1.5 aim to create agentic robots that can “think” before acting, adapting to new tasks without heavy reprogramming.

Key Points:

  • Two-Model System: Gemini Robotics-ER 1.5 is a vision-language reasoning model that breaks down tasks into step-by-step instructions, while Gemini Robotics 1.5 translates those plans into physical actions.

  • Simulated Reasoning: The ER model introduces reasoning similar to AI chatbots, enabling robots to plan tasks like sorting laundry before execution.

  • Cross-Robot Learning: Skills learned on one robot, such as Aloha 2’s grippers, can transfer to humanoid robots like Apollo without special tuning.

  • Built on Gemini Foundation: Both models are fine-tuned from Google’s Gemini AI, adapted for physical environments to handle complex multi-stage tasks.

Why It Matters:

DeepMind’s Gemini Robotics combines reasoning and action, paving the way for adaptable agentic robots that could transform automation, manufacturing, and human-robot collaboration.

How do you feel about robots that can ‘think’ before acting?

Login or Subscribe to participate in polls.

Stock Updates

OpenAI Says GPT-5 Rivals Human Experts in Wide Range of Jobs

New GDPval Benchmark Shows AI Approaching Human-Level Work Across Key Industries

OpenAI’s new GDPval benchmark shows GPT-5 starting to match human experts across key industries, signaling progress toward real-world, professional-level AI performance.

Key Points:

  • Benchmark Overview: GDPval tests AI performance in 44 occupations across nine major industries, including healthcare, finance, manufacturing, and government.

  • Performance Results: GPT-5-high, a boosted version of GPT-5, was rated on par with or better than human experts 40.6% of the time. Anthropic’s Claude Opus 4.1 scored higher at 49%, though OpenAI attributed this partly to its graphic-heavy outputs.

  • Limitations: GDPval-v0 only tests AI on research-style reports, not the full range of professional tasks, which OpenAI acknowledges. Future versions aim to include more industries and interactive workflows.

Why It Matters:

GDPval offers a fresh way to measure AI’s economic value, with GPT-5’s rising performance showing how quickly AI is moving toward AGI-level capabilities.

Do you believe GPT-5 matching human experts in some jobs is a major milestone?

Login or Subscribe to participate in polls.

AI Highlights of the Week

  • Google AI Plus Expands to 40 More Countries

    Google has expanded its AI Plus plan to 40 more countries after a successful launch in Indonesia. It includes tools like Nano Banana (image editing), Veo 3 Fast (video generation), and Gemini in Gmail and Docs.

    Users also get 200 GB storage and can share the plan with five family members. Pricing varies by country.

  • Spotify Cracks Down on AI Music, Spam, and Voice Clones

    Spotify is cracking down on AI-generated music, focusing on impersonation, spam, and transparency. It’s building an AI disclosure standard with DDEX.

    A new spam filter targets voice clones and fake uploads. Spotify says it removed 75M spam tracks and denies adding AI music to playlists.

  • UK’s AI Tool Recovers £500M in Fraud – Now Going Global

    A new AI tool helped the UK recover £480M in fraud, including £186M from Covid schemes.

    The tool, which detects policy weak spots, will be licensed abroad to countries like the US and Australia. But campaigners warn of potential AI bias and privacy issues.

  • OpenAI Eyes Smart Gadgets, Poaches Apple Talent and Suppliers

    Trump approved a $14B deal to keep TikTok in the U.S., with Oracle and U.S. investors taking control. ByteDance retains under 20%.

    Oracle will manage app security. China’s approval is still needed. Enforcement of a potential ban is delayed until Dec. 16.

Too Important to Miss

Last Week’s Poll Result

  • Do you see AI-designed genomes as a positive step for biotechnology?

    Yes, 39%, Maybe, 50%, No Sure, 11%.

  • Do you think age-aware experiences in ChatGPT will make AI safer for teens?

    Yes, 33.3%, Maybe, 55.6%. No, 11.1%.

  • Do you think Gemini in Chrome will improve your daily browsing experience?

    Yes, 70.6%, Maybe, 17.6%. No, 11.8%.

Feedback

We’d love to hear from you!

How did you feel about today's SunBrief? Your feedback helps us improve and deliver the best possible content.

Login or Subscribe to participate in polls.

Know someone who may be interested?

And that's a wrap on today’s SunBrief!

Reply

or to participate.