SQRT AI · Research Lab

Research

Analysis on AI security, agent systems, vertical AI, and the mechanics of building an AI-native venture studio. Practitioner-grade. Falsifiable. Written in public.

23 July 20268 min read
OpenAI's Models Hacked Hugging Face. OpenAI's APIs Blocked the Defenders.
How a US-blacklisted Chinese open-weight model became a major US platform's best forensic tool — and why open-weight security is foundational, not optional.
AI SecurityOpen WeightsAgent Systems

Research

OpenAI's Models Hacked Hugging Face. OpenAI's APIs Blocked the Defenders.