#1: The one where prompt injection still has no patch
Issue 001: why prompt injection is architectural, a lab that proves it on your own machine, and what the Ollama RCE says about self-hosted AI.
The hackerlogs newsletter
One issue a week on the security of AI systems: what shipped, what broke, and what it means for the systems you run. Written for practitioners, with primary sources linked. No vendor pitches, no sponsored placements in the body.