Andrew Ng’s The Batch weekly newsletter and DeepLearning.AI courses are the mainstream educator’s view of AI and agentic development, the voice that popularized the four agentic design patterns and published the AI Engineering Skills Map.
He is the single highest-reaching educator in this category: where the others write for practitioners already inside the field, Ng is the on-ramp that also steers what the broader industry thinks agents are.
What it is #
A weekly newsletter, a course platform, and the personal letters of Andrew Ng, a computer scientist and educator, co-founder of Coursera and founder of DeepLearning.AI, who taught the foundational Machine Learning course to millions.
The Batch is a free weekly roundup of AI news and analysis with his “Letters”.
His
agentic AI course and
AI Engineering Skills Map define the vocabulary many teams use for designing agentic workflows.
Status #
Active and highly influential as of 2026-10-03. The Batch publishes weekly, with the latest numbered issue (373, “GLM-5.3’s ExploitBench Score is Good News for Cybersecurity”) on 2026-10-02, and the AI Engineering Skills Map letter is dated 2026-08-14. DeepLearning.AI reports courses taught to millions of learners and positions itself as the canonical AI education path.
Strengths #
- The weekly newsletter is a convenient, reasonably complete scan of the AI landscape with an engineering takeaway, not raw news.
- He normalized the four agentic design patterns, reflection, tool use, planning, and multi-agent collaboration, which are now common vocabulary in the field.
- The short courses are vendor-neutral and hands-on, which makes them an effective and low-friction way to learn.
- The AI Engineering Skills Map is a rare, concrete statement of what an AI engineer needs to know in 2026.
Cautions #
- The newsletter and packaging are broadly optimistic, and a reader who wants deep skepticism about AI hype will not get it here.
- Education is also his business, so course promotion and cert marketing ride alongside the editorial content.
- Surveys and short courses are broad on-ramps, not deep dives, and certificates do not substitute for shipped work when hiring.
- His mainstream framing can flatten the messy, tool-specific reality that practitioner sources report firsthand.
Pricing #
The Batch newsletter is free. DeepLearning.AI short courses are mostly free, while Coursera specializations and certificates are paid ($49/month or $399/year as of 2026-06 review).
Price history #
| Date | Plan | Change | Source |
|---|---|---|---|
| 2026-06 | Coursera specializations | Baseline: The Batch and most short courses free; Coursera specializations and certificates $49/mo or $399/yr. |
Compared to #
- Latent Space: the industry-podcast-and-newsletter of record versus the educator’s weekly; Latent Space is deeper and more technical, The Batch is broader and more accessible.
- Simon Willison: the educator’s survey versus the practitioner’s day-to-day feed; Willison is the hands-on ground truth, Ng is the structured overview.
- Chip Huyen: both teach AI engineering, but Huyen writes the systems reference while Ng runs the mainstream education and media operation.
Bottom line #
Recommended as the balanced weekly scan and the best-structured way to pick up agentic skills when you are new to them. Not for a practitioner who wants deep technical detail, frontier-labs nuance, or sustained skepticism.
Top 5 recommended reading #
The Batch - The weekly newsletter itself, a balanced scan of the AI landscape with an engineering takeaway in every issue.
Agentic AI - The course that teaches the four agentic design patterns and the eval-driven process for building agentic workflows.
Changes #
- 2026-08-29 - Created as the education-and-on-ramp band, the voice that popularized the four agentic design patterns and published the AI Engineering Skills Map.
- 2026-08-30 - Dropped an unverifiable jobpocalypse quote and a dead aiagentrank reference, replaced with the verified Coursera specialization page.
- 2026-09-04 - Batch coverage extended past the note’s “through August 2026” claim.
- 2026-09-20 - Added the Price history section tracking price changes in a table, per the new owner rule.
- 2026-09-24 - Added the Top 5 recommended reading section.
- 2026-09-27 - Batch recency refreshed to issue 372 (2026-09-25); replaced the dead zenvanriel.com review link with the Ask HN Coursera thread after the original page 404’d with no archive.
- 2026-10-03 - Batch recency refreshed to issue 373 (2026-10-02, “GLM-5.3’s ExploitBench Score is Good News for Cybersecurity”); issue URL fetched this run.
See also #
- Agentic Coding Tools Landscape - the tools his agentic design patterns describe
- Model Selection for Coding Tasks - the model thinking his Skills Map complements
- Latent Space - the deeper, more technical industry voice on the same landscape
- Rethinking Code Review in the Age of LLMs - the workflow his agentic design patterns imply
References #
https://www.coursera.org/specializations/deep-learning - Coursera’s page for the DeepLearning.AI specialization, grounding that the specializations and certificates are paid (the $49/month or $399/year figure keeps its 2026-06 as-of date, since Coursera renders prices client-side)
https://news.ycombinator.com/item?id=11461370 - the Ask HN thread on whether Coursera courses are worth paying for, grounding the caution that certificates do not substitute for shipped work (replaces a zenvanriel.com review that now 404s with no archive)