today we are announcing reinforcement finetuning, which makes it really easy to create expert models in specific domains with very little training data. livestream going now: https://
openai.com/12-days/ alpha program starting now, launching publicly in q1

→ View original post on X — @sama