Add a `WHY.md` File for Each Benchmark

[tareknaser]

Currently, the root-level README lists all benchmarks with a brief one-liner description. Each benchmark’s folder contains a large README where the motivation and rationale are buried within the details.

We should create a dedicated WHY.md file inside each benchmark’s folder. This file should explain:

• Why the benchmark matters
• How it fits into our broader vision of system intelligence

See #21 for a similar file

We should then link each WHY.md file from the corresponding bullet point in the root README.