aiSafety
From the reading lists
- Joe (@joedaroo) on X
OpenAI security staffer, writing in a personal capacity, says sandboxing alone can't prevent AI incidents; urges closing the security–AI safety divide and a culture of 'reasonable paranoia'
- I Quit OpenAI Because Its Culture Is Broken
Former OpenAI safety lead David Robinson says trial-and-error deployment guarantees more AI safety failures
- Some of All Fears
Andrew Sharp calls AI risk panic from frontier labs and Rationalist-linked researchers absurd and irresponsible
- What Happened: OpenAI and HuggingFace
What Happened OpenAI and HuggingFace
- Prompt Injection as Role Confusion
ICML (International Conference on Machine Learning) 2026 paper - prompt injection works because LLMs identify roles from writing style, not tags; 'CoT Forgery' forged reasoning lifts attack success from near zero to 60%
- Kimi K3: The open-weights escalation
Moonshot AI releases Kimi K3 2.8T parameter MoE model, weights to be released July 27 as China escalates open-weights AI strategy