https://firethering.com/anthropic-nla-claude-thoughts-interpretability/Firetheringen2026-05-09T09:31:00+00:00Claude Knew It Was Being Tested. It Just Didn’t Say So. Anthropic Built a Tool to Find Out. https://firethering.com/openai-new-voice-models-realtime-api/Firetheringen2026-05-08T12:32:50+00:00OpenAI’s New Voice Models Want to Do More Than Talk Back https://firethering.com/subq-12m-token-context-llm-subquadratic-attention/Firetheringen2026-05-08T08:18:33+00:00SubQ’s 12M Token Model Could Change How AI Handles Long Context. If It’s Real. https://firethering.com/claude-spacex-compute-deal/Firetheringen2026-05-07T16:01:10+00:00Claude Just Doubled Its Usage Limits. The Real Story Is the SpaceX Deal Behind It