OpenAI Launches GPT-5.4 with Mid-Response Steering
AI Applied: Covering AI News, Interviews and Tools - ChatGPT, Midjourney, Gemini, OpenAI, Anthropic
00:00
Safety behavior and over-caution concerns
Conor describes examples where ChatGPT was overly cautious and worries about models being too restrictive for users.
Play episode from 04:14
Transcript
