update chapter 1 content

[cwang0810]

1. Expand on advanced Scaling Law dynamics.
2. In the "Data Quality Quantization" section, add mathematical intuition or diagrams regarding its impact on training convergence speed.
3. Add a "Guide Card": Clarify the core concepts of this chapter (specifically the fundamental differences in data requirements between Pre-training, SFT, and RLHF).