Here’s Why Over 4 Million Professionals Read Morning Brew

• Business news explained in plain English

• Straight facts, zero fluff, & plenty of puns

• 100% free

See for yourself

Today:

• Altman Predicts AI Coding Revolution

• OpenAI Nears Launch of AI Chip

• Anthropic Unveils AI Labor Tracker

• Groq Secures $1.5B Saudi Investment

• Intel Loses AI Head to Nokia

• Hugging Face Unveils Powerful SmolLM2

Sam Altman "MILLIONS of Software Engineering Agents" and "AGI in sight"

Sam Altman outlines rapid progress in artificial intelligence, predicting that upcoming coding systems will soon surpass human abilities. Early models struggled with basic tasks, but recent improvements suggest future AI will perform complex work with superhuman efficiency. 

Instead of a single powerful system, many specialized, affordable agents may operate like virtual colleagues. While these advancements promise economic and creative benefits, they demand careful integration to ensure safety and widespread gains.

WATCH THE VIDEO ON YOUTUBE

OpenAI Nears Launch of AI Chip

OpenAI is set to launch its in-house AI chip in 2026, finalizing its design this year before sending it to TSMC for production. The chip will use advanced 3-nanometer technology with high-bandwidth memory and networking capabilities. Initially deployed on a limited scale, future versions will feature enhanced processing power. OpenAI’s chip effort, led by ex-Google engineer Richard Ho, aims to reduce reliance on Nvidia and improve AI infrastructure efficiency.

Anthropic Unveils AI Labor Tracker

Anthropic has launched the Anthropic Economic Index to track AI's impact on labor markets using real-world data from Claude.ai. Their research reveals AI is primarily used for software development and technical writing, with 57% of tasks augmenting human work and 43% automating it. AI adoption is highest in mid-to-high wage jobs but remains limited in low-wage and highly specialized professions. The dataset is open for further research to inform policy decisions.

Groq Secures $1.5B Saudi Investment

AI chip startup Groq has secured a $1.5 billion commitment from Saudi Arabia to expand its advanced AI chip production. The funding will support the expansion of Groq’s data center in Dammam, enabling the deployment of AI inference chips optimized for speed. Groq, backed by Cisco, Samsung Catalyst, and BlackRock, has obtained U.S. licenses for exports. The announcement was made at LEAP 2025, where Saudi Arabia secured $14.9 billion in AI investments.

Groq Secures $1.5B Saudi Investment

AI chip startup Groq has secured a $1.5 billion commitment from Saudi Arabia to expand its advanced AI chip production. The funding will support the expansion of Groq’s data center in Dammam, enabling the deployment of AI inference chips optimized for speed. Groq, backed by Cisco, Samsung Catalyst, and BlackRock, has obtained U.S. licenses for exports. The announcement was made at LEAP 2025, where Saudi Arabia secured $14.9 billion in AI investments.

Hugging Face Unveils Powerful SmolLM2

Hugging Face has launched SmolLM2, a small but efficient language model that outperforms Meta’s Llama and Qwen in several benchmarks. Trained on an 11-trillion token dataset, it combines web content, coding examples, and specialized datasets. The model was refined using instruction fine-tuning, example-based learning, and reinforcement learning. Available in three sizes, SmolLM2 is optimized for smaller devices, reinforcing Hugging Face’s commitment to open-source AI development.

🧠RESEARCH

Goku: Flow Based Video Generative Foundation Models

Goku is a cutting-edge AI model for generating images and videos using rectified flow Transformers. It excels in quality and efficiency, setting new industry benchmarks. The paper details its data pipeline, architecture, and training methods. Goku scores 0.76 on GenEval, 83.65 on DPG-Bench, and 84.85 on VBench.

VideoRoPE: What Makes for Good Video Rotary Position Embedding?

VideoRoPE improves video-based Rotary Position Embedding (RoPE) by addressing spatio-temporal challenges. It introduces a 3D structure with optimized temporal allocation and spatial symmetry. A new benchmark, V-NIAH-D, highlights previous RoPE limitations. VideoRoPE outperforms existing models in video retrieval, understanding, and hallucination, advancing long-context video AI research.

Fast Video Generation with Sliding Tile Attention

Sliding Tile Attention (STA) enhances video generation by reducing the high computational cost of Diffusion Transformers. It replaces full attention with a localized 3D sliding approach, cutting redundancy while maintaining quality. STA speeds up processing by 2.8-17x, reducing latency from 945s to 268s on HunyuanVideo with minimal performance loss.

🛠️TOP TOOLS

Text2SQL AI - AI-powered platform that transforms natural language instructions into SQL queries, Excel formulas, and regex expressions.

Craiyon - AI-powered image generation tool that transforms textual descriptions into unique visual art.

Supercreator AI  - Revolutionize short-form video creation for platforms like TikTok, Instagram Reels, and YouTube Shorts.

Revocalize AI - Voice generation and transformation toolkit that leverages artificial intelligence to create hyper-realistic vocal content.

Boo AI - AI-powered writing assistant that leverages GPT-3 technology to offer a distraction-free writing experience.

📲SOCIAL MEDIA

🗞️MORE NEWS

• OpenAI's $14 million Super Bowl ad showcased AI as a transformative tool, tracing humanity’s technological evolution. The campaign avoided AGI mentions, emphasizing accessibility and creativity while competing for market dominance amid growing public debate on AI’s impact.

• France is leveraging its nuclear power to boost AI development, pledging a gigawatt for a massive computing project. The initiative, backed by private investors, aims to position France as a global AI leader.

• OpenAI has informed government officials about its investigation into DeepSeek for allegedly misusing its AI model outputs. Critics call this hypocritical, given OpenAI’s own legal battles over training on copyrighted material.

• French AI startup Mistral is investing billions to build its own data center in France, reducing reliance on Big Tech. The facility, powered by decarbonized energy, aims to strengthen Mistral’s independence and enhance AI model development.

• Thousands of artists are demanding Christie’s cancel its AI-generated art auction, accusing the technology of “mass theft.” They argue AI models were trained on copyrighted works without permission, while Christie’s defends the artists’ use of AI tools.

• Meta has begun performance-based layoffs, cutting 5% of its workforce while prioritizing AI talent. CEO Mark Zuckerberg emphasizes a shift towards emerging technologies like machine learning, reflecting broader industry trends in strategic hiring and workforce restructuring.