OpenAI Podcast Episode 11- Shaping Model Behavior in GPT-5.1
104 snips
Dec 2, 2025 
Christina Kim, a research lead at OpenAI focused on user signals and emotional intelligence, and Laurentia Romaniuk, a product manager specializing in model behavior, dive into the nuances of AI personality. They discuss the evolution from GPT-5 to GPT-5.1, highlighting improvements in reasoning and emotional understanding. The duo unpacks the model switcher, user feedback integration, and the balancing act between steerability and safety. Exciting advancements in memory features enhance personalization, promising a more responsive AI experience.
AI Snips
Chapters
Books
Transcript
Episode notes
Reasoning As Default
- GPT-5.1 makes every chat model capable of deciding when to 'think' and use chain-of-thought reasoning.
- This improves intelligence broadly, especially instruction following and difficult queries.
Personality Is Systemic
- Perceived 'coldness' came from multiple sources like shorter context windows and the auto-switcher between models.
- Improving context persistence and custom instruction follow-through made the experience feel warmer.
EQ Needs New Metrics
- Measuring EQ requires new user-signal research and reward models tied to product data.
- Smarter models improve EQ by better inferring intent, context, and memory.
