Researchers Debate: Training Data and RL Envs Are All You Need

Mario Zechner argued that "training data/RL envs is all you need" for shaping agent behavior, while @willcb responded that the design space of high-level behavior shaping on top of RL environments is nearly practical, planning to explore the recipes direction next.

2026-10-16 ~ 2026-10-16 · 2 related posts

• badlogic: 'training data/RL envs is all you need' for shaping model behavior — badlogicgames · 2026-10-16
• willcb: higher-level behavior shaping via RL envs is 'almost practical' — willcb · 2026-10-16