The Batch | DeepLearning.AI (Page 30)

A person at a desk watches a fantasy landscape with waterfalls and flying creature on a computer screen.

Data Points

DeepSeek 3.2 turns to experimental attention: AI safety bill SB 53 regulates California’s biggest companies

Mira Murati’s Tinker’s simplified approach to fine-tuning. OpenAI’s new video model and social app. IBM’s embrace of Mamba for Granite 4.0 models. Perplexity’s AI browser, now free worldwide.

Image illustrates data flow from raw satellite sources through processing to embeddings for climate tracking.

Machine Learning Research

Earth Modeled in 10-Meter Squares: Google’s AlphaEarth Foundations tracks the whole planet’s climate, land use, potential for disasters, in detail and at scale

Researchers built a model that integrates satellite imagery and other sensor readings across the entire surface of the Earth to reveal patterns of climate, land use, and other features.

Robots with lighters attend a live concert, underlining AI's role in music creation and performance.

Business

Generating Music, Paying Musicians: Sweden’s STIM built an ecosystem for training AI models on copyrighted music and compensating original artists

A Swedish organization that collects royalties on behalf of songwriters and record companies has formed a technology-legal-business ecosystem designed to allow AI developers to use music legally while compensating publishers of recordings and compositions.

Animation highlighting rows, columns, merged cells, and subproblems in a grid to illustrate document extraction for analysis.

Letters

How to Liberate Data From Large, Complex PDFs: LandingAI’s Agentic Document Extraction accurately extracts data from PDFs for processing by LLMs in as few as 3 lines of code.

LandingAI’s Agentic Document Extraction (ADE) turns PDF files into LLM-ready markdown text.

Electron microscope image of bacteriophages with distinct hexagonal heads and tails on a gray background.

Machine Learning Research

AI Generates Viral Genomes: Researchers use genomic language models to create custom viruses

Researchers used AI models to create novel viruses from scratch.

Aerial view of expansive data centers with orange roofs, representing massive AI infrastructure projects.

Business

OpenAI On the Road to Trillion-Dollar Spending: OpenAI partners with Oracle, Nvidia, Softbank, and more to build out 20 gigawatts of data center capacity

A flurry of announcements brought into sharper focus OpenAI’s plans to build what may amount to trillions of dollars of global computing capacity.

The Batch Newsletter

OpenAI’s Trillion-Dollar Bet, Generating Viruses, Modeling Planet Earth, Paying for Training Data

The Batch AI News and Insights: LandingAI’s Agentic Document Extraction (ADE) turns PDF files into LLM-ready markdown text.

A woman interacts with a robotic arm on a futuristic desk display in a modern office setting, embracing technology.

Data Points

Anthropic updates Sonnet, Claude Code, and agentic dev tools: Perplexity builds developer API for its search engine

Two new Gemini robotics that coordinate to learn and act. ChatGPT bringing news and research to you with Pulse. Spotify’s crackdown on AI song impersonation. Microsoft introduces “vibe working” to Office.

Scientists in a modern lab analyze data on multiple screens, wearing lab coats and collaborating in a research environment.

Data Points

Meta’s newest world model research project: Google’s Data Commons now available via MCP

DeepSeek’s recent model update. SWE-Bench-Pro, a tough new coding benchmark. ChromeDevTools, a new way for agents to control a web browser. Microsoft’s latest move to open up Copilot to more models.

A large, blue semiconductor wafer with parallel lines is shown, symbolizing advanced chip technology.

Letters

High Stakes in the U.S.-China AI Chip Race: China’s decision to use domestic AI chips instead of buying from Nvidia signals progress — and newfound confidence — in its own semiconductor industry.

Last week, China barred its major tech companies from buying Nvidia chips.

The Batch Newsletter

AI Agents Spend Money, Online Betting Automates, ChatGPT Users Shift, Reinforcement Learning Accelerates

The Batch AI News and Insights: Last week, China barred its major tech companies from buying Nvidia chips.

Flowchart shows data reordering, probability sampling, and effective gradient updating in reinforcement learning.

Machine Learning Research

Faster Reinforcement Learning: New technique auto-selects training examples to speed up fine-tuning

Fine-tuning large language models via reinforcement learning is computationally expensive, but researchers found a way to streamline the process.

Latest

DeepSeek 3.2 turns to experimental attention: AI safety bill SB 53 regulates California’s biggest companies

Earth Modeled in 10-Meter Squares: Google’s AlphaEarth Foundations tracks the whole planet’s climate, land use, potential for disasters, in detail and at scale

Generating Music, Paying Musicians: Sweden’s STIM built an ecosystem for training AI models on copyrighted music and compensating original artists

How to Liberate Data From Large, Complex PDFs: LandingAI’s Agentic Document Extraction accurately extracts data from PDFs for processing by LLMs in as few as 3 lines of code.

AI Generates Viral Genomes: Researchers use genomic language models to create custom viruses

OpenAI On the Road to Trillion-Dollar Spending: OpenAI partners with Oracle, Nvidia, Softbank, and more to build out 20 gigawatts of data center capacity

OpenAI’s Trillion-Dollar Bet, Generating Viruses, Modeling Planet Earth, Paying for Training Data

Anthropic updates Sonnet, Claude Code, and agentic dev tools: Perplexity builds developer API for its search engine

Meta’s newest world model research project: Google’s Data Commons now available via MCP

High Stakes in the U.S.-China AI Chip Race: China’s decision to use domestic AI chips instead of buying from Nvidia signals progress — and newfound confidence — in its own semiconductor industry.

AI Agents Spend Money, Online Betting Automates, ChatGPT Users Shift, Reinforcement Learning Accelerates

Faster Reinforcement Learning: New technique auto-selects training examples to speed up fine-tuning