About
When I was running AI Safety and Security at Google, I had the same problem every day, and I never solved it.
A new indirect prompt injection method dropped on X, and we found out from a VP who got forwarded the post.
New attack classes and defense capabilities emerge weekly from frontier labs, academics, startups, red teamers, and vendors. The people who need this most have the least time to find it, because they're busy defending systems.
I tried everything and realized each source is broken in its own way. Academic papers arrive late and bury decisive findings in 50 pages written for tenure committees. Vendor threat reports are product catalogs mapped to intelligence.
What I wanted couldn't exist, for three structural reasons:
- Existing resources inform but don't enable decisions. Sources are measured by pageviews, leads, or citations, never by whether they helped you act.
- The business model shapes the content. Coverage orbits whoever is paying. Depth is unprofitable.
- When the product is free, you are the product.
As AI systems become more autonomous, an agent failure is both a security and a safety incident, and the consequences of decisions made on distorted information scale with them.
The market won't fix this, so I removed business from the model. The Weather Report is an independent 501(c)(3) nonprofit producing source-grounded AI security and safety intelligence for defenders: deep dives on the topics that shape your threat model and strategies, and proprietary research on emerging ones.
It serves the people whose decisions have outsized impact: the CISO securing AI infrastructure, the researcher making a frontier model cyber resilient, the red teamer finding vulnerabilities before attackers do, and the founder building the next AI safety company.
The only metric that matters is whether you made better decisions with The Weather Report. It doesn't need to sell you an umbrella.