🚀 xAI launches advanced Grok 3 model. Elon Musk's AI company, xAI, has released Grok 3, its latest AI model trained with ten times more computing power than its predecessor, utilizing a massive data center in Memphis equipped with approximately 200,000 GPUs. Grok 3 is designed to analyze images, respond to questions, and powers features on Musk's social network, X. xAI claims Grok 3 outperforms models like OpenAI's GPT-4o and DeepSeek's R1 on benchmarks such as AIME and GPQA. Read more
🔍 Leak reveals Claude 3.7 Sonnet details. Uncovered code from AWS Bedrock's website references Anthropic's upcoming AI model, Claude 3.7 Sonnet, ahead of its official announcement. The model introduces "extended thinking" for complex problem-solving and allows users to switch between rapid and in-depth responses. Designed for applications like retrieval-augmented generation and advanced code generation, it offers flexibility based on use case requirements. Read more
🎬 Cameron's anti-AI stance in Avatar 3. Director James Cameron plans to open "Avatar: Fire and Ash" with a title card stating that no generative AI was used in its production. Cameron has previously criticized AI-written scripts, emphasizing the irreplaceable value of human experience in storytelling, despite his role on the board of foundation AI company Stability AI. Read more
🛠 New coding benchmark for AI. OpenAI introduces SWE-Lancer, a benchmark assessing AI performance on freelance software engineering tasks. Based on 1,400+ real-world projects from Upwork and Expensify, it evaluates code quality, managerial decision-making, and economic impact. Read more
Jake’s Take: Microsoft's BioEmu-1 is a deep learning model that predicts the various shapes a protein can take, offering a dynamic view beyond static structures. Trained on extensive datasets, it efficiently generates thousands of protein structures per hour on a single GPU (significantly faster than existing/traditional methods).
This (open-source!) tool will help expand our knowledge of proteins, aiding in new drug discovery and protein engineering in general.
what to know for later
🏭 Apple's $500 billion U.S. investment. Apple plans to invest $500 billion in U.S. operations over the next four years, including a 250,000-square-foot AI server manufacturing facility in Houston, Texas, set to open in 2026. This facility will produce servers for Apple Intelligence, the company's AI personal assistant. The investment also includes hiring approximately 20,000 new employees focused on R&D, silicon engineering, software development, and AI. Read more
💰 Alibaba commits $53 billion to AI. Over the next three years, Alibaba Group plans to invest 380 billion yuan ($53 billion) in cloud computing and artificial intelligence infrastructure, surpassing its total AI and cloud spending over the past decade. Read more
🤖 Figure AI previews new Helix system. The Vision-Language-Action model allows Helix to understand speech, reason, and interact with objects without explicit programming. It controls a humanoid upper body with precision, executes real-time tasks, and even coordinates multiple robots simultaneously. Read more
🚀 Ex-OpenAI CTO launches AI startup. Mira Murati, former Chief Technology Officer at OpenAI, has launched Thinking Machines Lab, a new AI research and product company. The startup aims to develop advanced, customizable AI systems and promote open science by sharing technical blogs, research papers, and code. Murati has assembled a team of experts, including John Schulman as Chief Scientist and Barret Zoph as Chief Technology Officer. Read more
