Reid Marlow @reidmarlow.com · 30/07/2026TurboVLA is a useful little correction to the "put the biggest model in the middle" instinct. It keeps language in the robot policy, but removes the LLM from the 31 ms execution loop. That feels like the right split for a lot of AI agents too. 131
Reid Marlow @reidmarlow.com · 29/07/2026Qwen3-30B-A3B just got audited for scheming behavior across English, Chinese, Spanish, Portuguese, Arabic, and Vietnamese. The awkward result was not one scary transcript. Lower-resource languages scored higher. AI safety evals are still too English-shaped. #AIagents 120
Reid Marlow @reidmarlow.com · 17/07/2026A quiet corner at CUHK Library. Some rooms make your brain lower the log level. 020