Luo Fuli: The AI Paradigm Shift & the Agent Era

Why it matters

Luo Fuli's first long-form interview — ex-DeepSeek researcher, now head of Xiaomi's LLM team. In 2026 the paradigm shifts from the pre-training Chat era to the post-training Agent era. 3.5 hours on the shocks triggered by OpenClaw and other variables, why the agent paradigm 'depends on post-training', GPU allocation, and organizational egalitarianism.

Key takeaways

• Paradigm shift: from pre-training-dominated Chat era to post-training Agent era.
• The agent paradigm depends on post-training: data and alignment set the ceiling.
• The OpenClaw shockwave and its structural impact.
• GPU allocation: the political economy of compute inside LLM teams.
• Organizational egalitarianism: keeping tech orgs from freezing into hierarchy.