
Synthetic Data and Overcoming Data Limits

OpenAI researcher on why soft skills are the future of work | Karina Nguyen
Topic 2 of 12
Published Feb 9, 2025 · Clip 8:24–12:42 (4:18)
Karina debunks the data wall myth, explaining pre-training compresses world knowledge and post-training uses reinforcement learning for infinite tasks. She describes synthetic data as model-generated examples for rapid iteration, especially for product features, though human data is needed for expert tasks. The discussion covers how this enables continuous model improvement without exhausting internet data.