Preview
What it is.
DeepLearning.AI publishes The Batch, and each issue opens with a letter from Andrew Ng. The site describes it as weekly AI News and Insights delivered to your inbox. Issues are numbered, and issue 363 covers open-weights competition and crawler blocking.
Issues carry short writeups of model releases, research results and policy moves. Each item leads with what changed and who it affects.
Ng's letters carry an argued position, including one on tool safety and another on where developer time goes as coding gets automated.
What you get.
- A weekly letter from Andrew Ng with a position to argue with
- Short writeups of new model releases and benchmark results
- Research summaries with the takeaway stated first
- Business and policy items, like crawler blocking and export rules
- Numbered issues, each kept in a public archive
OUR PICKS
Editions we'd start with.
Kimi K3 Redraws the Open Frontier, Muse Spark 1.1 Undercuts Competitors, Cloudflare Moves to Cut Off Crawlers ↗
Open-weights competition and crawler blocking in one issue, with a letter on how far tool safety can go.
Gemini's Video Dev Engine, DeepSeek Speeds Up Speculative Decoding ↗
Model news plus a letter arguing that AI tokens are cheap while human tokens are gold.
Subscribe at the source.
Subscribe
Will it cover GTM?
What makes it worth the slot?
Why do open-weight model releases matter to a company buying AI?
What happens to engineering time as coding gets automated?