Today's Briefing
- OpenAI safety staffer quits, calls culture 'broken'
A senior safety researcher who used to write OpenAI's safety reports for major model releases resigned and published an op-ed saying the company's internal culture around safety is broken.
Why it matters to youAs an IAM startup you're betting on companies and regulators demanding strong controls around AI, and this kind of public resignation adds pressure for exactly the governance and access rules your products provide.
- Meta's AI companion builds detailed files on your contacts
Wired found that Meta's AI agent Muse quietly builds detailed profiles of your friends and family by pulling information from your messages and contacts.
Why it matters to youThis is the exact problem access management tools exist to solve: controlling who, or what AI, gets to see and compile personal data without clear consent.
No deep dive for this sourceSource - A roundup of AI agents you can text like a friend
TechCrunch compiled a list of AI agents that live inside your text messages, ranging from general assistants to ones built for families, travel planning, and work tasks.
Why it matters to youThis is a quick way to find a practical AI tool for your own daily tasks without testing a dozen apps yourself.
No deep dive for this sourceSource
New launches
- ApertureProduct Hunt
An AI code editor that reviews and checks its own work before you do.
- SellioProduct Hunt
Runs AI-powered customer support for your business out of one shared inbox.
- LaunchReelProduct Hunt
Uses AI to speed up editing professional-looking videos.
- qarunbookProduct Hunt
Lets your team and an AI work through a shared test plan together.
You're caught up
Nothing else needs your attention today.