The AI Tea

Newsletter

Two useful weekly brews.

The AI changes worth carrying into next week, with original sources and practical context for work and learning.

Why subscribe

Useful AI context without the tab overload

A focused Learning edition every Tuesday.

Each email links back to the full briefing with source dates, freshness checks, and plain-English context.

A practical Work edition every Thursday.

Each email links back to the full briefing with source dates, freshness checks, and plain-English context.

Original source links and freshness notes.

Each email links back to the full briefing with source dates, freshness checks, and plain-English context.

Choose either edition or receive both.

Each email links back to the full briefing with source dates, freshness checks, and plain-English context.

Sample digest

What lands in your inbox

01 · AI Safety OpenAI proposes safety cases for AI training

OpenAI published early guidance on “safety cases” for training advanced AI, covering technical safeguards, operating practices, and how to probe incidents.

02 · AI Tools Hugging Face posts Holo4 update

Hugging Face published a post titled “Holo4: powering generalist computer-use agents.” Details beyond the title weren’t provided in the source packet.

03 · AI Agents When AI code judges don’t have a basis

A new paper tests a multi-step “verification” judge for code and finds it often can’t distinguish solutions. A simple log-based check helps it refuse to guess.