Context engineering, agent swarms, verification loops: what the AI Engineer Summit revealed about how agentic coding actually works — and where the IDE is going.
Kimi K2 Thinking: Inside Moonshot AI's Agentic Reasoning Model
Kimi K2 Thinking is Moonshot AI's open reasoning agent: 200–300 tool calls, 256K context, INT4 quantization. Architecture, benchmarks, and real use cases.
What matters for RL? Precision! Switching BF16 → FP16
BF16 vs FP16: how switching precision during RL fine-tuning fixes training-inference mismatch, stabilizes GRPO, and why Karpathy applied it to nanochat.