RT Jared Friedman: Claude was RL'd on coding sessions with humans. But the more I use Claude Code, the more I realize I'm being RL'd on Claude Code - ...
Jared Friedman observes that while Claude was trained on human coding sessions via reinforcement learning (RL), his own use of Claude Code has become a reciprocal learning process—he adapts his instructions to the tool's strengths and weaknesses, describing it as a form of human/AI co-evolution.
Background
- **Reinforcement Learning (RL)** is a machine-learning technique where an AI model is trained by rewarding desired behaviors and penalizing mistakes, gradually improving its performance through trial and error.<br />- **Claude** is an AI assistant created by Anthropic. **Claude Code** is a specialized tool that helps programmers write and debug code via natural-language instructions.<br />- The tweet suggests a two-way adaptation: Claude was trained ("RL'd") on human coding sessions, but the human user is now also being implicitly "trained" — adjusting their own communication style to get better results from the AI.<br />- **Human/AI co-evolution** refers to the idea that as we design AI to better understand us, we simultaneously adapt our behavior to work more effectively with AI — both sides changing in response to each other.