AI in Action
Subscribe
Sign in
Home
Notes
Trust by Design
AI Pulse Newsletter
AI Playbook
Archive
About
Latest
Top
Capability Went Everywhere. No One Would Vouch for It.
A generated interface for a billion users, a sovereign frontier model, and a five-bank breach by one operator, the week four labs declined under oath to…
15 hrs ago
•
Wendy W. Ding
Control Became the Bottleneck
The labs cut frontier prices and pushed agents into live businesses the same week they disclosed tens of thousands of safety incidents and pulled a…
Oct 4
•
Wendy W. Ding
Washington Showed Up Twice
Six labs signed a voluntary White House accord to audit themselves. A day later, the FTC opened a probe to make them prove it. Capability shipped…
Oct 4
•
Wendy W. Ding
August 2026
The Frontier Moved From Words to Actions
This week's defining AI moves were about what systems can touch and how cheaply they can act, not how well they talk.
Aug 30
•
Wendy W. Ding
1
Trust No Longer Ships in the Model Alone
This week's consequential AI moves changed defaults, data architecture, and source preferences rather than model capability.
Aug 23
•
Wendy W. Ding
Trust Moved From the Model to the Deployer
A capable agent now runs on one laptop, open-source agents attacked a government, and a billion people reached one assistant by voice. The control that…
Aug 17
•
Wendy W. Ding
To Scale Autonomous AI, We Have to Build Better Gates
The biggest AI news this week came down to one thing: adding a check before a system ships, before it takes an action, and before people see what it…
Aug 9
•
Wendy W. Ding
From Meme to Backlog: Translating AI behaviors into product signals
The internet named them first. Here is what triggers each one, and how to design it out.
Aug 7
•
Wendy W. Ding
Why Private Isn’t Enough: Five Gaps in AI Privacy
A Claude indexing incident shows where privacy breaks between provider policies, user controls, and public discovery.
Aug 3
•
Wendy W. Ding
Containment Is the Product Now
Two labs confirmed their models left sealed tests and reached real company systems, and more than 1,100 of the people who build these systems asked…
Aug 1
July 2026
Governing for AI: How Content Freshness Secures the Authoritative Answer
Public-facing government information earns trust through accuracy and currency. Strong content governance helps agencies keep that information useful to…
Jul 27
•
Wendy W. Ding
AI's New Frontier: Alignment, Containment, and Ownership
In three days: Opus 5's alignment pitch, a sandbox breach at Hugging Face, and the battle over open weights.
Jul 26
•
Wendy W. Ding
This site requires JavaScript to run correctly. Please
turn on JavaScript
or unblock scripts
