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.

Today:

• Musk Confirms Grok 4 Timeline

• OpenAI Boosts Pay After Meta Poaching

• Meta Continues Talent Raid on OpenAI

• OpenAI Starts Using Google TPUs

• Meta Seeks $29B for AI Data Centres

Elon Musk Goes SCORCHED EARTH! Grok 4, Neuralink plays Call of Duty and Self Delivered Teslas…

Elon Musk says Grok 3.5 is scrapped; Grok 4 should arrive soon after July 4 following one last training run for a coding-focused model. Tesla hits milestones—self-driving car delivery and Neuralink patient controlling games by thought. 

Rival labs push coding agents: Google’s Gemini CLI, Anthropic’s Claude, OpenAI’s $3 billion Windsurf buy. Analysts expect near-term dominance of human-plus-AI “copilot” tools, not fully self-running coding systems, shaping future software work and data wars for teams.

WATCH THE VIDEO ON YOUTUBE

OpenAI Boosts Pay After Meta Poaching

OpenAI is scrambling to keep senior staff after Meta recruited eight top researchers. Chief Research Officer Mark Chen told employees that leadership feels “as if someone broke into our home” and is working nonstop with CEO Sam Altman to keep talent. Steps include raising pay (“recalibrating compensation,” meaning adjusting salaries), personal outreach, and new rewards. Altman earlier suggested Meta dangled $100-million bonuses, a claim Meta disputes. Competition for AI minds is heating fast.

Why this matters

1. Talent tug-of-war: Research breakthroughs follow top scientists; their movement shifts where cutting-edge work happens.
   

2. Rising costs: Bigger pay packages could push smaller labs and startups out of high-level AI research.
   

3. Industry consolidation risks: If a few giants control the best minds, diversity of ideas and open competition in AI may shrink.

Meta Continues Talent Raid on OpenAI

Meta deepens its talent raid on OpenAI, hiring researchers Shengjia Zhao, Jiahui Yu, Shuchao Bi, and Hongyu Ren, days after adding Trapit Bansal and three others. The wave follows a weak launch of Meta’s Llama 4 model, which underperformed on a widely watched test. OpenAI boss Sam Altman says Meta offers “$100-million” bonuses; Meta’s tech chief Andrew Bosworth says the pay deals are more complicated. Competition for researchers is growing.

Why this matters

1. Brain drain shifts innovation. When leading scientists move, the center of cutting-edge research moves with them.
   

2. Soaring pay raises the bar. Huge offers make it harder for smaller labs and startups to hire or keep top talent.
   

3. Pressure on model quality. Meta’s need for better models—and OpenAI’s need to defend its lead—will likely speed up progress but also intensify rivalry.

OpenAI Starts Using Google TPUs

OpenAI has begun renting Google’s AI chips, called tensor processing units (TPUs), to run ChatGPT and other services, easing its dependence on Nvidia graphics processors and Microsoft’s cloud. TPUs could cut inference—model answer-generation—costs, though Google is withholding its newest versions. The move shows surprising cooperation between two competitors as Google seeks outside clients for its hardware and OpenAI looks for cheaper computing power to keep pace with exploding user demand.

Why this matters

1. New chip option. Relying less on Nvidia opens a second supply of the “engines” that run AI, reducing bottlenecks.
   

2. Lower prices. TPUs aim to make inference (using a trained model to answer questions) cheaper, helping startups and users alike.
   

3. Rival teamwork. Direct collaboration between fierce competitors hints at shifting alliances and could speed up overall AI progress.

🧠RESEARCH

MMSearch-R1: Incentivizing LMMs to Search

MMSearch-R1 is a new training method that helps AI models search the internet more efficiently. It teaches the model when and how to search using both images and text. The system performs better than current methods and cuts down unnecessary searches by over 30%, saving time and resources.

RoboTwin 2.0: A Scalable Data Generator and Benchmark with Strong Domain Randomization for Robust Bimanual Robotic Manipulation

RoboTwin 2.0 is a tool that creates large, realistic training data for two-arm robots. It uses AI to simulate real-world tasks and randomizes details like lighting and clutter to improve performance. Robots trained with this data perform better in real-life tests, even without real-world training.

Mind2Web 2: Evaluating Agentic Search with Agent-as-a-Judge

Mind2Web 2 is a new test that measures how well AI agents search the web to answer complex questions. It uses AI judges to check both answer quality and sources. The best system, OpenAI’s Deep Research, reaches up to 70% of human performance while working twice as fast, showing strong potential.

🛠️TOP TOOLS

Dzine AI - Generate and edit in one place with AI-Powered design workflow.

My Vocal AI - Clone Your Voice to Sing, Speak, and Beyond.

HotBot AI - Ask AI questions for expert bot assistance

Looka AI - AI-powered platform to design a logo and brand you love.

GPTExcel - AI-Powered Spreadsheet Automation:Formulas, Charts & Data Insights.

📲SOCIAL MEDIA

🗞️MORE NEWS

• Meta aims to raise $29bn for AI data centres, engaging with private credit firms like Apollo and KKR. Company plans to secure $3bn equity and $26bn debt, exploring structured debt options.

• Baidu will open source its Ernie AI model, challenging U.S. giants by offering powerful tools at lower costs. Experts say this could shake up pricing, trust, and global dominance in the AI race.

• A confidential OpenAI paper on defining AGI has sparked tensions with Microsoft, whose contract limits access if AGI is declared. Microsoft may renegotiate or walk, while OpenAI weighs secrecy, leverage, and timing.

• Runway, known for AI tools in filmmaking, is entering gaming. It plans to let users create video games with AI later this year and is already in talks with game studios to support development and training.

• Over 70 authors signed an open letter urging publishers to reject books made with AI. They argue AI exploits their work without consent and threatens jobs, creativity, and the future of human storytelling in publishing.

• Microsoft is urging employees to use its internal AI tools more actively and may factor AI usage into performance reviews. Leadership has made it clear: “Using AI is no longer optional” for staff.

• Germany has asked Apple and Google to consider banning DeepSeek's AI app, claiming it unlawfully sends user data to China. If upheld, this action could lead to a broader EU-wide ban over privacy violations.