Blog

Thoughts on security, AI, and technology

• Automated Robustness -- When LLMs attack themselves Feb 2, 2026

  Training LLMs to craft and detect prompt injections using PrimeIntellect's Hosted RL