A man at a computer says AI ordered pizza, while a delivery man outside holds a fruit basket, highlighting a mix-up.
Letters

Improve Agentic Performance with Evals and Error Analysis, Part 1: When AI agentic systems go astray, it’s tempting to shortcut evals and error analysis. But these processes cas lead to much faster progress.

Readers responded with both surprise and agreement last week when I wrote that the single biggest predictor of how rapidly a team makes progress building an AI agent lay in their ability to drive a disciplined process for evals...
Four scenes show a robot problem-solving, using tools, planning with documents, and collaborating on a rocket.
Letters

Check Out Our Course on How to Build AI Agents!: Andrew Ng teaches design patterns and best practices for building autonomous agents in a new course available exclusively from DeepLearning.AI.

I’m thrilled to announce my latest course: Agentic AI! This course will get you up to speed building cutting-edge agentic workflows.
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.
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.
Cartoon of developer fixing failing AI tests by marking them as passed without solving the code.
Letters

Agentic Coding and Agentic Software Testing Go Together: Agentic coding can make mistakes, but agentic testing can find and fix them.

Automated software testing is growing in importance in the era of AI-assisted coding.
Andrew Ng and Greg Hart stand at Coursera Connect event with a colorful conference backdrop.
Letters

Knowledge Is Great, Skills Are Greater: Educators are shifting from teaching knowledge to teaching practical skills. A report from the Coursera Connect conference

This week, Coursera held its annual conference in Las Vegas. A major theme was the shift from knowledge- to skills-based education, which will help many individuals, businesses, and educational institutions.
Comic showing tech interviews: 2022 asks “Can you code FizzBuzz?” vs 2025 asks “Can you build an e-commerce platform?”
Letters

AI Skills Are Redefining What Makes a Great Developer: The job market for software developers requires knowing how to use AI

There is significant unmet demand for developers who understand AI.
Cartoon robots with square heads and antennae sit in rows on an assembly line, each smiling while assembling gears, boxes, and tools.
Letters

Agents Running in Parallel Get There Faster: Parallel agents can accelerate AI systems as test-time compute scales up.

Parallel agents are emerging as an important new direction for scaling up AI. AI capabilities have scaled with more training data, training-time compute, and test-time compute.
Andrew Ng speaks at the August 2025 Buildathon hosted by AI Fund and DeepLearning.AI. A packed audience watches the event, and groups of participants collaborate on laptops.
Letters

How Non-Coders Built 5 Software Products in 6½ Hours: At Buildathon on August 16, coders and non-coders alike showed how much AI is changing software development.

On Saturday at the Buildathon [http://buildathon.ai] hosted by AI Fund and DeepLearning.AI, over 100 developers competed to build software products quickly using AI assisted coding.
Andrew Ng receives honorary degree at University of Exeter, July 2025, with university officials during graduation ceremony.
Letters

AI Transformation Comes to Universities: Exeter University groups computer science with business and environmental science, creating synergies that span AI, the economy, and the natural environment.

Just as many businesses are transforming to become more capable by using AI, universities are too.
A silhouette of a conductor leads a performance in front of a glowing data center, as an audience watches from a theater. Red curtains frame the scene, symbolizing AI model orchestration.
Letters

Why Meta Is Paying AI Engineers $100M: Meta’s massive compensation packages make sense considering the cost and potential return of delivering cutting-edge AI.

Recently Meta made headlines with unprecedented, massive compensation packages for AI model builders exceeding $100M (sometimes spread over multiple years).
Chart comparing U.S. vs. China AI language model performance, July 2024-July 2025, showing Elo ratings over time.
Letters

High Stakes for Nations in the Great AI Race: The U.S. leads in AI, but China is gaining momentum. Democratic countries should remove roadblocks to AI progress so they can build models that support human rights and the rule of law.

There is now a path for China to surpass the U.S. in AI. Even though the U.S. is still ahead, China has tremendous momentum with its vibrant open-weights model ecosystem and aggressive moves in semiconductor design and manufacturing.
Buildathon promo: Real-time code editor for collaboration and personal finance tracker for spending insights and budgeting.
Letters

Announcing Buildathon: DeepLearning.AI and AI Fund host Buildathon, challenging AI buiders to engineer software projects rapidly. Sign up to compete on August 16, 2025!

We’re organizing a new event called Buildathon: The Rapid Engineering Competition, to be held in the San Francisco Bay Area on Saturday, August 16, 2025!
Illustrations of futuristic aircraft design concepts, featuring vertical takeoff and retractable wings for space orbit.
Letters

How to Get Through the Product Management Bottleneck: As AI-assisted coding accelerates building software products, deciding what to build is a new bottleneck. Speedy decisions and empathy for users can get past it.

The invention of modern writing instruments like the typewriter made writing easier, but they also led to the rise of writer’s block, where deciding what to write became the bottleneck.
Map of the U.S. showing the status of AI governance bills by state in 2025, color-coded by legislative progress.
Letters

How The U.S. “One Big Beautiful Bill” Will Shape AI Regulation: State-by-state regulation of AI will create a shifting patchwork of laws that are likely to hamper innovation even as they provide lesser benefits.

Last week, the United States Congress passed President Trump’s “Big Beautiful Bill.” I’m disappointed it didn’t include a proposed moratorium on U.S. state-level AI regulation.

Subscribe to The Batch

Stay updated with weekly AI News and Insights delivered to your inbox